From gfrost at openjdk.org  Wed Oct  1 10:53:09 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Wed, 1 Oct 2025 10:53:09 GMT
Subject: git: openjdk/babylon: code-reflection: Hat move config to core
Message-ID: <d86a9517-cf39-4365-872b-ecc570b567d4@openjdk.org>

Changeset: d288efa1
Branch: code-reflection
Author:    Gary Frost <gfrost at openjdk.org>
Date:      2025-10-01 10:51:36 +0000
URL:       https://git.openjdk.org/babylon/commit/d288efa194a2eac83061760abd9ae053e2289199

Hat move config to core

! hat/backends/ffi/cuda/src/main/java/hat/backend/ffi/CudaBackend.java
! hat/backends/ffi/mock/src/main/java/hat/backend/ffi/MockBackend.java
! hat/backends/ffi/opencl/src/main/java/hat/backend/ffi/OpenCLBackend.java
! hat/backends/ffi/shared/src/main/java/hat/backend/ffi/C99FFIBackend.java
! hat/backends/ffi/shared/src/main/java/hat/backend/ffi/FFIBackend.java
! hat/backends/ffi/shared/src/main/java/hat/backend/ffi/FFIBackendDriver.java
- hat/backends/ffi/shared/src/main/java/hat/backend/ffi/FFIConfig.java
! hat/backends/ffi/shared/src/main/native/include/config.h
! hat/core/src/main/java/hat/Config.java
+ hat/core/src/main/java/hat/FFIConfigCreator.java
+ hat/core/src/main/java/hat/codebuilders/C99HATConfigBuilder.java
! hat/core/src/main/java/hat/codebuilders/CodeBuilder.java
! hat/core/src/main/java/hat/codebuilders/HATCodeBuilder.java
! hat/examples/experiments/src/main/java/experiments/Mesh.java
! hat/examples/experiments/src/main/java/experiments/MinBufferTest.java
! hat/hat/bld
! hat/hat/clean
= hat/hat/obld
= hat/hat/oclean
= hat/hat/orun
= hat/hat/osanity
= hat/hat/otest
! hat/hat/run
! hat/hat/sanity
! hat/hat/test


From gfrost at openjdk.org  Wed Oct  1 10:54:51 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Wed, 1 Oct 2025 10:54:51 GMT
Subject: [code-reflection] Integrated: Hat move config to core
Message-ID: <Z0h8Ng9fZ5Vk_Mq9q4U4thByLw78CrMqn5g7S4HuHxk=.631bcd71-5edd-4da9-9915-b2d3bcf0302b@github.com>

By moving Config up to core (from FFI) we can take advantage of it from anywhere in the HAT code. 

At present we still only access it from FFI.  I plan to make more general changes in future. 

Also switched `java @hat/bld` to actually use the new `job.jar` based build (same for @hat/run ... sanity .. test clean etc) 

Have not removed the Script based tools. But will soon.  If you see a problem you can revert to the prev version using  `java @hat/obld ....` instead of `java @hat/bld ...`

-------------

Commit messages:
 - Moved config out of ffi also switched to using job.jar based bld,run etc
 - Moved config out of ffi also switched to using job.jar based bld,run etc

Changes: https://git.openjdk.org/babylon/pull/589/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=589&range=00
  Stats: 804 lines in 25 files changed: 382 ins; 325 del; 97 mod
  Patch: https://git.openjdk.org/babylon/pull/589.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/589/head:pull/589

PR: https://git.openjdk.org/babylon/pull/589

From gfrost at openjdk.org  Wed Oct  1 10:54:53 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Wed, 1 Oct 2025 10:54:53 GMT
Subject: [code-reflection] Integrated: Hat move config to core
In-Reply-To: <Z0h8Ng9fZ5Vk_Mq9q4U4thByLw78CrMqn5g7S4HuHxk=.631bcd71-5edd-4da9-9915-b2d3bcf0302b@github.com>
References: <Z0h8Ng9fZ5Vk_Mq9q4U4thByLw78CrMqn5g7S4HuHxk=.631bcd71-5edd-4da9-9915-b2d3bcf0302b@github.com>
Message-ID: <GAAStKGy4d0vbuLoy-QDipIHUS7JGiOnwALoeFiXPak=.bfc333ad-af61-4deb-aabc-bd00d33a473e@github.com>

On Wed, 1 Oct 2025 10:48:03 GMT, Gary Frost <gfrost at openjdk.org> wrote:

> By moving Config up to core (from FFI) we can take advantage of it from anywhere in the HAT code. 
> 
> At present we still only access it from FFI.  I plan to make more general changes in future. 
> 
> Also switched `java @hat/bld` to actually use the new `job.jar` based build (same for @hat/run ... sanity .. test clean etc) 
> 
> Have not removed the Script based tools. But will soon.  If you see a problem you can revert to the prev version using  `java @hat/obld ....` instead of `java @hat/bld ...`

This pull request has now been integrated.

Changeset: d288efa1
Author:    Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/d288efa194a2eac83061760abd9ae053e2289199
Stats:     804 lines in 25 files changed: 382 ins; 325 del; 97 mod

Hat move config to core

-------------

PR: https://git.openjdk.org/babylon/pull/589

From jfumero at openjdk.org  Wed Oct  1 11:21:18 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Wed, 1 Oct 2025 11:21:18 GMT
Subject: [code-reflection] RFR: [hat] Example of HAT for 2D Matmul with
 register tiling
Message-ID: <d78Kd2izQKFqw4ThoHp5szbEiDlu4OzKqAWkTPVuzf4=.ca129223-8990-4d23-940c-7c99dc8c0014@github.com>

There was an example with accesses to local memory specific to the CUDA programming model. This has been fixed and now it is compatible with both the OpenCL and CUDA backends of HAT. 

How to run?

HAT=SHOW_CODE java @hat/run ffi-opencl matmul 2DREGISTERTILING

-------------

Commit messages:
 - [hat] Example of HAT for 2D Matmul with register tiling

Changes: https://git.openjdk.org/babylon/pull/590/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=590&range=00
  Stats: 295 lines in 5 files changed: 196 ins; 75 del; 24 mod
  Patch: https://git.openjdk.org/babylon/pull/590.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/590/head:pull/590

PR: https://git.openjdk.org/babylon/pull/590

From jfumero at openjdk.org  Wed Oct  1 11:54:42 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Wed, 1 Oct 2025 11:54:42 GMT
Subject: git: openjdk/babylon: code-reflection: [hat] Example of HAT for 2D
 Matmul with register tiling
Message-ID: <4ca230e5-4c86-444a-b77d-bdb6894cda19@openjdk.org>

Changeset: 8e9bc5b2
Branch: code-reflection
Author:    Juan Fumero <jfumero at openjdk.org>
Date:      2025-10-01 11:52:48 +0000
URL:       https://git.openjdk.org/babylon/commit/8e9bc5b2db96aa7b2f0879722237ecbfe98b841d

[hat] Example of HAT for 2D Matmul with register tiling

! hat/backends/ffi/shared/src/main/java/hat/backend/ffi/C99FFIBackend.java
- hat/core/src/main/java/hat/CUDARange.java
! hat/examples/matmul/src/main/java/matmul/Main.java
! hat/tests/src/main/java/oracle/code/hat/TestArrays.java
! hat/tests/src/main/java/oracle/code/hat/TestMatMul.java


From jfumero at openjdk.org  Wed Oct  1 11:55:29 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Wed, 1 Oct 2025 11:55:29 GMT
Subject: [code-reflection] Integrated: [hat] Example of HAT for 2D Matmul with
 register tiling
In-Reply-To: <d78Kd2izQKFqw4ThoHp5szbEiDlu4OzKqAWkTPVuzf4=.ca129223-8990-4d23-940c-7c99dc8c0014@github.com>
References: <d78Kd2izQKFqw4ThoHp5szbEiDlu4OzKqAWkTPVuzf4=.ca129223-8990-4d23-940c-7c99dc8c0014@github.com>
Message-ID: <Qf1dBlgqR23smNmOP56iw0Y4ZpRsUGjYFI0qV87B-I0=.b6432d25-fb66-42fc-a3c6-256313b4f25b@github.com>

On Wed, 1 Oct 2025 11:14:43 GMT, Juan Fumero <jfumero at openjdk.org> wrote:

> There was an example with accesses to local memory specific to the CUDA programming model. This has been fixed and now it is compatible with both the OpenCL and CUDA backends of HAT. 
> 
> How to run?
> 
> HAT=SHOW_CODE java @hat/run ffi-opencl matmul 2DREGISTERTILING

This pull request has now been integrated.

Changeset: 8e9bc5b2
Author:    Juan Fumero <jfumero at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/8e9bc5b2db96aa7b2f0879722237ecbfe98b841d
Stats:     295 lines in 5 files changed: 196 ins; 75 del; 24 mod

[hat] Example of HAT for 2D Matmul with register tiling

-------------

PR: https://git.openjdk.org/babylon/pull/590

From jfumero at openjdk.org  Wed Oct  1 13:30:34 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Wed, 1 Oct 2025 13:30:34 GMT
Subject: [code-reflection] RFR: [hat] Fix cuda set context
Message-ID: <3mMhtRkpi8-TxcFbC8yxFs11XX9wFZhNFoCnN21OCC4=.72529812-b27d-463e-bbc8-f59d20dfc184@github.com>

When running Java multi-threaded applications with the HAT CUDA Backend we need to set the current context before building the kernel. This patch fixes this. 

Heal application should work now for the CUDA backend.

-------------

Commit messages:
 - [hat] Fix cuda set context

Changes: https://git.openjdk.org/babylon/pull/591/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=591&range=00
  Stats: 4 lines in 3 files changed: 1 ins; 1 del; 2 mod
  Patch: https://git.openjdk.org/babylon/pull/591.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/591/head:pull/591

PR: https://git.openjdk.org/babylon/pull/591

From jfumero at openjdk.org  Wed Oct  1 13:48:27 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Wed, 1 Oct 2025 13:48:27 GMT
Subject: git: openjdk/babylon: code-reflection: [hat] Fix cuda set context
Message-ID: <879a0ee0-42f4-4706-9bba-490d3f172254@openjdk.org>

Changeset: 0afb4ad7
Branch: code-reflection
Author:    Juan Fumero <jfumero at openjdk.org>
Date:      2025-10-01 13:45:50 +0000
URL:       https://git.openjdk.org/babylon/commit/0afb4ad7b6c65c923288299abf3b72e0c324244f

[hat] Fix cuda set context

! hat/backends/ffi/cuda/src/main/native/cpp/cuda_backend.cpp
! hat/backends/ffi/cuda/src/main/native/cpp/cuda_backend_queue.cpp
! hat/backends/ffi/shared/src/main/java/hat/backend/ffi/FFIBackendDriver.java


From jfumero at openjdk.org  Wed Oct  1 13:48:57 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Wed, 1 Oct 2025 13:48:57 GMT
Subject: [code-reflection] Integrated: [hat] Fix cuda set context
In-Reply-To: <3mMhtRkpi8-TxcFbC8yxFs11XX9wFZhNFoCnN21OCC4=.72529812-b27d-463e-bbc8-f59d20dfc184@github.com>
References: <3mMhtRkpi8-TxcFbC8yxFs11XX9wFZhNFoCnN21OCC4=.72529812-b27d-463e-bbc8-f59d20dfc184@github.com>
Message-ID: <O92he_fMx3KGD7KAYnBcraklF6Nwf8E50WGaOF-vqXw=.dfe99045-ced8-4e78-a201-6ceda3466c0d@github.com>

On Wed, 1 Oct 2025 13:24:47 GMT, Juan Fumero <jfumero at openjdk.org> wrote:

> When running Java multi-threaded applications with the HAT CUDA Backend we need to set the current context before building the kernel. This patch fixes this. 
> 
> Heal application should work now for the CUDA backend.

This pull request has now been integrated.

Changeset: 0afb4ad7
Author:    Juan Fumero <jfumero at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/0afb4ad7b6c65c923288299abf3b72e0c324244f
Stats:     4 lines in 3 files changed: 1 ins; 1 del; 2 mod

[hat] Fix cuda set context

-------------

PR: https://git.openjdk.org/babylon/pull/591

From jfumero at openjdk.org  Wed Oct  1 14:09:20 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Wed, 1 Oct 2025 14:09:20 GMT
Subject: [code-reflection] RFR: [hat] rollback CUDA specific scheduler
Message-ID: <7BDFWP9sWCcL0ZXx19oZcjuiVMnRwRcRXXeBHCOcMIY=.a3fdc886-33f7-4e22-ba86-0db2cd3130c5@github.com>

Remove the CUDA Specific scheduler from the kernel context.

-------------

Commit messages:
 - [hat] rollback CUDA specific scheduler

Changes: https://git.openjdk.org/babylon/pull/592/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=592&range=00
  Stats: 17 lines in 3 files changed: 0 ins; 16 del; 1 mod
  Patch: https://git.openjdk.org/babylon/pull/592.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/592/head:pull/592

PR: https://git.openjdk.org/babylon/pull/592

From gfrost at openjdk.org  Wed Oct  1 14:25:04 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Wed, 1 Oct 2025 14:25:04 GMT
Subject: [code-reflection] Integrated: initial prototype of prefix sum
 experiment
Message-ID: <ORChY7_f2PKENVttsmmf-vdKbhfBDzvZyv6LTtPCxJ4=.a974f513-8048-446b-ad2d-25cc802dfe3b@github.com>

This will be a good demonstrator for recent  array view and local memory work. 

Also the basis for a reduction based violajones implementation. 

I also added matmul.iml so we can view matmul in intellij project

-------------

Commit messages:
 - initial prototype of prefix sum experiment

Changes: https://git.openjdk.org/babylon/pull/593/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=593&range=00
  Stats: 341 lines in 3 files changed: 341 ins; 0 del; 0 mod
  Patch: https://git.openjdk.org/babylon/pull/593.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/593/head:pull/593

PR: https://git.openjdk.org/babylon/pull/593

From gfrost at openjdk.org  Wed Oct  1 14:25:05 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Wed, 1 Oct 2025 14:25:05 GMT
Subject: [code-reflection] Integrated: initial prototype of prefix sum
 experiment
In-Reply-To: <ORChY7_f2PKENVttsmmf-vdKbhfBDzvZyv6LTtPCxJ4=.a974f513-8048-446b-ad2d-25cc802dfe3b@github.com>
References: <ORChY7_f2PKENVttsmmf-vdKbhfBDzvZyv6LTtPCxJ4=.a974f513-8048-446b-ad2d-25cc802dfe3b@github.com>
Message-ID: <p_YUMzuYb7eYMFHKezzR9QtAL1cbPgdUqZ9YWlBxY44=.07abeb4c-6190-4c8d-bf47-4eee045cd8be@github.com>

On Wed, 1 Oct 2025 14:04:03 GMT, Gary Frost <gfrost at openjdk.org> wrote:

> This will be a good demonstrator for recent  array view and local memory work. 
> 
> Also the basis for a reduction based violajones implementation. 
> 
> I also added matmul.iml so we can view matmul in intellij project

This pull request has now been integrated.

Changeset: 161d5835
Author:    Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/161d583585ceb2e2ecca2c95649acd59560c1bf9
Stats:     341 lines in 3 files changed: 341 ins; 0 del; 0 mod

initial prototype of prefix sum experiment

-------------

PR: https://git.openjdk.org/babylon/pull/593

From gfrost at openjdk.org  Wed Oct  1 14:25:59 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Wed, 1 Oct 2025 14:25:59 GMT
Subject: git: openjdk/babylon: code-reflection: initial prototype of prefix sum
 experiment
Message-ID: <0278c455-c966-47cd-b5c3-521f2fb460b3@openjdk.org>

Changeset: 161d5835
Branch: code-reflection
Author:    Gary Frost <gfrost at openjdk.org>
Date:      2025-10-01 14:22:01 +0000
URL:       https://git.openjdk.org/babylon/commit/161d583585ceb2e2ecca2c95649acd59560c1bf9

initial prototype of prefix sum experiment

+ hat/examples/experiments/src/main/java/experiments/PrefixSum.java
! hat/intellij/.idea/modules.xml
+ hat/intellij/example_matmul.iml


From jfumero at openjdk.org  Wed Oct  1 14:33:12 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Wed, 1 Oct 2025 14:33:12 GMT
Subject: git: openjdk/babylon: code-reflection: [hat] rollback CUDA specific
 scheduler
Message-ID: <5ccfb1dc-f47a-455b-a6a9-6c0fbea96b6c@openjdk.org>

Changeset: 6aa54542
Branch: code-reflection
Author:    Juan Fumero <jfumero at openjdk.org>
Date:      2025-10-01 14:30:54 +0000
URL:       https://git.openjdk.org/babylon/commit/6aa54542593a463a36d1e8b54b190519ab05e86c

[hat] rollback CUDA specific scheduler

! hat/backends/ffi/cuda/src/main/native/cpp/cuda_backend_queue.cpp
! hat/backends/ffi/shared/src/main/native/include/shared.h
! hat/core/src/main/java/hat/buffer/KernelContext.java


From jfumero at openjdk.org  Wed Oct  1 14:33:35 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Wed, 1 Oct 2025 14:33:35 GMT
Subject: [code-reflection] Integrated: [hat] rollback CUDA specific scheduler
In-Reply-To: <7BDFWP9sWCcL0ZXx19oZcjuiVMnRwRcRXXeBHCOcMIY=.a3fdc886-33f7-4e22-ba86-0db2cd3130c5@github.com>
References: <7BDFWP9sWCcL0ZXx19oZcjuiVMnRwRcRXXeBHCOcMIY=.a3fdc886-33f7-4e22-ba86-0db2cd3130c5@github.com>
Message-ID: <y06bnl3Vq66mGTuMyMJXhTY3G-H1JcSmr4mqQMYXmzs=.0200e056-3f17-4cf0-ba87-3419acfc42bf@github.com>

On Wed, 1 Oct 2025 14:01:00 GMT, Juan Fumero <jfumero at openjdk.org> wrote:

> Remove the CUDA Specific scheduler from the kernel context.

This pull request has now been integrated.

Changeset: 6aa54542
Author:    Juan Fumero <jfumero at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/6aa54542593a463a36d1e8b54b190519ab05e86c
Stats:     17 lines in 3 files changed: 0 ins; 16 del; 1 mod

[hat] rollback CUDA specific scheduler

-------------

PR: https://git.openjdk.org/babylon/pull/592

From jfumero at openjdk.org  Wed Oct  1 16:47:58 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Wed, 1 Oct 2025 16:47:58 GMT
Subject: [code-reflection] RFR: [hat] clean-up - removed duplicated sample
Message-ID: <k8t58c2O9X6KDABHpul3yjO2kppzBKntDPrUcynzLyw=.0e38f58d-01bb-4750-b194-7a10a7f54b5f@github.com>

Clean-up samples - removed duplicated version

-------------

Commit messages:
 - [hat] clean-up - removed duplicated sample

Changes: https://git.openjdk.org/babylon/pull/594/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=594&range=00
  Stats: 445 lines in 2 files changed: 0 ins; 444 del; 1 mod
  Patch: https://git.openjdk.org/babylon/pull/594.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/594/head:pull/594

PR: https://git.openjdk.org/babylon/pull/594

From jfumero at openjdk.org  Wed Oct  1 16:48:55 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Wed, 1 Oct 2025 16:48:55 GMT
Subject: git: openjdk/babylon: code-reflection: [hat] clean-up - removed
 duplicated sample
Message-ID: <4fe7ae54-2e97-4c41-b63c-872694d4c5c5@openjdk.org>

Changeset: 1cfba3b3
Branch: code-reflection
Author:    Juan Fumero <jfumero at openjdk.org>
Date:      2025-10-01 16:48:05 +0000
URL:       https://git.openjdk.org/babylon/commit/1cfba3b3a7acb15babae5b55b4b3a1eefd48d8c3

[hat] clean-up - removed duplicated sample

! hat/examples/matmul/src/main/java/matmul/Main.java
! hat/tests/src/main/java/oracle/code/hat/TestMatMul.java


From jfumero at openjdk.org  Wed Oct  1 16:50:29 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Wed, 1 Oct 2025 16:50:29 GMT
Subject: [code-reflection] Integrated: [hat] clean-up - removed duplicated
 sample
In-Reply-To: <k8t58c2O9X6KDABHpul3yjO2kppzBKntDPrUcynzLyw=.0e38f58d-01bb-4750-b194-7a10a7f54b5f@github.com>
References: <k8t58c2O9X6KDABHpul3yjO2kppzBKntDPrUcynzLyw=.0e38f58d-01bb-4750-b194-7a10a7f54b5f@github.com>
Message-ID: <Gy0LmfuaQ7TeyCikoVEr9fTO8K07723GYMlG3ov6Sak=.3238a9ed-b87e-42c3-8843-dc999f562bb7@github.com>

On Wed, 1 Oct 2025 16:42:19 GMT, Juan Fumero <jfumero at openjdk.org> wrote:

> Clean-up samples - removed duplicated version

This pull request has now been integrated.

Changeset: 1cfba3b3
Author:    Juan Fumero <jfumero at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/1cfba3b3a7acb15babae5b55b4b3a1eefd48d8c3
Stats:     445 lines in 2 files changed: 0 ins; 444 del; 1 mod

[hat] clean-up - removed duplicated sample

-------------

PR: https://git.openjdk.org/babylon/pull/594

From jfumero at openjdk.org  Thu Oct  2 08:27:02 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Thu, 2 Oct 2025 08:27:02 GMT
Subject: [code-reflection] RFR: [hat] testing framework updated to use the new
 scripts
Message-ID: <sHdLO5GOwWWaMupJEFj1Lt_kJTstznwNYbEkjH_jXKY=.db04cf89-8ed7-4a31-a441-c965cc364b54@github.com>

Testing framework updated to use the new scripts

-------------

Commit messages:
 - [hat] testing framework updated to use the new scripts

Changes: https://git.openjdk.org/babylon/pull/595/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=595&range=00
  Stats: 2 lines in 1 file changed: 0 ins; 0 del; 2 mod
  Patch: https://git.openjdk.org/babylon/pull/595.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/595/head:pull/595

PR: https://git.openjdk.org/babylon/pull/595

From jfumero at openjdk.org  Thu Oct  2 08:33:59 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Thu, 2 Oct 2025 08:33:59 GMT
Subject: [code-reflection] Integrated: [hat] testing framework updated to use
 the new scripts
In-Reply-To: <sHdLO5GOwWWaMupJEFj1Lt_kJTstznwNYbEkjH_jXKY=.db04cf89-8ed7-4a31-a441-c965cc364b54@github.com>
References: <sHdLO5GOwWWaMupJEFj1Lt_kJTstznwNYbEkjH_jXKY=.db04cf89-8ed7-4a31-a441-c965cc364b54@github.com>
Message-ID: <FVcWJ1_lv-bWYh9lLPGiZ__dpJR2wlXQjQlHuRHyz5o=.de0c7507-f475-41b2-bdd9-b50bd4b2d384@github.com>

On Thu, 2 Oct 2025 08:21:13 GMT, Juan Fumero <jfumero at openjdk.org> wrote:

> Testing framework updated to use the new scripts

This pull request has now been integrated.

Changeset: 788baa20
Author:    Juan Fumero <jfumero at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/788baa204e68c0099f9856814f38ce45f2e3aa4e
Stats:     2 lines in 1 file changed: 0 ins; 0 del; 2 mod

[hat] testing framework updated to use the new scripts

-------------

PR: https://git.openjdk.org/babylon/pull/595

From jfumero at openjdk.org  Thu Oct  2 08:34:39 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Thu, 2 Oct 2025 08:34:39 GMT
Subject: git: openjdk/babylon: code-reflection: [hat] testing framework updated
 to use the new scripts
Message-ID: <f7368c90-3e4c-42ff-9a07-331287acf8f7@openjdk.org>

Changeset: 788baa20
Branch: code-reflection
Author:    Juan Fumero <jfumero at openjdk.org>
Date:      2025-10-02 08:31:13 +0000
URL:       https://git.openjdk.org/babylon/commit/788baa204e68c0099f9856814f38ce45f2e3aa4e

[hat] testing framework updated to use the new scripts

! hat/scripts/remoteTesting.sh


From jfumero at openjdk.org  Thu Oct  2 10:41:36 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Thu, 2 Oct 2025 10:41:36 GMT
Subject: [code-reflection] Integrated: [hat] Remote tests script fixed + add
 ArrayViews
Message-ID: <u_pmFb-V-TOhkWTREoSjROaNe2BzJxy7HkLsqs7IURI=.32f9f45f-78b6-452b-b36e-e9482283ffd0@github.com>

- Remote tests script fixed
- Add ArrayViews in the test suite

-------------

Commit messages:
 - [hat] Remote tests script fixed + add ArrayViews

Changes: https://git.openjdk.org/babylon/pull/596/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=596&range=00
  Stats: 6 lines in 2 files changed: 0 ins; 0 del; 6 mod
  Patch: https://git.openjdk.org/babylon/pull/596.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/596/head:pull/596

PR: https://git.openjdk.org/babylon/pull/596

From jfumero at openjdk.org  Thu Oct  2 10:41:44 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Thu, 2 Oct 2025 10:41:44 GMT
Subject: git: openjdk/babylon: code-reflection: [hat] Remote tests script fixed
 + add ArrayViews
Message-ID: <7ffc19bb-a90a-4856-8654-9bb6c3f6149a@openjdk.org>

Changeset: 5a905ae8
Branch: code-reflection
Author:    Juan Fumero <jfumero at openjdk.org>
Date:      2025-10-02 10:38:38 +0000
URL:       https://git.openjdk.org/babylon/commit/5a905ae8e211ab0aff77cd31eaa0f6fe65b2cfa0

[hat] Remote tests script fixed + add ArrayViews

! hat/hat.java
! hat/scripts/remoteTesting.sh


From jfumero at openjdk.org  Thu Oct  2 10:41:37 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Thu, 2 Oct 2025 10:41:37 GMT
Subject: [code-reflection] Integrated: [hat] Remote tests script fixed + add
 ArrayViews
In-Reply-To: <u_pmFb-V-TOhkWTREoSjROaNe2BzJxy7HkLsqs7IURI=.32f9f45f-78b6-452b-b36e-e9482283ffd0@github.com>
References: <u_pmFb-V-TOhkWTREoSjROaNe2BzJxy7HkLsqs7IURI=.32f9f45f-78b6-452b-b36e-e9482283ffd0@github.com>
Message-ID: <duICBmQWeovFIGdL6daAVVnj4iP1g4Pqij2VJ9U8-w8=.c5bb7110-bba5-49d2-8188-01aebec35bed@github.com>

On Thu, 2 Oct 2025 10:32:40 GMT, Juan Fumero <jfumero at openjdk.org> wrote:

> - Remote tests script fixed
> - Add ArrayViews in the test suite

This pull request has now been integrated.

Changeset: 5a905ae8
Author:    Juan Fumero <jfumero at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/5a905ae8e211ab0aff77cd31eaa0f6fe65b2cfa0
Stats:     6 lines in 2 files changed: 0 ins; 0 del; 6 mod

[hat] Remote tests script fixed + add ArrayViews

-------------

PR: https://git.openjdk.org/babylon/pull/596

From gfrost at openjdk.org  Thu Oct  2 13:23:14 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Thu, 2 Oct 2025 13:23:14 GMT
Subject: [code-reflection] Integrated: config now accessible from all backends
Message-ID: <nsK5PdomtkjLcxXNl1CgoYPi0qHd_6b7eN5N4XuvsK8=.2c91fde0-58a7-4ac9-9ee6-fa3ecc5791c9@github.com>

Config now setup for all backends. 

This allows us to test capabilities (such as SHOW_CODE) in all backends

We should lean on this rather than Boolean.get(propName).

-------------

Commit messages:
 - config now accessible from all backends

Changes: https://git.openjdk.org/babylon/pull/597/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=597&range=00
  Stats: 290 lines in 30 files changed: 153 ins; 52 del; 85 mod
  Patch: https://git.openjdk.org/babylon/pull/597.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/597/head:pull/597

PR: https://git.openjdk.org/babylon/pull/597

From gfrost at openjdk.org  Thu Oct  2 13:23:14 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Thu, 2 Oct 2025 13:23:14 GMT
Subject: [code-reflection] Integrated: config now accessible from all backends
In-Reply-To: <nsK5PdomtkjLcxXNl1CgoYPi0qHd_6b7eN5N4XuvsK8=.2c91fde0-58a7-4ac9-9ee6-fa3ecc5791c9@github.com>
References: <nsK5PdomtkjLcxXNl1CgoYPi0qHd_6b7eN5N4XuvsK8=.2c91fde0-58a7-4ac9-9ee6-fa3ecc5791c9@github.com>
Message-ID: <heNxiusDtDevAhdVw_mY2z6ehD-WkAKKCrQCiyGjXxk=.5094a914-ab90-495f-a051-8d16ab5ff21d@github.com>

On Thu, 2 Oct 2025 13:15:53 GMT, Gary Frost <gfrost at openjdk.org> wrote:

> Config now setup for all backends. 
> 
> This allows us to test capabilities (such as SHOW_CODE) in all backends
> 
> We should lean on this rather than Boolean.get(propName).

This pull request has now been integrated.

Changeset: 469e6ab3
Author:    Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/469e6ab36eafbcc3badc999c66e59d6ef825a61d
Stats:     290 lines in 30 files changed: 153 ins; 52 del; 85 mod

config now accessible from all backends

-------------

PR: https://git.openjdk.org/babylon/pull/597

From gfrost at openjdk.org  Thu Oct  2 13:24:36 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Thu, 2 Oct 2025 13:24:36 GMT
Subject: git: openjdk/babylon: code-reflection: config now accessible from all
 backends
Message-ID: <aed8020b-d586-4afe-ba46-ebcb3c6a239d@openjdk.org>

Changeset: 469e6ab3
Branch: code-reflection
Author:    Gary Frost <gfrost at openjdk.org>
Date:      2025-10-02 13:20:05 +0000
URL:       https://git.openjdk.org/babylon/commit/469e6ab36eafbcc3badc999c66e59d6ef825a61d

config now accessible from all backends

! hat/backends/ffi/cuda/src/main/java/hat/backend/ffi/CudaBackend.java
! hat/backends/ffi/cuda/src/main/java/hat/backend/ffi/CudaDeviceInfo.java
! hat/backends/ffi/mock/src/main/java/hat/backend/ffi/MockBackend.java
! hat/backends/ffi/opencl/src/main/java/hat/backend/ffi/OpenCLBackend.java
! hat/backends/ffi/opencl/src/main/java/hat/backend/ffi/OpenCLDeviceInfo.java
! hat/backends/ffi/shared/src/main/java/hat/backend/ffi/C99FFIBackend.java
! hat/backends/ffi/shared/src/main/java/hat/backend/ffi/FFIBackend.java
! hat/backends/ffi/shared/src/main/java/hat/backend/ffi/FFIBackendDriver.java
! hat/backends/ffi/shared/src/main/native/include/config.h
! hat/backends/java/seq/src/main/java/hat/backend/java/JavaSequentialBackend.java
! hat/backends/jextracted/opencl/src/main/java/hat/backend/jextracted/OpenCLBackend.java
! hat/backends/jextracted/shared/src/main/java/hat/backend/jextracted/C99JExtractedBackend.java
! hat/backends/jextracted/shared/src/main/java/hat/backend/jextracted/JExtractedBackend.java
! hat/backends/jextracted/shared/src/main/java/hat/backend/jextracted/JExtractedBackendDriver.java
! hat/core/src/main/java/hat/Accelerator.java
! hat/core/src/main/java/hat/ComputeContext.java
! hat/core/src/main/java/hat/Config.java
! hat/core/src/main/java/hat/FFIConfigCreator.java
! hat/core/src/main/java/hat/backend/Backend.java
! hat/core/src/main/java/hat/backend/BackendAdaptor.java
! hat/core/src/main/java/hat/backend/DebugBackend.java
! hat/core/src/main/java/hat/backend/java/JavaBackend.java
! hat/core/src/main/java/hat/buffer/ArgArray.java
! hat/core/src/main/java/hat/callgraph/CallGraph.java
! hat/core/src/main/java/hat/callgraph/ComputeCallGraph.java
! hat/core/src/main/java/hat/codebuilders/C99HATConfigBuilder.java
! hat/examples/experiments/src/main/java/experiments/Mesh.java
! hat/examples/experiments/src/main/java/experiments/MinBufferTest.java
! hat/examples/mandel/src/main/java/mandel/Main.java
! hat/examples/violajones/src/main/java/violajones/Main.java


From gfrost at openjdk.org  Thu Oct  2 13:32:07 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Thu, 2 Oct 2025 13:32:07 GMT
Subject: [code-reflection] Integrated: hat.java now scans the tests.jar for
 tests to run
Message-ID: <jGcTz1Ojkct8jAbI41f9mfi-Wkjww3KIgr6WXRl0Fks=.38fb26ef-815c-4216-9e52-67ad3842581c@github.com>

Previously we ha dto add each new test suite to `hat.java` now we scan the jar file for test suites.

-------------

Commit messages:
 - hat.java now scans the tests.jar for tests to run
 - hat.java now scans the tests.jar for tests to run

Changes: https://git.openjdk.org/babylon/pull/598/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=598&range=00
  Stats: 10 lines in 1 file changed: 0 ins; 0 del; 10 mod
  Patch: https://git.openjdk.org/babylon/pull/598.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/598/head:pull/598

PR: https://git.openjdk.org/babylon/pull/598

From gfrost at openjdk.org  Thu Oct  2 13:32:07 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Thu, 2 Oct 2025 13:32:07 GMT
Subject: [code-reflection] Integrated: hat.java now scans the tests.jar for
 tests to run
In-Reply-To: <jGcTz1Ojkct8jAbI41f9mfi-Wkjww3KIgr6WXRl0Fks=.38fb26ef-815c-4216-9e52-67ad3842581c@github.com>
References: <jGcTz1Ojkct8jAbI41f9mfi-Wkjww3KIgr6WXRl0Fks=.38fb26ef-815c-4216-9e52-67ad3842581c@github.com>
Message-ID: <dsR2Cha8ZIOcOaGPzCUFVvA8H4AX1VeswFR21fytGCc=.0147d786-0dc8-4782-ade3-cf57000a0748@github.com>

On Thu, 2 Oct 2025 13:23:46 GMT, Gary Frost <gfrost at openjdk.org> wrote:

> Previously we ha dto add each new test suite to `hat.java` now we scan the jar file for test suites.

This pull request has now been integrated.

Changeset: 4fbdcd4a
Author:    Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/4fbdcd4a421cf23dfaf33ee3d9a799d5abc6ccaa
Stats:     10 lines in 1 file changed: 0 ins; 0 del; 10 mod

hat.java now scans the tests.jar for tests to run

-------------

PR: https://git.openjdk.org/babylon/pull/598

From gfrost at openjdk.org  Thu Oct  2 13:32:59 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Thu, 2 Oct 2025 13:32:59 GMT
Subject: git: openjdk/babylon: code-reflection: hat.java now scans the
 tests.jar for tests to run
Message-ID: <7e9b1775-c17a-4381-91a1-6698c62eb5cb@openjdk.org>

Changeset: 4fbdcd4a
Branch: code-reflection
Author:    Gary Frost <gfrost at openjdk.org>
Date:      2025-10-02 13:29:03 +0000
URL:       https://git.openjdk.org/babylon/commit/4fbdcd4a421cf23dfaf33ee3d9a799d5abc6ccaa

hat.java now scans the tests.jar for tests to run

! hat/hat.java


From duke at openjdk.org  Thu Oct  2 23:12:47 2025
From: duke at openjdk.org (Ruby Chen)
Date: Thu, 2 Oct 2025 23:12:47 GMT
Subject: [code-reflection] RFR: Remove noModuleOp and bufferTagging flags
Message-ID: <INshjDUX6UYGezCE6_NIHloBB-aSBSdfd0H2BNwiWMY=.04153fa9-b6ce-4753-bed4-7a04d1c2aca2@github.com>

Phase out the backend that doesn't use `ModuleOp`. Also make `bufferTagging` run by default.

-------------

Commit messages:
 - Remove noModuleOp and bufferTagging flags

Changes: https://git.openjdk.org/babylon/pull/599/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=599&range=00
  Stats: 112 lines in 12 files changed: 15 ins; 77 del; 20 mod
  Patch: https://git.openjdk.org/babylon/pull/599.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/599/head:pull/599

PR: https://git.openjdk.org/babylon/pull/599

From duke at openjdk.org  Fri Oct  3 14:52:06 2025
From: duke at openjdk.org (duke)
Date: Fri, 3 Oct 2025 14:52:06 GMT
Subject: [code-reflection] RFR: Remove noModuleOp and bufferTagging flags
In-Reply-To: <INshjDUX6UYGezCE6_NIHloBB-aSBSdfd0H2BNwiWMY=.04153fa9-b6ce-4753-bed4-7a04d1c2aca2@github.com>
References: <INshjDUX6UYGezCE6_NIHloBB-aSBSdfd0H2BNwiWMY=.04153fa9-b6ce-4753-bed4-7a04d1c2aca2@github.com>
Message-ID: <4rWTe4AnWQ0pbxoRMdP-bnN4LiSNXQSUbHKCgeA7l1M=.91249acc-7ddb-4080-b2bb-a1d3b806fcad@github.com>

On Thu, 2 Oct 2025 23:06:07 GMT, Ruby Chen <duke at openjdk.org> wrote:

> Phase out the backend that doesn't use `ModuleOp`. Also make `bufferTagging` run by default.

@rbrchen 
Your change (at version 56b48cf35bb0b559eabe55b072772f107ebc66cd) is now ready to be sponsored by a Committer.

-------------

PR Comment: https://git.openjdk.org/babylon/pull/599#issuecomment-3366013209

From gfrost at openjdk.org  Fri Oct  3 15:03:37 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Fri, 3 Oct 2025 15:03:37 GMT
Subject: git: openjdk/babylon: code-reflection: Remove noModuleOp and
 bufferTagging flags
Message-ID: <7941606c-8edf-411f-8a46-a4978ae0c01e@openjdk.org>

Changeset: 09e34f28
Branch: code-reflection
Author:    Ruby Chen <ruby.r.chen at gmail.com>
Committer: Gary Frost <gfrost at openjdk.org>
Date:      2025-10-03 15:01:47 +0000
URL:       https://git.openjdk.org/babylon/commit/09e34f28441ed95fcc218a0c8b65e3d8e0c9190d

Remove noModuleOp and bufferTagging flags

! hat/backends/ffi/cuda/src/main/java/hat/backend/ffi/CudaBackend.java
! hat/backends/ffi/mock/src/main/java/hat/backend/ffi/MockBackend.java
! hat/backends/ffi/shared/src/main/java/hat/backend/ffi/C99FFIBackend.java
! hat/core/src/main/java/hat/BufferTagger.java
! hat/core/src/main/java/hat/buffer/ArgArray.java
! hat/core/src/main/java/hat/callgraph/CallGraph.java
! hat/core/src/main/java/hat/callgraph/ComputeCallGraph.java
! hat/core/src/main/java/hat/callgraph/KernelCallGraph.java
! hat/hat.java
! hat/hat/Script.java
! hat/hat/run.java
! hat/hat/test.java


From duke at openjdk.org  Fri Oct  3 15:04:14 2025
From: duke at openjdk.org (Ruby Chen)
Date: Fri, 3 Oct 2025 15:04:14 GMT
Subject: [code-reflection] Integrated: Remove noModuleOp and bufferTagging
 flags
In-Reply-To: <INshjDUX6UYGezCE6_NIHloBB-aSBSdfd0H2BNwiWMY=.04153fa9-b6ce-4753-bed4-7a04d1c2aca2@github.com>
References: <INshjDUX6UYGezCE6_NIHloBB-aSBSdfd0H2BNwiWMY=.04153fa9-b6ce-4753-bed4-7a04d1c2aca2@github.com>
Message-ID: <2CQM8KIJPGxvo2KSFjVQ7OBpeDFX6VJqqPoAy5xgpYw=.ab9bd535-4d84-4022-9570-0e415b7ea3f1@github.com>

On Thu, 2 Oct 2025 23:06:07 GMT, Ruby Chen <duke at openjdk.org> wrote:

> Phase out the backend that doesn't use `ModuleOp`. Also make `bufferTagging` run by default.

This pull request has now been integrated.

Changeset: 09e34f28
Author:    Ruby Chen <ruby.r.chen at gmail.com>
Committer: Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/09e34f28441ed95fcc218a0c8b65e3d8e0c9190d
Stats:     112 lines in 12 files changed: 15 ins; 77 del; 20 mod

Remove noModuleOp and bufferTagging flags

-------------

PR: https://git.openjdk.org/babylon/pull/599

From gfrost at openjdk.org  Fri Oct  3 15:29:01 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Fri, 3 Oct 2025 15:29:01 GMT
Subject: [code-reflection] Integrated: implemented preprocessor hook for OpenCL
 code gen
Message-ID: <NyteK_1sNbpAxS4kNzMJlzFTtM0JLHkMlmq0ObBlR6w=.df140bfa-61d1-48aa-a642-14cdc5ecfade@github.com>

Here we leverage OpenCL's ability to lean on the C99 preprocessor to generate C99 code. 

Instead of having OpenCL generator differentiate, we get the OpenCL code builder to define a set of #defines and generate backend 'neutral' C99 code (common between Cuda and OpenCL). 

This PR just has the OpenCL changes.  Cuda changes will be next

-------------

Commit messages:
 - implemented preprocessor hook for OpenCL code gen

Changes: https://git.openjdk.org/babylon/pull/600/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=600&range=00
  Stats: 250 lines in 10 files changed: 90 ins; 44 del; 116 mod
  Patch: https://git.openjdk.org/babylon/pull/600.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/600/head:pull/600

PR: https://git.openjdk.org/babylon/pull/600

From gfrost at openjdk.org  Fri Oct  3 15:29:02 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Fri, 3 Oct 2025 15:29:02 GMT
Subject: [code-reflection] Integrated: implemented preprocessor hook for OpenCL
 code gen
In-Reply-To: <NyteK_1sNbpAxS4kNzMJlzFTtM0JLHkMlmq0ObBlR6w=.df140bfa-61d1-48aa-a642-14cdc5ecfade@github.com>
References: <NyteK_1sNbpAxS4kNzMJlzFTtM0JLHkMlmq0ObBlR6w=.df140bfa-61d1-48aa-a642-14cdc5ecfade@github.com>
Message-ID: <kYVaH0Pagyrgs7R1N66XHrAN9yY5gKN0rliMOpngwXI=.9470a15d-0c87-45c5-bb63-3519e62b39d4@github.com>

On Fri, 3 Oct 2025 15:22:05 GMT, Gary Frost <gfrost at openjdk.org> wrote:

> Here we leverage OpenCL's ability to lean on the C99 preprocessor to generate C99 code. 
> 
> Instead of having OpenCL generator differentiate, we get the OpenCL code builder to define a set of #defines and generate backend 'neutral' C99 code (common between Cuda and OpenCL). 
> 
> This PR just has the OpenCL changes.  Cuda changes will be next

This pull request has now been integrated.

Changeset: d1bd5d57
Author:    Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/d1bd5d57d38573c536ec5c4d48b248c0e4b4a1cb
Stats:     250 lines in 10 files changed: 90 ins; 44 del; 116 mod

implemented preprocessor hook for OpenCL code gen

-------------

PR: https://git.openjdk.org/babylon/pull/600

From gfrost at openjdk.org  Fri Oct  3 15:29:11 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Fri, 3 Oct 2025 15:29:11 GMT
Subject: git: openjdk/babylon: code-reflection: implemented preprocessor hook
 for OpenCL code gen
Message-ID: <1fbd1449-f0dd-45a6-b596-74f9a95fcae1@openjdk.org>

Changeset: d1bd5d57
Branch: code-reflection
Author:    Gary Frost <gfrost at openjdk.org>
Date:      2025-10-03 15:25:23 +0000
URL:       https://git.openjdk.org/babylon/commit/d1bd5d57d38573c536ec5c4d48b248c0e4b4a1cb

implemented preprocessor hook for OpenCL code gen

! hat/backends/ffi/cuda/src/main/java/hat/backend/ffi/CudaHATKernelBuilder.java
! hat/backends/ffi/opencl/src/main/java/hat/backend/ffi/OpenCLHATKernelBuilder.java
! hat/backends/ffi/shared/src/main/java/hat/backend/ffi/C99FFIBackend.java
! hat/backends/jextracted/opencl/src/main/java/hat/backend/jextracted/OpenCLHatKernelBuilder.java
! hat/backends/jextracted/shared/src/main/java/hat/backend/jextracted/C99JExtractedBackend.java
! hat/core/src/main/java/hat/codebuilders/C99HATKernelBuilder.java
! hat/core/src/main/java/hat/codebuilders/CodeBuilder.java
! hat/core/src/main/java/hat/codebuilders/HATCodeBuilder.java
! hat/examples/experiments/src/main/java/experiments/PrefixSum.java
! hat/examples/nbody/src/main/java/nbody/opencl/OpenCLNBodyGLWindow.java


From gfrost at openjdk.org  Fri Oct  3 17:03:17 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Fri, 3 Oct 2025 17:03:17 GMT
Subject: git: openjdk/babylon: code-reflection: cuda now also leverages c99
 preprocessor
Message-ID: <ec843faa-9694-4a0f-8728-5dd86d38c706@openjdk.org>

Changeset: 86a9485a
Branch: code-reflection
Author:    Gary Frost <gfrost at openjdk.org>
Date:      2025-10-03 17:01:32 +0000
URL:       https://git.openjdk.org/babylon/commit/86a9485a0428af125533e7d2a01ef683d1598c28

cuda now also leverages c99 preprocessor

! hat/backends/ffi/cuda/src/main/java/hat/backend/ffi/CudaHATKernelBuilder.java
! hat/backends/ffi/opencl/src/main/java/hat/backend/ffi/OpenCLHATKernelBuilder.java
! hat/core/src/main/java/hat/codebuilders/C99HATKernelBuilder.java
+ hat/hat/exp


From gfrost at openjdk.org  Fri Oct  3 17:04:53 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Fri, 3 Oct 2025 17:04:53 GMT
Subject: [code-reflection] Integrated: cuda now also leverages c99 preprocessor
Message-ID: <gEKAIyF3O2-h7g_Ifne4Kw6EyqXllPysHrg1ZDr-wJ8=.996fb39d-0868-4e54-8673-4cafd8aa731e@github.com>

Merged common HATKernelCodeGen.  Now both CUDA and OpenCL lean on the c99 preprocessor.

-------------

Commit messages:
 - cuda now also leverages c99 preprocessor

Changes: https://git.openjdk.org/babylon/pull/601/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=601&range=00
  Stats: 203 lines in 4 files changed: 75 ins; 99 del; 29 mod
  Patch: https://git.openjdk.org/babylon/pull/601.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/601/head:pull/601

PR: https://git.openjdk.org/babylon/pull/601

From gfrost at openjdk.org  Fri Oct  3 17:04:53 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Fri, 3 Oct 2025 17:04:53 GMT
Subject: [code-reflection] Integrated: cuda now also leverages c99 preprocessor
In-Reply-To: <gEKAIyF3O2-h7g_Ifne4Kw6EyqXllPysHrg1ZDr-wJ8=.996fb39d-0868-4e54-8673-4cafd8aa731e@github.com>
References: <gEKAIyF3O2-h7g_Ifne4Kw6EyqXllPysHrg1ZDr-wJ8=.996fb39d-0868-4e54-8673-4cafd8aa731e@github.com>
Message-ID: <5hTVIZvyBeQtjmGEopReTVIPdTH4_6VXY3FOumLLIwA=.50b271a9-ded4-4854-b250-d3f1f9651b6f@github.com>

On Fri, 3 Oct 2025 16:56:58 GMT, Gary Frost <gfrost at openjdk.org> wrote:

> Merged common HATKernelCodeGen.  Now both CUDA and OpenCL lean on the c99 preprocessor.

This pull request has now been integrated.

Changeset: 86a9485a
Author:    Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/86a9485a0428af125533e7d2a01ef683d1598c28
Stats:     203 lines in 4 files changed: 75 ins; 99 del; 29 mod

cuda now also leverages c99 preprocessor

-------------

PR: https://git.openjdk.org/babylon/pull/601

From gfrost at openjdk.org  Sat Oct  4 10:26:32 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Sat, 4 Oct 2025 10:26:32 GMT
Subject: git: openjdk/babylon: code-reflection: converging cuda/opencl code gen
 using preprocessor
Message-ID: <6533cdf7-3d36-4601-af00-8ed7a4815d34@openjdk.org>

Changeset: 09d27d9b
Branch: code-reflection
Author:    Gary Frost <gfrost at openjdk.org>
Date:      2025-10-04 10:24:15 +0000
URL:       https://git.openjdk.org/babylon/commit/09d27d9b887a7d573855924abd12b9a558e9e201

converging cuda/opencl code gen using preprocessor

! hat/backends/ffi/cuda/src/main/java/hat/backend/ffi/CudaHATKernelBuilder.java
! hat/backends/ffi/opencl/src/main/java/hat/backend/ffi/OpenCLHATKernelBuilder.java
! hat/core/src/main/java/hat/codebuilders/C99HATKernelBuilder.java


From gfrost at openjdk.org  Sat Oct  4 10:27:46 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Sat, 4 Oct 2025 10:27:46 GMT
Subject: [code-reflection] Integrated: converging cuda/opencl code gen using
 preprocessor
Message-ID: <0o51AeV6_u0xdq69mH2MDBNpzZDqbkSlze_xpVDgZbg=.05462cfe-7167-4a82-868e-a4eb66073261@github.com>

Took opportunity to move more c99 code gen to preprocessor.

-------------

Commit messages:
 - converging cuda/opencl code gen using preprocessor

Changes: https://git.openjdk.org/babylon/pull/602/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=602&range=00
  Stats: 79 lines in 3 files changed: 20 ins; 25 del; 34 mod
  Patch: https://git.openjdk.org/babylon/pull/602.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/602/head:pull/602

PR: https://git.openjdk.org/babylon/pull/602

From gfrost at openjdk.org  Sat Oct  4 10:27:46 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Sat, 4 Oct 2025 10:27:46 GMT
Subject: [code-reflection] Integrated: converging cuda/opencl code gen using
 preprocessor
In-Reply-To: <0o51AeV6_u0xdq69mH2MDBNpzZDqbkSlze_xpVDgZbg=.05462cfe-7167-4a82-868e-a4eb66073261@github.com>
References: <0o51AeV6_u0xdq69mH2MDBNpzZDqbkSlze_xpVDgZbg=.05462cfe-7167-4a82-868e-a4eb66073261@github.com>
Message-ID: <W7IaOhzJc7I3o0A4yKuO7oSrhfT2c-s5t4vGB_z3fdU=.1871261f-d1fc-40ed-95a0-b605c3cfd2fd@github.com>

On Sat, 4 Oct 2025 10:22:02 GMT, Gary Frost <gfrost at openjdk.org> wrote:

> Took opportunity to move more c99 code gen to preprocessor.

This pull request has now been integrated.

Changeset: 09d27d9b
Author:    Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/09d27d9b887a7d573855924abd12b9a558e9e201
Stats:     79 lines in 3 files changed: 20 ins; 25 del; 34 mod

converging cuda/opencl code gen using preprocessor

-------------

PR: https://git.openjdk.org/babylon/pull/602

From gfrost at openjdk.org  Sat Oct  4 11:06:31 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Sat, 4 Oct 2025 11:06:31 GMT
Subject: [code-reflection] Integrated: CUDA and OpenCL C99 code now identical
 except for handling of atomics
Message-ID: <CeUYAezxTvw4tqRtefifgs8u7zAsy-yEes77tamEOzk=.f0c98842-e870-403f-8967-1f08fc510373@github.com>

Generated code essentially same for CUDA and OpenCL.  

Not only does this lend itself to simpler codegen, it also opens the doorway to a HAT C99 dialect which we can use for prototyping and even intrinsics

Check the code generated for 

HAT=SHOW_CODE java @hat/test ffi-opencl
HAT=SHOW_CODE java @hat/test ffi-cuda
``` 
The only differences should be the generated preprocessor macros.

We need to deal with atomics .... somehow. Still thinking about that.

-------------

Commit messages:
 - CUDA and OpenCL C99 code now identical except for handling of atomics

Changes: https://git.openjdk.org/babylon/pull/603/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=603&range=00
  Stats: 90 lines in 3 files changed: 0 ins; 28 del; 62 mod
  Patch: https://git.openjdk.org/babylon/pull/603.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/603/head:pull/603

PR: https://git.openjdk.org/babylon/pull/603

From gfrost at openjdk.org  Sat Oct  4 11:06:31 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Sat, 4 Oct 2025 11:06:31 GMT
Subject: [code-reflection] Integrated: CUDA and OpenCL C99 code now identical
 except for handling of atomics
In-Reply-To: <CeUYAezxTvw4tqRtefifgs8u7zAsy-yEes77tamEOzk=.f0c98842-e870-403f-8967-1f08fc510373@github.com>
References: <CeUYAezxTvw4tqRtefifgs8u7zAsy-yEes77tamEOzk=.f0c98842-e870-403f-8967-1f08fc510373@github.com>
Message-ID: <dLgLV_uLlYyowbF-a2sF1fJvn2gA9xEc932689cv6v8=.96734aa0-aa53-452c-a6a6-6ec8a8958b97@github.com>

On Sat, 4 Oct 2025 11:00:19 GMT, Gary Frost <gfrost at openjdk.org> wrote:

> Generated code essentially same for CUDA and OpenCL.  
> 
> Not only does this lend itself to simpler codegen, it also opens the doorway to a HAT C99 dialect which we can use for prototyping and even intrinsics
> 
> Check the code generated for 
> 
> HAT=SHOW_CODE java @hat/test ffi-opencl
> HAT=SHOW_CODE java @hat/test ffi-cuda
> ``` 
> The only differences should be the generated preprocessor macros.
> 
> We need to deal with atomics .... somehow. Still thinking about that.

This pull request has now been integrated.

Changeset: b46f103b
Author:    Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/b46f103b7cafa6216ee82fbc2c557a10883b1728
Stats:     90 lines in 3 files changed: 0 ins; 28 del; 62 mod

CUDA and OpenCL C99 code now identical except for handling of atomics

-------------

PR: https://git.openjdk.org/babylon/pull/603

From gfrost at openjdk.org  Sat Oct  4 11:06:48 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Sat, 4 Oct 2025 11:06:48 GMT
Subject: git: openjdk/babylon: code-reflection: CUDA and OpenCL C99 code now
 identical except for handling of atomics
Message-ID: <e023ddf3-fa8d-4f74-aa40-e5b27e8868e5@openjdk.org>

Changeset: b46f103b
Branch: code-reflection
Author:    Gary Frost <gfrost at openjdk.org>
Date:      2025-10-04 11:03:49 +0000
URL:       https://git.openjdk.org/babylon/commit/b46f103b7cafa6216ee82fbc2c557a10883b1728

CUDA and OpenCL C99 code now identical except for handling of atomics

! hat/backends/ffi/cuda/src/main/java/hat/backend/ffi/CudaHATKernelBuilder.java
! hat/backends/ffi/opencl/src/main/java/hat/backend/ffi/OpenCLHATKernelBuilder.java
! hat/core/src/main/java/hat/codebuilders/C99HATKernelBuilder.java


From jfumero at openjdk.org  Mon Oct  6 10:43:51 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Mon, 6 Oct 2025 10:43:51 GMT
Subject: [code-reflection] Integrated: [hat][docs] Some useful links added
Message-ID: <rUYIR3qDjZY-FHo9aC63bbXD06_OYKLsC6rl9zRsfdE=.6b6272e8-80c8-476b-aa7a-d2d87385671e@github.com>

Some useful links added

-------------

Commit messages:
 - [hat][docs] Some useful links added

Changes: https://git.openjdk.org/babylon/pull/604/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=604&range=00
  Stats: 4 lines in 1 file changed: 4 ins; 0 del; 0 mod
  Patch: https://git.openjdk.org/babylon/pull/604.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/604/head:pull/604

PR: https://git.openjdk.org/babylon/pull/604

From jfumero at openjdk.org  Mon Oct  6 10:43:52 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Mon, 6 Oct 2025 10:43:52 GMT
Subject: [code-reflection] Integrated: [hat][docs] Some useful links added
In-Reply-To: <rUYIR3qDjZY-FHo9aC63bbXD06_OYKLsC6rl9zRsfdE=.6b6272e8-80c8-476b-aa7a-d2d87385671e@github.com>
References: <rUYIR3qDjZY-FHo9aC63bbXD06_OYKLsC6rl9zRsfdE=.6b6272e8-80c8-476b-aa7a-d2d87385671e@github.com>
Message-ID: <lUBcEan8D0OL4lht_4gkpqPc09lC4jMGMZbKuFgbP3U=.4514e6c1-8a7c-4879-9ab5-82930ac7b49a@github.com>

On Mon, 6 Oct 2025 10:35:29 GMT, Juan Fumero <jfumero at openjdk.org> wrote:

> Some useful links added

This pull request has now been integrated.

Changeset: cd9577b0
Author:    Juan Fumero <jfumero at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/cd9577b0fb12a17a3a672e58225a89c0bb8f23b9
Stats:     4 lines in 1 file changed: 4 ins; 0 del; 0 mod

[hat][docs] Some useful links added

-------------

PR: https://git.openjdk.org/babylon/pull/604

From jfumero at openjdk.org  Mon Oct  6 10:44:40 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Mon, 6 Oct 2025 10:44:40 GMT
Subject: git: openjdk/babylon: code-reflection: [hat][docs] Some useful links
 added
Message-ID: <5e4a160a-aecb-4420-94e7-3ef6f7137e08@openjdk.org>

Changeset: cd9577b0
Branch: code-reflection
Author:    Juan Fumero <jfumero at openjdk.org>
Date:      2025-10-06 10:40:50 +0000
URL:       https://git.openjdk.org/babylon/commit/cd9577b0fb12a17a3a672e58225a89c0bb8f23b9

[hat][docs] Some useful links added

! hat/docs/hat-notes-and-links.md


From duke at openjdk.org  Tue Oct  7 21:53:17 2025
From: duke at openjdk.org (Ruby Chen)
Date: Tue, 7 Oct 2025 21:53:17 GMT
Subject: [code-reflection] RFR: Add kernel entrypoint to ModuleOp
Message-ID: <UkyuDt6Ei-D4V2Ly2ukTEV-Sx9qLIbN6ih3KSJ1BHYY=.45a19d55-ee90-46ec-b41c-28ec98117c5b@github.com>

Include the kernel (and compute) entrypoints in their respective ModuleOps.

-------------

Commit messages:
 - Add kernel entrypoint to ModuleOp

Changes: https://git.openjdk.org/babylon/pull/605/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=605&range=00
  Stats: 43 lines in 2 files changed: 9 ins; 22 del; 12 mod
  Patch: https://git.openjdk.org/babylon/pull/605.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/605/head:pull/605

PR: https://git.openjdk.org/babylon/pull/605

From duke at openjdk.org  Fri Oct 10 18:19:12 2025
From: duke at openjdk.org (Ruby Chen)
Date: Fri, 10 Oct 2025 18:19:12 GMT
Subject: [code-reflection] RFR: Add private and local arrayView support
Message-ID: <-E-v636NObf8LpGdZj4eYCphUYCSvkh7rgeShc4ORSA=.584358e1-13c4-4d7b-8de8-8876c7e32e34@github.com>

Private and local arrayViews can now also be accessed using `.privateArrayView()` and `.localArrayView()`, respectively.

Implementation of `.privateArrayView()` and `.localArrayView()` is left to the user, and accessing global arrays is still kept as `.arrayView()`.

-------------

Commit messages:
 - add private and local arrayview support

Changes: https://git.openjdk.org/babylon/pull/607/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=607&range=00
  Stats: 188 lines in 4 files changed: 123 ins; 25 del; 40 mod
  Patch: https://git.openjdk.org/babylon/pull/607.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/607/head:pull/607

PR: https://git.openjdk.org/babylon/pull/607

From duke at openjdk.org  Tue Oct 14 01:10:18 2025
From: duke at openjdk.org (Ruby Chen)
Date: Tue, 14 Oct 2025 01:10:18 GMT
Subject: [code-reflection] RFR: Add private and local arrayView support
 [v2]
In-Reply-To: <-E-v636NObf8LpGdZj4eYCphUYCSvkh7rgeShc4ORSA=.584358e1-13c4-4d7b-8de8-8876c7e32e34@github.com>
References: <-E-v636NObf8LpGdZj4eYCphUYCSvkh7rgeShc4ORSA=.584358e1-13c4-4d7b-8de8-8876c7e32e34@github.com>
Message-ID: <zFHU6ThrrBR6Qo9ilBTNVw-YXdXYXRRSIbGqF8XMUD4=.8f01ef7c-80a6-4ba7-9391-dc0b304f90d9@github.com>

> Private and local arrayViews can now also be accessed using `.privateArrayView()` and `.localArrayView()`, respectively.
> 
> Implementation of `.privateArrayView()` and `.localArrayView()` is left to the user, and accessing global arrays is still kept as `.arrayView()`.

Ruby Chen has updated the pull request incrementally with one additional commit since the last revision:

  cleaning up arrayView code

-------------

Changes:
  - all: https://git.openjdk.org/babylon/pull/607/files
  - new: https://git.openjdk.org/babylon/pull/607/files/9f8c9ec8..1411d7ed

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=babylon&pr=607&range=01
 - incr: https://webrevs.openjdk.org/?repo=babylon&pr=607&range=00-01

  Stats: 28 lines in 1 file changed: 0 ins; 10 del; 18 mod
  Patch: https://git.openjdk.org/babylon/pull/607.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/607/head:pull/607

PR: https://git.openjdk.org/babylon/pull/607

From duke at openjdk.org  Tue Oct 14 01:10:19 2025
From: duke at openjdk.org (duke)
Date: Tue, 14 Oct 2025 01:10:19 GMT
Subject: [code-reflection] RFR: Add private and local arrayView support
In-Reply-To: <-E-v636NObf8LpGdZj4eYCphUYCSvkh7rgeShc4ORSA=.584358e1-13c4-4d7b-8de8-8876c7e32e34@github.com>
References: <-E-v636NObf8LpGdZj4eYCphUYCSvkh7rgeShc4ORSA=.584358e1-13c4-4d7b-8de8-8876c7e32e34@github.com>
Message-ID: <RnyvozSV_sf78039jJif4xr4uQUsb2d9D9riGpYjNPM=.dc24708b-dcb1-4b10-940e-5e70778d5767@github.com>

On Fri, 10 Oct 2025 18:13:23 GMT, Ruby Chen <duke at openjdk.org> wrote:

> Private and local arrayViews can now also be accessed using `.privateArrayView()` and `.localArrayView()`, respectively.
> 
> Implementation of `.privateArrayView()` and `.localArrayView()` is left to the user, and accessing global arrays is still kept as `.arrayView()`.

@rbrchen 
Your change (at version 1411d7ed716943c84d56d965649d35381b360005) is now ready to be sponsored by a Committer.

-------------

PR Comment: https://git.openjdk.org/babylon/pull/607#issuecomment-3399586999

From gfrost at openjdk.org  Tue Oct 14 07:46:35 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Tue, 14 Oct 2025 07:46:35 GMT
Subject: git: openjdk/babylon: code-reflection: Add private and local arrayView
 support
Message-ID: <8b4b92f9-5093-4397-9051-5b223cbe703e@openjdk.org>

Changeset: e5dfd23b
Branch: code-reflection
Author:    Ruby Chen <ruby.r.chen at gmail.com>
Committer: Gary Frost <gfrost at openjdk.org>
Date:      2025-10-14 07:45:38 +0000
URL:       https://git.openjdk.org/babylon/commit/e5dfd23bac5e313870ca3dabcc02321d74ab59e5

Add private and local arrayView support

! hat/core/src/main/java/hat/callgraph/KernelCallGraph.java
! hat/core/src/main/java/hat/optools/OpTk.java
! hat/examples/arrayview/src/main/java/arrayview/Main.java
! hat/tests/src/main/java/oracle/code/hat/TestArrayView.java


From duke at openjdk.org  Tue Oct 14 07:48:39 2025
From: duke at openjdk.org (Ruby Chen)
Date: Tue, 14 Oct 2025 07:48:39 GMT
Subject: [code-reflection] Integrated: Add private and local arrayView support
In-Reply-To: <-E-v636NObf8LpGdZj4eYCphUYCSvkh7rgeShc4ORSA=.584358e1-13c4-4d7b-8de8-8876c7e32e34@github.com>
References: <-E-v636NObf8LpGdZj4eYCphUYCSvkh7rgeShc4ORSA=.584358e1-13c4-4d7b-8de8-8876c7e32e34@github.com>
Message-ID: <dWUDVcdFsVJh9AIfHKC2_FbDx6Fbi39NsnHKq5Q-MZ8=.a62697db-4c5b-4e8c-b0ab-e8ec0630598b@github.com>

On Fri, 10 Oct 2025 18:13:23 GMT, Ruby Chen <duke at openjdk.org> wrote:

> Private and local arrayViews can now also be accessed using `.privateArrayView()` and `.localArrayView()`, respectively.
> 
> Implementation of `.privateArrayView()` and `.localArrayView()` is left to the user, and accessing global arrays is still kept as `.arrayView()`.

This pull request has now been integrated.

Changeset: e5dfd23b
Author:    Ruby Chen <ruby.r.chen at gmail.com>
Committer: Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/e5dfd23bac5e313870ca3dabcc02321d74ab59e5
Stats:     199 lines in 4 files changed: 120 ins; 32 del; 47 mod

Add private and local arrayView support

-------------

PR: https://git.openjdk.org/babylon/pull/607

From jfumero at openjdk.org  Tue Oct 14 09:41:14 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Tue, 14 Oct 2025 09:41:14 GMT
Subject: [code-reflection] RFR: Proposal for Vector Types for GPUs (OpenCL and
 CUDA backends)
Message-ID: <laZ7WwpiImcpodRE2bHwYkw3HunfrRMVvNTTsoHKxmI=.a13dbc47-6d88-412a-b57a-ea2f9d92d783@github.com>

This PR opens a new proposal to integrate an API for representing explicit vector types in HAT. 

The way to operate is as follows:


Float4 vA = a.float4View(index * 4);
Float4 vB = b.float4View(index * 4);
Float4 vC = Float4.add(vA, vB);
c.storeFloat4View(vC, index * 4);


Java programmers can load/store float4 chunks within one instruction by using array views. These views operate on GPU's global memory, but we can add similar concepts for local/shared and private memory.  

Additionally, Java/HAT developers can access lanes within the vector instance by name:


vA.x(myNewValue);
float foo = vA.z();
``` 

We will iterate over this proposal. 

How to test?

For CUDA

HAT=SHOW_CODE,INFO,SHOW_COMPILATION_PHASES java @hat/otest ffi-cuda oracle.code.hat.TestVectorTypes


For OpenCL:



HAT=SHOW_CODE,INFO,SHOW_COMPILATION_PHASES java @hat/otest ffi-opencl oracle.code.hat.TestVectorTypes


This PR also introduces a new option to debug the HAT Dialect Phases. 


HAT=SHOW_COMPILATION_PHASES ...

-------------

Commit messages:
 - [hat] clean-up old code
 - [hat] float4 support for the CUDA backend types generalized
 - [hat] Padding for copyFrom/copyTo functions
 - Merge branch 'code-reflection' into hat/vector/types
 - [hat] Example of matmul with 2D RT + vector loads
 - [hat] Matmul with 2D register tiling + vector loads/stores working
 - [hat] clean-up test
 - [hat] CUDA backend with the right accessor for loading/storing float4 views
 - [hat] Allow float4 to/from local/private memory
 - [hat] Offsets for S32Arrays fixed
 - ... and 14 more: https://git.openjdk.org/babylon/compare/e5dfd23b...5f37c6f7

Changes: https://git.openjdk.org/babylon/pull/606/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=606&range=00
  Stats: 3350 lines in 44 files changed: 3219 ins; 86 del; 45 mod
  Patch: https://git.openjdk.org/babylon/pull/606.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/606/head:pull/606

PR: https://git.openjdk.org/babylon/pull/606

From jfumero at openjdk.org  Tue Oct 14 11:26:25 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Tue, 14 Oct 2025 11:26:25 GMT
Subject: [code-reflection] RFR: Proposal for Vector Types for GPUs (OpenCL
 and CUDA backends) [v2]
In-Reply-To: <laZ7WwpiImcpodRE2bHwYkw3HunfrRMVvNTTsoHKxmI=.a13dbc47-6d88-412a-b57a-ea2f9d92d783@github.com>
References: <laZ7WwpiImcpodRE2bHwYkw3HunfrRMVvNTTsoHKxmI=.a13dbc47-6d88-412a-b57a-ea2f9d92d783@github.com>
Message-ID: <ORPE-lBsa7AqSKjwbWnTXUlK8hGGpo6xS5BF1rkHaYw=.c8ceae40-2e9a-44fa-a0b6-45b1b1560209@github.com>

> This PR opens a new proposal to integrate an API for representing explicit vector types in HAT. 
> 
> The way to operate is as follows:
> 
> 
> Float4 vA = a.float4View(index * 4);
> Float4 vB = b.float4View(index * 4);
> Float4 vC = Float4.add(vA, vB);
> c.storeFloat4View(vC, index * 4);
> 
> 
> Java programmers can load/store float4 chunks within one instruction by using array views. These views operate on GPU's global memory, but we can add similar concepts for local/shared and private memory.  
> 
> Additionally, Java/HAT developers can access lanes within the vector instance by name:
> 
> 
> vA.x(myNewValue);
> float foo = vA.z();
> ``` 
> 
> We will iterate over this proposal. 
> 
> How to test?
> 
> For CUDA
> 
> HAT=SHOW_CODE,INFO,SHOW_COMPILATION_PHASES java @hat/otest ffi-cuda oracle.code.hat.TestVectorTypes
> 
> 
> For OpenCL:
> 
> 
> 
> HAT=SHOW_CODE,INFO,SHOW_COMPILATION_PHASES java @hat/otest ffi-opencl oracle.code.hat.TestVectorTypes
> 
> 
> This PR also introduces a new option to debug the HAT Dialect Phases. 
> 
> 
> HAT=SHOW_COMPILATION_PHASES ...

Juan Fumero has updated the pull request incrementally with one additional commit since the last revision:

  [hat] F32ArrayPadded DS added

-------------

Changes:
  - all: https://git.openjdk.org/babylon/pull/606/files
  - new: https://git.openjdk.org/babylon/pull/606/files/5f37c6f7..7fbc18a7

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=babylon&pr=606&range=01
 - incr: https://webrevs.openjdk.org/?repo=babylon&pr=606&range=00-01

  Stats: 232 lines in 6 files changed: 137 ins; 14 del; 81 mod
  Patch: https://git.openjdk.org/babylon/pull/606.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/606/head:pull/606

PR: https://git.openjdk.org/babylon/pull/606

From jfumero at openjdk.org  Tue Oct 14 12:12:17 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Tue, 14 Oct 2025 12:12:17 GMT
Subject: git: openjdk/babylon: code-reflection: Proposal for Vector Types for
 GPUs (OpenCL and CUDA backends)
Message-ID: <bcb95f03-ac11-453b-a30c-c670e7cc1ca4@openjdk.org>

Changeset: 26afd5b8
Branch: code-reflection
Author:    Juan Fumero <jfumero at openjdk.org>
Date:      2025-10-14 12:09:46 +0000
URL:       https://git.openjdk.org/babylon/commit/26afd5b8e08e01fa126f88467ebdbb3d79f27a51

Proposal for Vector Types for GPUs (OpenCL and CUDA backends)

! hat/backends/ffi/cuda/src/main/java/hat/backend/ffi/CudaHATKernelBuilder.java
! hat/backends/ffi/opencl/src/main/java/hat/backend/ffi/OpenCLHATKernelBuilder.java
! hat/core/src/main/java/hat/Config.java
! hat/core/src/main/java/hat/buffer/F32Array.java
! hat/core/src/main/java/hat/buffer/F32Array2D.java
+ hat/core/src/main/java/hat/buffer/F32ArrayPadded.java
+ hat/core/src/main/java/hat/buffer/Float4.java
+ hat/core/src/main/java/hat/buffer/HatVector.java
! hat/core/src/main/java/hat/buffer/S08x3RGBImage.java
! hat/core/src/main/java/hat/buffer/S32Array.java
- hat/core/src/main/java/hat/buffer/S32LocalArray.java
! hat/core/src/main/java/hat/callgraph/KernelCallGraph.java
! hat/core/src/main/java/hat/codebuilders/BabylonOpBuilder.java
! hat/core/src/main/java/hat/codebuilders/C99HATComputeBuilder.java
! hat/core/src/main/java/hat/codebuilders/HATCodeBuilder.java
! hat/core/src/main/java/hat/codebuilders/HATCodeBuilderWithContext.java
! hat/core/src/main/java/hat/codebuilders/ScopedCodeBuilderContext.java
+ hat/core/src/main/java/hat/dialect/HatVSelectLoadOp.java
+ hat/core/src/main/java/hat/dialect/HatVSelectStoreOp.java
+ hat/core/src/main/java/hat/dialect/HatVectorAddOp.java
+ hat/core/src/main/java/hat/dialect/HatVectorBinaryOp.java
+ hat/core/src/main/java/hat/dialect/HatVectorDivOp.java
+ hat/core/src/main/java/hat/dialect/HatVectorLoadOp.java
+ hat/core/src/main/java/hat/dialect/HatVectorMulOp.java
+ hat/core/src/main/java/hat/dialect/HatVectorStoreView.java
+ hat/core/src/main/java/hat/dialect/HatVectorSubOp.java
+ hat/core/src/main/java/hat/dialect/HatVectorVarLoadOp.java
+ hat/core/src/main/java/hat/dialect/HatVectorVarOp.java
+ hat/core/src/main/java/hat/dialect/HatVectorViewOp.java
! hat/core/src/main/java/hat/optools/OpTk.java
+ hat/core/src/main/java/hat/phases/HatCompilationTier.java
+ hat/core/src/main/java/hat/phases/HatDialectAbstractPhase.java
! hat/core/src/main/java/hat/phases/HatDialectifyBarrierPhase.java
! hat/core/src/main/java/hat/phases/HatDialectifyMemoryPhase.java
! hat/core/src/main/java/hat/phases/HatDialectifyThreadsPhase.java
! hat/core/src/main/java/hat/phases/HatDialectifyTier.java
+ hat/core/src/main/java/hat/phases/HatDialectifyVSelectPhase.java
+ hat/core/src/main/java/hat/phases/HatDialectifyVectorOpPhase.java
+ hat/core/src/main/java/hat/phases/HatDialectifyVectorStorePhase.java
! hat/examples/matmul/src/main/java/matmul/Main.java
! hat/hat/test.java
! hat/tests/src/main/java/oracle/code/hat/TestArrays.java
! hat/tests/src/main/java/oracle/code/hat/TestMatMul.java
+ hat/tests/src/main/java/oracle/code/hat/TestVectorTypes.java
! hat/tools/src/main/java/hat/tools/text/JavaHATCodeBuilder.java


From jfumero at openjdk.org  Tue Oct 14 12:12:49 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Tue, 14 Oct 2025 12:12:49 GMT
Subject: [code-reflection] Integrated: Proposal for Vector Types for GPUs
 (OpenCL and CUDA backends)
In-Reply-To: <laZ7WwpiImcpodRE2bHwYkw3HunfrRMVvNTTsoHKxmI=.a13dbc47-6d88-412a-b57a-ea2f9d92d783@github.com>
References: <laZ7WwpiImcpodRE2bHwYkw3HunfrRMVvNTTsoHKxmI=.a13dbc47-6d88-412a-b57a-ea2f9d92d783@github.com>
Message-ID: <KtdxUdWBtcvJR41SQLaovpAnpnQ95b30aHty8W9KiLg=.d1d8d0fb-e0ee-44b0-85cc-d7120c22e1da@github.com>

On Thu, 9 Oct 2025 14:34:08 GMT, Juan Fumero <jfumero at openjdk.org> wrote:

> This PR opens a new proposal to integrate an API for representing explicit vector types in HAT. 
> 
> The way to operate is as follows:
> 
> 
> Float4 vA = a.float4View(index * 4);
> Float4 vB = b.float4View(index * 4);
> Float4 vC = Float4.add(vA, vB);
> c.storeFloat4View(vC, index * 4);
> 
> 
> Java programmers can load/store float4 chunks within one instruction by using array views. These views operate on GPU's global memory, but we can add similar concepts for local/shared and private memory.  
> 
> Additionally, Java/HAT developers can access lanes within the vector instance by name:
> 
> 
> vA.x(myNewValue);
> float foo = vA.z();
> ``` 
> 
> This PR also extends the HAT API to allow vector functions in a chain.  For example:
> 
> 
> va.add(vb).mul(vc).div(va.mul(va));
> 
> 
> We will iterate over this proposal. 
> 
> How to test?
> 
> For CUDA
> 
> HAT=SHOW_CODE,INFO,SHOW_COMPILATION_PHASES java @hat/otest ffi-cuda oracle.code.hat.TestVectorTypes
> 
> 
> For OpenCL:
> 
> 
> HAT=SHOW_CODE,INFO,SHOW_COMPILATION_PHASES java @hat/otest ffi-opencl oracle.code.hat.TestVectorTypes
> 
> 
> This PR also introduces a new option to debug the HAT Dialect Phases. 
> 
> 
> HAT=SHOW_COMPILATION_PHASES ...

This pull request has now been integrated.

Changeset: 26afd5b8
Author:    Juan Fumero <jfumero at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/26afd5b8e08e01fa126f88467ebdbb3d79f27a51
Stats:     3489 lines in 45 files changed: 3344 ins; 88 del; 57 mod

Proposal for Vector Types for GPUs (OpenCL and CUDA backends)

-------------

PR: https://git.openjdk.org/babylon/pull/606

From duke at openjdk.org  Tue Oct 14 16:38:40 2025
From: duke at openjdk.org (Ruby Chen)
Date: Tue, 14 Oct 2025 16:38:40 GMT
Subject: [code-reflection] RFR: Add kernel entrypoint to ModuleOp [v2]
In-Reply-To: <UkyuDt6Ei-D4V2Ly2ukTEV-Sx9qLIbN6ih3KSJ1BHYY=.45a19d55-ee90-46ec-b41c-28ec98117c5b@github.com>
References: <UkyuDt6Ei-D4V2Ly2ukTEV-Sx9qLIbN6ih3KSJ1BHYY=.45a19d55-ee90-46ec-b41c-28ec98117c5b@github.com>
Message-ID: <B-rHQDhNdzTsMptMH8YoDnPvu36qEy0H-8-iclLcdcQ=.0f808bc3-f393-4a2d-bfe7-6ec8453b8469@github.com>

> Include the kernel (and compute) entrypoints in their respective ModuleOps.

Ruby Chen has updated the pull request with a new target base due to a merge or a rebase. The pull request now contains two commits:

 - Merge branch 'code-reflection' into cleanup
 - Add kernel entrypoint to ModuleOp

-------------

Changes: https://git.openjdk.org/babylon/pull/605/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=605&range=01
  Stats: 45 lines in 2 files changed: 11 ins; 22 del; 12 mod
  Patch: https://git.openjdk.org/babylon/pull/605.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/605/head:pull/605

PR: https://git.openjdk.org/babylon/pull/605

From psandoz at openjdk.org  Tue Oct 14 16:42:27 2025
From: psandoz at openjdk.org (Paul Sandoz)
Date: Tue, 14 Oct 2025 16:42:27 GMT
Subject: [babylon-docs:master] Integrated: Links to latest presentations
Message-ID: <t30aYQJQlqqflCb2mZ615zYE3IEklVmlNim7vZzUrTY=.2d2b22b4-0a5c-46fb-845a-ca787fc936c5@github.com>

Add links to JVMLS 2025 and Devoxx Belgium 2025 talks.

-------------

Commit messages:
 - Links to latest presentations.

Changes: https://git.openjdk.org/babylon-docs/pull/8/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon-docs&pr=8&range=00
  Stats: 7 lines in 1 file changed: 7 ins; 0 del; 0 mod
  Patch: https://git.openjdk.org/babylon-docs/pull/8.diff
  Fetch: git fetch https://git.openjdk.org/babylon-docs.git pull/8/head:pull/8

PR: https://git.openjdk.org/babylon-docs/pull/8

From psandoz at openjdk.org  Tue Oct 14 16:42:28 2025
From: psandoz at openjdk.org (Paul Sandoz)
Date: Tue, 14 Oct 2025 16:42:28 GMT
Subject: [babylon-docs:master] Integrated: Links to latest presentations
In-Reply-To: <t30aYQJQlqqflCb2mZ615zYE3IEklVmlNim7vZzUrTY=.2d2b22b4-0a5c-46fb-845a-ca787fc936c5@github.com>
References: <t30aYQJQlqqflCb2mZ615zYE3IEklVmlNim7vZzUrTY=.2d2b22b4-0a5c-46fb-845a-ca787fc936c5@github.com>
Message-ID: <CY6-AqcYBQqzK9D0JjK0XQcktPxi0RyGmn4R6eRJDwo=.5ab39efa-aeac-4662-8fd2-e0d6c0cf5a76@github.com>

On Tue, 14 Oct 2025 16:34:31 GMT, Paul Sandoz <psandoz at openjdk.org> wrote:

> Add links to JVMLS 2025 and Devoxx Belgium 2025 talks.

This pull request has now been integrated.

Changeset: bb5f9e25
Author:    Paul Sandoz <psandoz at openjdk.org>
URL:       https://git.openjdk.org/babylon-docs/commit/bb5f9e2562cf85fc7e33e22c587bbdd6fc020e40
Stats:     7 lines in 1 file changed: 7 ins; 0 del; 0 mod

Links to latest presentations

-------------

PR: https://git.openjdk.org/babylon-docs/pull/8

From duke at openjdk.org  Tue Oct 14 16:52:50 2025
From: duke at openjdk.org (duke)
Date: Tue, 14 Oct 2025 16:52:50 GMT
Subject: [code-reflection] RFR: Add kernel entrypoint to ModuleOp [v2]
In-Reply-To: <B-rHQDhNdzTsMptMH8YoDnPvu36qEy0H-8-iclLcdcQ=.0f808bc3-f393-4a2d-bfe7-6ec8453b8469@github.com>
References: <UkyuDt6Ei-D4V2Ly2ukTEV-Sx9qLIbN6ih3KSJ1BHYY=.45a19d55-ee90-46ec-b41c-28ec98117c5b@github.com>
 <B-rHQDhNdzTsMptMH8YoDnPvu36qEy0H-8-iclLcdcQ=.0f808bc3-f393-4a2d-bfe7-6ec8453b8469@github.com>
Message-ID: <dsRGpMz7hzSFfCy6fg_ZRnLih3e0qI1O_JAeMeLyodE=.003ea07d-d494-402e-9b69-c1aa9e5147c8@github.com>

On Tue, 14 Oct 2025 16:38:40 GMT, Ruby Chen <duke at openjdk.org> wrote:

>> Include the kernel (and compute) entrypoints in their respective ModuleOps.
>
> Ruby Chen has updated the pull request with a new target base due to a merge or a rebase. The pull request now contains two commits:
> 
>  - Merge branch 'code-reflection' into cleanup
>  - Add kernel entrypoint to ModuleOp

@rbrchen 
Your change (at version 408391ce04fe44752f7acb6df30140a68178fd78) is now ready to be sponsored by a Committer.

-------------

PR Comment: https://git.openjdk.org/babylon/pull/605#issuecomment-3402793105

From psandoz at openjdk.org  Tue Oct 14 16:58:35 2025
From: psandoz at openjdk.org (Paul Sandoz)
Date: Tue, 14 Oct 2025 16:58:35 GMT
Subject: [babylon-docs:master] Integrated: Add link to J1 preso
Message-ID: <uwxhO-wQHCmk3oB7DPkAV2AhQd7j4rQ8AKGWg2PM5lk=.5bbf3c59-010c-4160-b46a-097e98f99a24@github.com>

Add link to Gary's J1 preso.

-------------

Commit messages:
 - Add J1 preso

Changes: https://git.openjdk.org/babylon-docs/pull/9/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon-docs&pr=9&range=00
  Stats: 3 lines in 1 file changed: 3 ins; 0 del; 0 mod
  Patch: https://git.openjdk.org/babylon-docs/pull/9.diff
  Fetch: git fetch https://git.openjdk.org/babylon-docs.git pull/9/head:pull/9

PR: https://git.openjdk.org/babylon-docs/pull/9

From psandoz at openjdk.org  Tue Oct 14 16:58:35 2025
From: psandoz at openjdk.org (Paul Sandoz)
Date: Tue, 14 Oct 2025 16:58:35 GMT
Subject: [babylon-docs:master] Integrated: Add link to J1 preso
In-Reply-To: <uwxhO-wQHCmk3oB7DPkAV2AhQd7j4rQ8AKGWg2PM5lk=.5bbf3c59-010c-4160-b46a-097e98f99a24@github.com>
References: <uwxhO-wQHCmk3oB7DPkAV2AhQd7j4rQ8AKGWg2PM5lk=.5bbf3c59-010c-4160-b46a-097e98f99a24@github.com>
Message-ID: <c4tp65W5bX4tJhvM-kJOPu0HWR6k3Oc7zMcKal2C5vc=.2ec5d58f-6553-40e3-ac3a-a47d672a4136@github.com>

On Tue, 14 Oct 2025 16:48:22 GMT, Paul Sandoz <psandoz at openjdk.org> wrote:

> Add link to Gary's J1 preso.

This pull request has now been integrated.

Changeset: 26829e89
Author:    Paul Sandoz <psandoz at openjdk.org>
URL:       https://git.openjdk.org/babylon-docs/commit/26829e89513b8522957aa60775a749a77b930293
Stats:     3 lines in 1 file changed: 3 ins; 0 del; 0 mod

Add link to J1 preso

-------------

PR: https://git.openjdk.org/babylon-docs/pull/9

From gfrost at openjdk.org  Wed Oct 15 12:15:46 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Wed, 15 Oct 2025 12:15:46 GMT
Subject: git: openjdk/babylon: code-reflection: Decouple example-experiments
 from opencl + add example_arrayview to intellij
Message-ID: <de43e457-ea65-4e36-a37d-7d2690e61c5b@openjdk.org>

Changeset: cdb0f4a9
Branch: code-reflection
Author:    Gary Frost <gfrost at openjdk.org>
Date:      2025-10-15 12:13:18 +0000
URL:       https://git.openjdk.org/babylon/commit/cdb0f4a9d1cccb68c36df0dfcfffa9ef22f39cb9

Decouple example-experiments from opencl + add example_arrayview to intellij

! hat/backends/jextracted/opencl/src/main/java/hat/backend/jextracted/OpenCLHatKernelBuilder.java
+ hat/docs/duke-hat.svg
! hat/examples/experiments/src/main/java/experiments/Mesh.java
! hat/examples/experiments/src/main/java/experiments/MinBufferTest.java
! hat/hat.java
! hat/intellij/.idea/modules.xml
+ hat/intellij/example_arrayview.iml
! hat/intellij/example_experiments.iml


From gfrost at openjdk.org  Wed Oct 15 12:16:47 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Wed, 15 Oct 2025 12:16:47 GMT
Subject: [code-reflection] Integrated: Decouple example-experiments from opencl
 + add example_arrayview to intellij
Message-ID: <V_BMPAAVzBDakPIdME3rLdGwhZqYcs-EpLU0zgvVO2Y=.f95e2bc5-66dc-42f7-9c8b-2fa6650b1103@github.com>

Some housekeeping.  

We should never have example-experiments which depend on a specific backend.  We had two Mesh.java and MinBufferTest.java 

Otherwide `@hat/exp yada experiment` will always pick up the specific backend and will ignor ethe one provided. 

I also fixed the HatOpenCLKernelBuilder in jextracted opendl.  I think this needs removing, but at least it compiles now.

-------------

Commit messages:
 - Decouple example-experiments from opencl backend + add example_arrayview to intellij proj

Changes: https://git.openjdk.org/babylon/pull/608/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=608&range=00
  Stats: 256 lines in 8 files changed: 233 ins; 18 del; 5 mod
  Patch: https://git.openjdk.org/babylon/pull/608.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/608/head:pull/608

PR: https://git.openjdk.org/babylon/pull/608

From gfrost at openjdk.org  Wed Oct 15 12:16:49 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Wed, 15 Oct 2025 12:16:49 GMT
Subject: [code-reflection] Integrated: Decouple example-experiments from opencl
 + add example_arrayview to intellij
In-Reply-To: <V_BMPAAVzBDakPIdME3rLdGwhZqYcs-EpLU0zgvVO2Y=.f95e2bc5-66dc-42f7-9c8b-2fa6650b1103@github.com>
References: <V_BMPAAVzBDakPIdME3rLdGwhZqYcs-EpLU0zgvVO2Y=.f95e2bc5-66dc-42f7-9c8b-2fa6650b1103@github.com>
Message-ID: <QYFTjy3C0QS0srcon_FBXNnjfPrGjccMUhGTPzL9WQw=.d2530607-f834-4ae8-a8a3-6eb8fb9e1492@github.com>

On Wed, 15 Oct 2025 12:08:16 GMT, Gary Frost <gfrost at openjdk.org> wrote:

> Some housekeeping.  
> 
> We should never have example-experiments which depend on a specific backend.  We had two Mesh.java and MinBufferTest.java 
> 
> Otherwide `@hat/exp yada experiment` will always pick up the specific backend and will ignor ethe one provided. 
> 
> I also fixed the HatOpenCLKernelBuilder in jextracted opendl.  I think this needs removing, but at least it compiles now.

This pull request has now been integrated.

Changeset: cdb0f4a9
Author:    Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/cdb0f4a9d1cccb68c36df0dfcfffa9ef22f39cb9
Stats:     256 lines in 8 files changed: 233 ins; 18 del; 5 mod

Decouple example-experiments from opencl + add example_arrayview to intellij

-------------

PR: https://git.openjdk.org/babylon/pull/608

From gfrost at openjdk.org  Wed Oct 15 12:43:41 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Wed, 15 Oct 2025 12:43:41 GMT
Subject: [code-reflection] Integrated: Updated Config
Message-ID: <J-ZyyCljweamFFXpPX_9QxCuv4mtVg0aNVsJ0fnzMVY=.6c01b3ed-698a-4c7c-b8b7-6b591df4da4f@github.com>

Added some config bits in prep for Backend changes.

-------------

Commit messages:
 - Updated Config

Changes: https://git.openjdk.org/babylon/pull/609/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=609&range=00
  Stats: 35 lines in 2 files changed: 10 ins; 15 del; 10 mod
  Patch: https://git.openjdk.org/babylon/pull/609.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/609/head:pull/609

PR: https://git.openjdk.org/babylon/pull/609

From gfrost at openjdk.org  Wed Oct 15 12:43:42 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Wed, 15 Oct 2025 12:43:42 GMT
Subject: [code-reflection] Integrated: Updated Config
In-Reply-To: <J-ZyyCljweamFFXpPX_9QxCuv4mtVg0aNVsJ0fnzMVY=.6c01b3ed-698a-4c7c-b8b7-6b591df4da4f@github.com>
References: <J-ZyyCljweamFFXpPX_9QxCuv4mtVg0aNVsJ0fnzMVY=.6c01b3ed-698a-4c7c-b8b7-6b591df4da4f@github.com>
Message-ID: <x27M3znyakgeAst040MNzz1lxgsRsdTtYUovlzBOp94=.0b76b688-1733-4ba3-848b-1e22161408ba@github.com>

On Wed, 15 Oct 2025 12:36:22 GMT, Gary Frost <gfrost at openjdk.org> wrote:

> Added some config bits in prep for Backend changes.

This pull request has now been integrated.

Changeset: 31ee3dc4
Author:    Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/31ee3dc417fabdcc9581598b0fe7cd70391732d9
Stats:     35 lines in 2 files changed: 10 ins; 15 del; 10 mod

Updated Config

-------------

PR: https://git.openjdk.org/babylon/pull/609

From gfrost at openjdk.org  Wed Oct 15 12:44:28 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Wed, 15 Oct 2025 12:44:28 GMT
Subject: git: openjdk/babylon: code-reflection: Updated Config
Message-ID: <70cee9bf-06c8-48ab-b4d0-5ed3eba2f5ee@openjdk.org>

Changeset: 31ee3dc4
Branch: code-reflection
Author:    Gary Frost <gfrost at openjdk.org>
Date:      2025-10-15 12:40:07 +0000
URL:       https://git.openjdk.org/babylon/commit/31ee3dc417fabdcc9581598b0fe7cd70391732d9

Updated Config

! hat/backends/ffi/shared/src/main/native/include/config.h
! hat/core/src/main/java/hat/Config.java


From duke at openjdk.org  Wed Oct 15 18:56:56 2025
From: duke at openjdk.org (Ruby Chen)
Date: Wed, 15 Oct 2025 18:56:56 GMT
Subject: [code-reflection] RFR: Update buffer tagging with arrayview support
Message-ID: <Q9zdEEQTILTtqvItQQ26aC5TCXabT0YwpnlNFFY-fZg=.e0f9b877-1021-4d2f-b8e0-b687928a61dc@github.com>

Add buffer tagging for array views, and also use `elements()` instead of `traverse()`.

-------------

Commit messages:
 - Update buffer tagging with arrayview support

Changes: https://git.openjdk.org/babylon/pull/610/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=610&range=00
  Stats: 39 lines in 2 files changed: 17 ins; 3 del; 19 mod
  Patch: https://git.openjdk.org/babylon/pull/610.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/610/head:pull/610

PR: https://git.openjdk.org/babylon/pull/610

From duke at openjdk.org  Wed Oct 15 19:01:50 2025
From: duke at openjdk.org (duke)
Date: Wed, 15 Oct 2025 19:01:50 GMT
Subject: [code-reflection] RFR: Update buffer tagging with arrayview
 support
In-Reply-To: <Q9zdEEQTILTtqvItQQ26aC5TCXabT0YwpnlNFFY-fZg=.e0f9b877-1021-4d2f-b8e0-b687928a61dc@github.com>
References: <Q9zdEEQTILTtqvItQQ26aC5TCXabT0YwpnlNFFY-fZg=.e0f9b877-1021-4d2f-b8e0-b687928a61dc@github.com>
Message-ID: <L-4mBTyjnEGVXIeNfhCXhRTc3RMM7apAFgWiw8-BHwU=.438aa854-2997-4917-838d-052ca42a7ce5@github.com>

On Wed, 15 Oct 2025 18:50:16 GMT, Ruby Chen <duke at openjdk.org> wrote:

> Add buffer tagging for array views, and also use `elements()` instead of `traverse()`.

@rbrchen 
Your change (at version 0da1fe6ddea18a405bc365ca095766178b33d718) is now ready to be sponsored by a Committer.

-------------

PR Comment: https://git.openjdk.org/babylon/pull/610#issuecomment-3407838458

From jfumero at openjdk.org  Thu Oct 16 08:26:48 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Thu, 16 Oct 2025 08:26:48 GMT
Subject: [code-reflection] RFR: [hat] KernelContext from the hat.buffer package
 renamed to KernelBufferContext
Message-ID: <P3FSKATVw8kNxeEJwKYUYA4WD_OydJOODyrmO9TgRCE=.ad3d5069-b088-4d10-809d-af6ba87476cc@github.com>

`KernelContext` from the hat.buffer package renamed to `KernelBufferContext`.

-------------

Commit messages:
 - [hat] KernelContext from the hat.buffer package renamed to KernelBufferContext

Changes: https://git.openjdk.org/babylon/pull/611/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=611&range=00
  Stats: 314 lines in 4 files changed: 146 ins; 147 del; 21 mod
  Patch: https://git.openjdk.org/babylon/pull/611.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/611/head:pull/611

PR: https://git.openjdk.org/babylon/pull/611

From jfumero at openjdk.org  Thu Oct 16 08:34:50 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Thu, 16 Oct 2025 08:34:50 GMT
Subject: [code-reflection] Integrated: [hat] KernelContext from the hat.buffer
 package renamed to KernelBufferContext
In-Reply-To: <P3FSKATVw8kNxeEJwKYUYA4WD_OydJOODyrmO9TgRCE=.ad3d5069-b088-4d10-809d-af6ba87476cc@github.com>
References: <P3FSKATVw8kNxeEJwKYUYA4WD_OydJOODyrmO9TgRCE=.ad3d5069-b088-4d10-809d-af6ba87476cc@github.com>
Message-ID: <CeJyO0cj7LoZM4kIwdlbiXNoLGjEXIfMgvAMIJBtsq8=.c10d6135-93c3-40bf-8554-785af1e5fa0d@github.com>

On Thu, 16 Oct 2025 08:18:45 GMT, Juan Fumero <jfumero at openjdk.org> wrote:

> `KernelContext` from the hat.buffer package renamed to `KernelBufferContext`.

This pull request has now been integrated.

Changeset: eacbfd77
Author:    Juan Fumero <jfumero at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/eacbfd7760416e06088012400a2565b88124ee90
Stats:     314 lines in 4 files changed: 146 ins; 147 del; 21 mod

[hat] KernelContext from the hat.buffer package renamed to KernelBufferContext

-------------

PR: https://git.openjdk.org/babylon/pull/611

From jfumero at openjdk.org  Thu Oct 16 08:36:22 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Thu, 16 Oct 2025 08:36:22 GMT
Subject: git: openjdk/babylon: code-reflection: [hat] KernelContext from the
 hat.buffer package renamed to KernelBufferContext
Message-ID: <78628dbe-9574-44e1-8b2c-1a88e0b216d7@openjdk.org>

Changeset: eacbfd77
Branch: code-reflection
Author:    Juan Fumero <jfumero at openjdk.org>
Date:      2025-10-16 08:31:45 +0000
URL:       https://git.openjdk.org/babylon/commit/eacbfd7760416e06088012400a2565b88124ee90

[hat] KernelContext from the hat.buffer package renamed to KernelBufferContext

! hat/backends/ffi/shared/src/main/java/hat/backend/ffi/C99FFIBackend.java
+ hat/core/src/main/java/hat/buffer/KernelBufferContext.java
- hat/core/src/main/java/hat/buffer/KernelContext.java
! hat/core/src/main/java/hat/optools/OpTk.java


From gfrost at openjdk.org  Thu Oct 16 13:06:23 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Thu, 16 Oct 2025 13:06:23 GMT
Subject: git: openjdk/babylon: code-reflection: Hat add kernel code annotation
Message-ID: <8789e213-43f4-414a-bfc7-6dee728914ca@openjdk.org>

Changeset: 0ce11a2c
Branch: code-reflection
Author:    Gary Frost <gfrost at openjdk.org>
Date:      2025-10-16 13:05:19 +0000
URL:       https://git.openjdk.org/babylon/commit/0ce11a2cc11c3c7a3541043bb5bb418cc8094080

Hat add kernel code annotation

! hat/backends/ffi/cuda/src/main/java/hat/backend/ffi/CudaBackend.java
! hat/backends/ffi/cuda/src/main/java/hat/backend/ffi/CudaHATKernelBuilder.java
! hat/backends/ffi/cuda/src/main/java/hat/backend/ffi/PTXHATKernelBuilder.java
! hat/backends/ffi/mock/src/main/java/hat/backend/ffi/MockBackend.java
! hat/backends/ffi/opencl/src/main/java/hat/backend/ffi/OpenCLHATKernelBuilder.java
! hat/backends/ffi/shared/src/main/java/hat/backend/ffi/C99FFIBackend.java
! hat/backends/ffi/shared/src/main/java/hat/backend/ffi/FFIBackend.java
! hat/backends/jextracted/opencl/src/main/java/hat/backend/jextracted/OpenCLHatKernelBuilder.java
! hat/backends/jextracted/shared/src/main/java/hat/backend/jextracted/C99JExtractedBackend.java
! hat/backends/jextracted/shared/src/main/java/hat/backend/jextracted/JExtractedBackend.java
! hat/core/src/main/java/hat/BufferTagger.java
! hat/core/src/main/java/hat/ComputeContext.java
! hat/core/src/main/java/hat/backend/DebugBackend.java
! hat/core/src/main/java/hat/callgraph/CallGraph.java
! hat/core/src/main/java/hat/callgraph/ComputeCallGraph.java
! hat/core/src/main/java/hat/callgraph/KernelCallGraph.java
! hat/core/src/main/java/hat/codebuilders/BabylonOpBuilder.java
! hat/core/src/main/java/hat/codebuilders/C99HATKernelBuilder.java
! hat/core/src/main/java/hat/codebuilders/HATCodeBuilder.java
! hat/core/src/main/java/hat/codebuilders/HATCodeBuilderWithContext.java
! hat/core/src/main/java/hat/codebuilders/ScopedCodeBuilderContext.java
= hat/core/src/main/java/hat/dialect/HATBarrierOp.java
+ hat/core/src/main/java/hat/dialect/HATBlockThreadIdOp.java
+ hat/core/src/main/java/hat/dialect/HATGlobalSizeOp.java
+ hat/core/src/main/java/hat/dialect/HATGlobalThreadIdOp.java
+ hat/core/src/main/java/hat/dialect/HATLocalSizeOp.java
+ hat/core/src/main/java/hat/dialect/HATLocalThreadIdOp.java
+ hat/core/src/main/java/hat/dialect/HATLocalVarOp.java
= hat/core/src/main/java/hat/dialect/HATMemoryOp.java
+ hat/core/src/main/java/hat/dialect/HATOp.java
= hat/core/src/main/java/hat/dialect/HATPrivateVarOp.java
+ hat/core/src/main/java/hat/dialect/HATThreadOp.java
= hat/core/src/main/java/hat/dialect/HATVSelectLoadOp.java
= hat/core/src/main/java/hat/dialect/HATVSelectStoreOp.java
+ hat/core/src/main/java/hat/dialect/HATVectorAddOp.java
= hat/core/src/main/java/hat/dialect/HATVectorBinaryOp.java
+ hat/core/src/main/java/hat/dialect/HATVectorDivOp.java
= hat/core/src/main/java/hat/dialect/HATVectorLoadOp.java
+ hat/core/src/main/java/hat/dialect/HATVectorMulOp.java
= hat/core/src/main/java/hat/dialect/HATVectorStoreView.java
+ hat/core/src/main/java/hat/dialect/HATVectorSubOp.java
+ hat/core/src/main/java/hat/dialect/HATVectorVarLoadOp.java
= hat/core/src/main/java/hat/dialect/HATVectorVarOp.java
= hat/core/src/main/java/hat/dialect/HATVectorViewOp.java
- hat/core/src/main/java/hat/dialect/HatBlockThreadIdOp.java
- hat/core/src/main/java/hat/dialect/HatGlobalSizeOp.java
- hat/core/src/main/java/hat/dialect/HatGlobalThreadIdOp.java
- hat/core/src/main/java/hat/dialect/HatLocalSizeOp.java
- hat/core/src/main/java/hat/dialect/HatLocalThreadIdOp.java
- hat/core/src/main/java/hat/dialect/HatLocalVarOp.java
- hat/core/src/main/java/hat/dialect/HatOP.java
- hat/core/src/main/java/hat/dialect/HatThreadOP.java
- hat/core/src/main/java/hat/dialect/HatVectorAddOp.java
- hat/core/src/main/java/hat/dialect/HatVectorDivOp.java
- hat/core/src/main/java/hat/dialect/HatVectorMulOp.java
- hat/core/src/main/java/hat/dialect/HatVectorSubOp.java
- hat/core/src/main/java/hat/dialect/HatVectorVarLoadOp.java
! hat/core/src/main/java/hat/optools/OpTk.java
= hat/core/src/main/java/hat/phases/HATCompilationTier.java
+ hat/core/src/main/java/hat/phases/HATDialectAbstractPhase.java
+ hat/core/src/main/java/hat/phases/HATDialectifyBarrierPhase.java
+ hat/core/src/main/java/hat/phases/HATDialectifyMemoryPhase.java
= hat/core/src/main/java/hat/phases/HATDialectifyPhase.java
+ hat/core/src/main/java/hat/phases/HATDialectifyThreadsPhase.java
+ hat/core/src/main/java/hat/phases/HATDialectifyTier.java
+ hat/core/src/main/java/hat/phases/HATDialectifyVSelectPhase.java
+ hat/core/src/main/java/hat/phases/HATDialectifyVectorOpPhase.java
+ hat/core/src/main/java/hat/phases/HATDialectifyVectorStorePhase.java
= hat/core/src/main/java/hat/phases/HATFinalDetectionPhase.java
= hat/core/src/main/java/hat/phases/HATPhase.java
- hat/core/src/main/java/hat/phases/HatDialectAbstractPhase.java
- hat/core/src/main/java/hat/phases/HatDialectifyBarrierPhase.java
- hat/core/src/main/java/hat/phases/HatDialectifyMemoryPhase.java
- hat/core/src/main/java/hat/phases/HatDialectifyThreadsPhase.java
- hat/core/src/main/java/hat/phases/HatDialectifyTier.java
- hat/core/src/main/java/hat/phases/HatDialectifyVSelectPhase.java
- hat/core/src/main/java/hat/phases/HatDialectifyVectorOpPhase.java
- hat/core/src/main/java/hat/phases/HatDialectifyVectorStorePhase.java
! hat/examples/experiments/src/main/java/experiments/DNA.java
! hat/examples/experiments/src/main/java/experiments/LayoutExample.java
! hat/examples/experiments/src/main/java/experiments/PrefixSum.java
! hat/examples/experiments/src/main/java/experiments/RawLayout.java
! hat/examples/experiments/src/main/java/experiments/Transform.java
! hat/tools/src/main/java/hat/tools/text/JavaHATCodeBuilder.java


From gfrost at openjdk.org  Thu Oct 16 13:08:41 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Thu, 16 Oct 2025 13:08:41 GMT
Subject: [code-reflection] Integrated: Hat add kernel code annotation
Message-ID: <l79C8292W3MuGlutmSJldxR_k-RdCqm7K8ukoouonSI=.bee7246b-b53b-4ea8-a6ba-46fef1501da0@github.com>

Tracking side table issue. 

Not fixed yet, but homing in.  

 I have moved all transforms (and traverses and stream access to ops) into central OpTK apis in preparation for having a single place for updating sidetables.
 
 I added a tracing mechanism for reporting accesses to transform, element stream access and traverses of funcops . This is useful for seeing where we are 'inadvertantly' transforming and possibly trashing sidetable.
 

java @hat/run ffi-opencl -DTRACE_CALLSITES=true life 


This PR was getting too large and I was worried about future merges. So I am pushing now.
This code still has the side table issue, but we have some tooling to help find it

-------------

Commit messages:
 - Merge branch 'code-reflection' into hat-add-kernel-code-annotation
 - Tracking down issue with side tables
 - Add annotation and fixup OpTk

Changes: https://git.openjdk.org/babylon/pull/612/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=612&range=00
  Stats: 4946 lines in 84 files changed: 2327 ins; 2203 del; 416 mod
  Patch: https://git.openjdk.org/babylon/pull/612.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/612/head:pull/612

PR: https://git.openjdk.org/babylon/pull/612

From gfrost at openjdk.org  Thu Oct 16 13:08:41 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Thu, 16 Oct 2025 13:08:41 GMT
Subject: [code-reflection] Integrated: Hat add kernel code annotation
In-Reply-To: <l79C8292W3MuGlutmSJldxR_k-RdCqm7K8ukoouonSI=.bee7246b-b53b-4ea8-a6ba-46fef1501da0@github.com>
References: <l79C8292W3MuGlutmSJldxR_k-RdCqm7K8ukoouonSI=.bee7246b-b53b-4ea8-a6ba-46fef1501da0@github.com>
Message-ID: <Q5Hm1x22ETtXfYqAhwNMtW-M7CQC8RNoSSMvh2yI3DQ=.1d15d649-6080-4d34-8130-19951991d431@github.com>

On Thu, 16 Oct 2025 12:58:45 GMT, Gary Frost <gfrost at openjdk.org> wrote:

> Tracking side table issue. 
> 
> Not fixed yet, but homing in.  
> 
>  I have moved all transforms (and traverses and stream access to ops) into central OpTK apis in preparation for having a single place for updating sidetables.
>  
>  I added a tracing mechanism for reporting accesses to transform, element stream access and traverses of funcops . This is useful for seeing where we are 'inadvertantly' transforming and possibly trashing sidetable.
>  
> 
> java @hat/run ffi-opencl -DTRACE_CALLSITES=true life 
> 
> 
> This PR was getting too large and I was worried about future merges. So I am pushing now.
> This code still has the side table issue, but we have some tooling to help find it

This pull request has now been integrated.

Changeset: 0ce11a2c
Author:    Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/0ce11a2cc11c3c7a3541043bb5bb418cc8094080
Stats:     4946 lines in 84 files changed: 2327 ins; 2203 del; 416 mod

Hat add kernel code annotation

-------------

PR: https://git.openjdk.org/babylon/pull/612

From gfrost at openjdk.org  Thu Oct 16 13:18:57 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Thu, 16 Oct 2025 13:18:57 GMT
Subject: [code-reflection] Integrated: Fixed merge issues from last PR
Message-ID: <ueHaanqum-U2-o57Wyo8CgaWx6Iv1xsrR-sa6IIONLE=.ae4e2c31-0a41-4dae-8f9b-c7cbee2321ec@github.com>

I messed up the previous merge.. 

This PR has fixes

-------------

Commit messages:
 - Fixed merge issues from last PR

Changes: https://git.openjdk.org/babylon/pull/613/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=613&range=00
  Stats: 12 lines in 3 files changed: 2 ins; 5 del; 5 mod
  Patch: https://git.openjdk.org/babylon/pull/613.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/613/head:pull/613

PR: https://git.openjdk.org/babylon/pull/613

From gfrost at openjdk.org  Thu Oct 16 13:19:03 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Thu, 16 Oct 2025 13:19:03 GMT
Subject: git: openjdk/babylon: code-reflection: Fixed merge issues from last PR
Message-ID: <434fbdcc-5c2c-4026-83b0-c9f4f134db84@openjdk.org>

Changeset: bc4faed7
Branch: code-reflection
Author:    Gary Frost <gfrost at openjdk.org>
Date:      2025-10-16 13:16:10 +0000
URL:       https://git.openjdk.org/babylon/commit/bc4faed7aae7b56f68c5420098e4d26f69eaed07

Fixed merge issues from last PR

! hat/backends/jextracted/shared/src/main/java/hat/backend/jextracted/C99JExtractedBackend.java
= hat/core/src/main/java/hat/annotations/Kernel.java
! hat/core/src/main/java/hat/optools/OpTk.java


From gfrost at openjdk.org  Thu Oct 16 13:18:57 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Thu, 16 Oct 2025 13:18:57 GMT
Subject: [code-reflection] Integrated: Fixed merge issues from last PR
In-Reply-To: <ueHaanqum-U2-o57Wyo8CgaWx6Iv1xsrR-sa6IIONLE=.ae4e2c31-0a41-4dae-8f9b-c7cbee2321ec@github.com>
References: <ueHaanqum-U2-o57Wyo8CgaWx6Iv1xsrR-sa6IIONLE=.ae4e2c31-0a41-4dae-8f9b-c7cbee2321ec@github.com>
Message-ID: <NtgqtfsVYKQjJnajUU1cWJETUkKlXuZksTrc3DF7UWI=.ba606381-932a-4d66-bf96-df85fde12bef@github.com>

On Thu, 16 Oct 2025 13:12:38 GMT, Gary Frost <gfrost at openjdk.org> wrote:

> I messed up the previous merge.. 
> 
> This PR has fixes

This pull request has now been integrated.

Changeset: bc4faed7
Author:    Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/bc4faed7aae7b56f68c5420098e4d26f69eaed07
Stats:     12 lines in 3 files changed: 2 ins; 5 del; 5 mod

Fixed merge issues from last PR

-------------

PR: https://git.openjdk.org/babylon/pull/613

From asotona at openjdk.org  Thu Oct 16 13:43:10 2025
From: asotona at openjdk.org (Adam Sotona)
Date: Thu, 16 Oct 2025 13:43:10 GMT
Subject: [code-reflection] RFR: OnnxProtoBuilder fixes
Message-ID: <WFdUPj68hK2RfeloCpVSTCOu_PurIelHv-EE5S1jW8c=.559ab9e8-d555-4759-b2ba-7e6ac26fce77@github.com>

- fixes ONNX graph output name of returned tuple element
- fixes String AttributeProto value

-------------

Commit messages:
 - OnnxProtoBuilder fixes

Changes: https://git.openjdk.org/babylon/pull/614/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=614&range=00
  Stats: 7 lines in 1 file changed: 6 ins; 1 del; 0 mod
  Patch: https://git.openjdk.org/babylon/pull/614.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/614/head:pull/614

PR: https://git.openjdk.org/babylon/pull/614

From gfrost at openjdk.org  Thu Oct 16 14:44:47 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Thu, 16 Oct 2025 14:44:47 GMT
Subject: [code-reflection] Integrated: Moved dialect and arrayview transforms
 to KernelCallGraph constructor
Message-ID: <otrXBvOjRblA7pHp34BOPlTAhKNC8ObsDCOkIqsf8Gs=.fcbffaac-1cc1-4224-8c31-694188ddfbcb@github.com>

OK so this gets us performance back in HAT. 

I moved the dialect op mapping and array view mappings to KernelCallGraph creation (from first dispatch) 

Mandle now back to 180 GPS.

-------------

Commit messages:
 - Moved  dialect and arrayview transforms to KernelCallGraph constructor

Changes: https://git.openjdk.org/babylon/pull/615/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=615&range=00
  Stats: 80 lines in 4 files changed: 22 ins; 44 del; 14 mod
  Patch: https://git.openjdk.org/babylon/pull/615.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/615/head:pull/615

PR: https://git.openjdk.org/babylon/pull/615

From gfrost at openjdk.org  Thu Oct 16 14:44:47 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Thu, 16 Oct 2025 14:44:47 GMT
Subject: [code-reflection] Integrated: Moved dialect and arrayview transforms
 to KernelCallGraph constructor
In-Reply-To: <otrXBvOjRblA7pHp34BOPlTAhKNC8ObsDCOkIqsf8Gs=.fcbffaac-1cc1-4224-8c31-694188ddfbcb@github.com>
References: <otrXBvOjRblA7pHp34BOPlTAhKNC8ObsDCOkIqsf8Gs=.fcbffaac-1cc1-4224-8c31-694188ddfbcb@github.com>
Message-ID: <zS5qnVQah5X0GBe0h546-Jtpd7megB5oeNt7uzwT6sU=.97349a77-76a5-41ec-a925-c56cc51e11c1@github.com>

On Thu, 16 Oct 2025 14:38:03 GMT, Gary Frost <gfrost at openjdk.org> wrote:

> OK so this gets us performance back in HAT. 
> 
> I moved the dialect op mapping and array view mappings to KernelCallGraph creation (from first dispatch) 
> 
> Mandle now back to 180 GPS.

This pull request has now been integrated.

Changeset: fc5c4720
Author:    Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/fc5c47206b7ae71bf203564e8ad1e79e50adbb21
Stats:     80 lines in 4 files changed: 22 ins; 44 del; 14 mod

Moved  dialect and arrayview transforms to KernelCallGraph constructor

-------------

PR: https://git.openjdk.org/babylon/pull/615

From gfrost at openjdk.org  Thu Oct 16 14:42:48 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Thu, 16 Oct 2025 14:42:48 GMT
Subject: git: openjdk/babylon: code-reflection: Moved  dialect and arrayview
 transforms to KernelCallGraph constructor
Message-ID: <c5fd3518-4977-4b1c-9734-5d904d3d3ed4@openjdk.org>

Changeset: fc5c4720
Branch: code-reflection
Author:    Gary Frost <gfrost at openjdk.org>
Date:      2025-10-16 14:41:17 +0000
URL:       https://git.openjdk.org/babylon/commit/fc5c47206b7ae71bf203564e8ad1e79e50adbb21

Moved  dialect and arrayview transforms to KernelCallGraph constructor

! hat/core/src/main/java/hat/ComputeContext.java
! hat/core/src/main/java/hat/callgraph/CallGraph.java
! hat/core/src/main/java/hat/callgraph/ComputeCallGraph.java
! hat/core/src/main/java/hat/callgraph/KernelCallGraph.java


From jfumero at openjdk.org  Fri Oct 17 09:12:59 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Fri, 17 Oct 2025 09:12:59 GMT
Subject: git: openjdk/babylon: code-reflection: [hat] Testing framework moved
 to hat.test
Message-ID: <34e51898-3df4-4559-8ae5-c331349b89b9@openjdk.org>

Changeset: 9570005d
Branch: code-reflection
Author:    Juan Fumero <jfumero at openjdk.org>
Date:      2025-10-17 09:12:26 +0000
URL:       https://git.openjdk.org/babylon/commit/9570005d172674cfd0f5d07509903479d734dd07

[hat] Testing framework moved to hat.test

! hat/hat.java
! hat/hat/test.java
= hat/tests/src/main/java/hat/test/TestArrayView.java
= hat/tests/src/main/java/hat/test/TestArrays.java
= hat/tests/src/main/java/hat/test/TestBlackscholes.java
= hat/tests/src/main/java/hat/test/TestConstants.java
= hat/tests/src/main/java/hat/test/TestLocal.java
= hat/tests/src/main/java/hat/test/TestMandel.java
= hat/tests/src/main/java/hat/test/TestMatMul.java
= hat/tests/src/main/java/hat/test/TestNbody.java
= hat/tests/src/main/java/hat/test/TestParenthesis.java
= hat/tests/src/main/java/hat/test/TestPrivate.java
= hat/tests/src/main/java/hat/test/TestReductions.java
= hat/tests/src/main/java/hat/test/TestVectorTypes.java
= hat/tests/src/main/java/hat/test/annotation/HatTest.java
= hat/tests/src/main/java/hat/test/engine/Colours.java
= hat/tests/src/main/java/hat/test/engine/HatAssertionError.java
= hat/tests/src/main/java/hat/test/engine/HatAsserts.java
= hat/tests/src/main/java/hat/test/engine/HatTestEngine.java
= hat/tests/src/main/java/hat/test/engine/HatTestFormatter.java


From jfumero at openjdk.org  Fri Oct 17 09:16:59 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Fri, 17 Oct 2025 09:16:59 GMT
Subject: [code-reflection] Integrated: [hat] Testing framework moved to
 hat.test
Message-ID: <eB86ADICG7mCvF4N1dTM7dhTgdWfEr8JJCurgunrhBU=.b706488b-8782-45da-ada6-bba7cab2728e@github.com>

Refactor: testing framework moved from `oracle.code.hat` to `hat.test`

-------------

Commit messages:
 - [hat] Testing framework moved to hat.test

Changes: https://git.openjdk.org/babylon/pull/616/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=616&range=00
  Stats: 60 lines in 20 files changed: 0 ins; 1 del; 59 mod
  Patch: https://git.openjdk.org/babylon/pull/616.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/616/head:pull/616

PR: https://git.openjdk.org/babylon/pull/616

From jfumero at openjdk.org  Fri Oct 17 09:17:00 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Fri, 17 Oct 2025 09:17:00 GMT
Subject: [code-reflection] Integrated: [hat] Testing framework moved to
 hat.test
In-Reply-To: <eB86ADICG7mCvF4N1dTM7dhTgdWfEr8JJCurgunrhBU=.b706488b-8782-45da-ada6-bba7cab2728e@github.com>
References: <eB86ADICG7mCvF4N1dTM7dhTgdWfEr8JJCurgunrhBU=.b706488b-8782-45da-ada6-bba7cab2728e@github.com>
Message-ID: <YIpi_7Mw4C77LjoiDXWtXnM7KQ45oN21QrGLdJQeiJs=.10b5c322-de3a-4a2a-a85b-e101c4048b2e@github.com>

On Fri, 17 Oct 2025 09:07:06 GMT, Juan Fumero <jfumero at openjdk.org> wrote:

> Refactor: testing framework moved from `oracle.code.hat` to `hat.test`

This pull request has now been integrated.

Changeset: 9570005d
Author:    Juan Fumero <jfumero at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/9570005d172674cfd0f5d07509903479d734dd07
Stats:     60 lines in 20 files changed: 0 ins; 1 del; 59 mod

[hat] Testing framework moved to hat.test

-------------

PR: https://git.openjdk.org/babylon/pull/616

From jfumero at openjdk.org  Fri Oct 17 10:06:56 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Fri, 17 Oct 2025 10:06:56 GMT
Subject: git: openjdk/babylon: code-reflection: [hat] Runner for a single test
Message-ID: <9b787dd3-e0e7-4564-80e9-46944c13c45e@openjdk.org>

Changeset: e02d79f2
Branch: code-reflection
Author:    Juan Fumero <jfumero at openjdk.org>
Date:      2025-10-17 10:02:41 +0000
URL:       https://git.openjdk.org/babylon/commit/e02d79f2522c05117ceb7878c65023ef4bd43502

[hat] Runner for a single test

! hat/hat.java
! hat/scripts/remoteTesting.sh


From jfumero at openjdk.org  Fri Oct 17 10:07:57 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Fri, 17 Oct 2025 10:07:57 GMT
Subject: [code-reflection] Integrated: [hat] Runner for a single test
Message-ID: <Gv-ZDCDOWvk5ziHYAA_AyEK5YAqgGzNx9VGqo5h61XI=.7a1b3e6b-fd7f-48f3-9da4-08d5c34be1a5@github.com>

This patch extends the testing framework to execute single tests using the new hat launcher. 

Run the test suite:


java -cp hat/job.jar hat.java test-suite ffi-opencl


Run a single test class:


java -cp hat/job.jar hat.java test ffi-opencl hat.test.TestMatMul


Run a single test method:


java -cp hat/job.jar hat.java test ffi-opencl hat.test.TestMatMul#testMatrixMultiply1D

-------------

Commit messages:
 - [hat] Update remote testing with the test-suite
 - [hat] Runner for a single test

Changes: https://git.openjdk.org/babylon/pull/617/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=617&range=00
  Stats: 26 lines in 2 files changed: 24 ins; 0 del; 2 mod
  Patch: https://git.openjdk.org/babylon/pull/617.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/617/head:pull/617

PR: https://git.openjdk.org/babylon/pull/617

From jfumero at openjdk.org  Fri Oct 17 10:07:59 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Fri, 17 Oct 2025 10:07:59 GMT
Subject: [code-reflection] Integrated: [hat] Runner for a single test
In-Reply-To: <Gv-ZDCDOWvk5ziHYAA_AyEK5YAqgGzNx9VGqo5h61XI=.7a1b3e6b-fd7f-48f3-9da4-08d5c34be1a5@github.com>
References: <Gv-ZDCDOWvk5ziHYAA_AyEK5YAqgGzNx9VGqo5h61XI=.7a1b3e6b-fd7f-48f3-9da4-08d5c34be1a5@github.com>
Message-ID: <Q10epG828rG66mmsHLBrIP1AFWdcpwlSSOGQaTuKdTk=.08a5a67c-4062-4287-b747-35da9f995f2a@github.com>

On Fri, 17 Oct 2025 09:40:44 GMT, Juan Fumero <jfumero at openjdk.org> wrote:

> This patch extends the testing framework to execute single tests using the new hat launcher. 
> 
> Run the test suite:
> 
> 
> java -cp hat/job.jar hat.java test-suite ffi-opencl
> 
> 
> Run a single test class:
> 
> 
> java -cp hat/job.jar hat.java test ffi-opencl hat.test.TestMatMul
> 
> 
> Run a single test method:
> 
> 
> java -cp hat/job.jar hat.java test ffi-opencl hat.test.TestMatMul#testMatrixMultiply1D

This pull request has now been integrated.

Changeset: e02d79f2
Author:    Juan Fumero <jfumero at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/e02d79f2522c05117ceb7878c65023ef4bd43502
Stats:     26 lines in 2 files changed: 24 ins; 0 del; 2 mod

[hat] Runner for a single test

-------------

PR: https://git.openjdk.org/babylon/pull/617

From gfrost at openjdk.org  Fri Oct 17 12:05:09 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Fri, 17 Oct 2025 12:05:09 GMT
Subject: [code-reflection] Integrated: Added @hat/help and @hat/test-suite
Message-ID: <Ej33LmixallPX7OY4UOpTOPDoZj7wSwUyb3f99hjqFI=.dbc2b56c-0668-450a-b679-72184c5580d7@github.com>

Added @hat/help (display help) 

Also @hat/test-suite to access recent test updates. 

So 

  @hat/help 
  
  @hat/test-suite:  [ffi|my|seq]-[opencl|java|cuda|mock|hip]
             @hat/test-suite ffi-opencl

   @hat/test:  [ffi|my|seq]-[opencl|java|cuda|mock|hip] classToTest (also classToTest#method)
              @hat/test ffi-opencl  hat.test.TestMatMul
               @hat/test ffi-opencl  hat.test.TestMatMul#method (note at shell we need to escape #!! )

-------------

Commit messages:
 - Grrrr whitespace
 - Added @hat/help and @hat/test-suite

Changes: https://git.openjdk.org/babylon/pull/619/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=619&range=00
  Stats: 24 lines in 3 files changed: 10 ins; 10 del; 4 mod
  Patch: https://git.openjdk.org/babylon/pull/619.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/619/head:pull/619

PR: https://git.openjdk.org/babylon/pull/619

From gfrost at openjdk.org  Fri Oct 17 12:06:06 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Fri, 17 Oct 2025 12:06:06 GMT
Subject: git: openjdk/babylon: code-reflection: Added @hat/help and
 @hat/test-suite
Message-ID: <bc742a0e-761c-4877-a9c8-3018bd869a6a@openjdk.org>

Changeset: b944761f
Branch: code-reflection
Author:    Gary Frost <gfrost at openjdk.org>
Date:      2025-10-17 12:02:17 +0000
URL:       https://git.openjdk.org/babylon/commit/b944761feb0000e8b47051cbf16c8e9b207f3fa2

Added @hat/help and @hat/test-suite

! hat/hat.java
+ hat/hat/help
+ hat/hat/test-suite


From gfrost at openjdk.org  Fri Oct 17 12:05:10 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Fri, 17 Oct 2025 12:05:10 GMT
Subject: [code-reflection] Integrated: Added @hat/help and @hat/test-suite
In-Reply-To: <Ej33LmixallPX7OY4UOpTOPDoZj7wSwUyb3f99hjqFI=.dbc2b56c-0668-450a-b679-72184c5580d7@github.com>
References: <Ej33LmixallPX7OY4UOpTOPDoZj7wSwUyb3f99hjqFI=.dbc2b56c-0668-450a-b679-72184c5580d7@github.com>
Message-ID: <PHuR6Q0Em55xZOrMS7uNZPgmv-4NlvCrX3gYys3MsRI=.7c79a898-1c6e-45ac-9498-bcbfa3b2b4a5@github.com>

On Fri, 17 Oct 2025 11:53:47 GMT, Gary Frost <gfrost at openjdk.org> wrote:

> Added @hat/help (display help) 
> 
> Also @hat/test-suite to access recent test updates. 
> 
> So 
> 
>   @hat/help 
>   
>   @hat/test-suite:  [ffi|my|seq]-[opencl|java|cuda|mock|hip]
>              @hat/test-suite ffi-opencl
> 
>    @hat/test:  [ffi|my|seq]-[opencl|java|cuda|mock|hip] classToTest (also classToTest#method)
>               @hat/test ffi-opencl  hat.test.TestMatMul
>                @hat/test ffi-opencl  hat.test.TestMatMul#method (note at shell we need to escape #!! )

This pull request has now been integrated.

Changeset: b944761f
Author:    Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/b944761feb0000e8b47051cbf16c8e9b207f3fa2
Stats:     24 lines in 3 files changed: 10 ins; 10 del; 4 mod

Added @hat/help and @hat/test-suite

-------------

PR: https://git.openjdk.org/babylon/pull/619

From duke at openjdk.org  Fri Oct 17 12:48:25 2025
From: duke at openjdk.org (Ana Maria Mihalceanu)
Date: Fri, 17 Oct 2025 12:48:25 GMT
Subject: [code-reflection] RFR: Add support for other onnx providers (CoreML)
Message-ID: <e7Q32XrF2lzb5NdbJDAbklKlodvEyxR9Ds38h_GSDvA=.96e74c03-8271-461d-bb73-638d83d07ea0@github.com>

This PR contains a series of changes meant to integrate other ONNX providers.

* It contains generated Java bindings for CoreML provider: `oracle.code.onnx.foreign.coreml_provider_factory_h`
* Configurations of providers are present in `oracle.code.onnx.provider`.
* Provider configuration should occur via `SessionOptions` on inference session, prior to any work done by the native library. Hence why `OnnxRuntime` has an `executeWithOptions` method.
* There is a small UI example that use [Facial Emotion Recognition](https://github.com/onnx/models/tree/main/validated/vision/body_analysis/emotion_ferplus) model on how use the CoreML provider under `test` directory, inside `oracle.code.onnx.fer`.
* The `README.md` file contains instructions on how to run the example and how to regenerate the bindings for the provider. Instructions and scripts to regenerate the bindings were provided for macOS users only since CoreML is an Apple software.

Work done in this PR is done together with Lize Raes.

-------------

Commit messages:
 - Another attempt to remove all tabs.
 - Reconvert modifier of Session createSession(Arena arena, byte[] model, SessionOptions options) to private.
 - Attempt to remove whitespaces in OnnxProtoBuilder.java
 - Add support for other onnx providers (CoreML)

Changes: https://git.openjdk.org/babylon/pull/618/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=618&range=00
  Stats: 1084 lines in 14 files changed: 1075 ins; 4 del; 5 mod
  Patch: https://git.openjdk.org/babylon/pull/618.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/618/head:pull/618

PR: https://git.openjdk.org/babylon/pull/618

From duke at openjdk.org  Fri Oct 17 12:53:44 2025
From: duke at openjdk.org (Ana Maria Mihalceanu)
Date: Fri, 17 Oct 2025 12:53:44 GMT
Subject: [code-reflection] RFR: Add support for other onnx providers
 (CoreML) [v2]
In-Reply-To: <e7Q32XrF2lzb5NdbJDAbklKlodvEyxR9Ds38h_GSDvA=.96e74c03-8271-461d-bb73-638d83d07ea0@github.com>
References: <e7Q32XrF2lzb5NdbJDAbklKlodvEyxR9Ds38h_GSDvA=.96e74c03-8271-461d-bb73-638d83d07ea0@github.com>
Message-ID: <zRusgia_NhzaeDkL6yHylgNoQqoXqW_UpujwB7slSo0=.280afd0b-199b-4bcd-bae4-c8d3da47ddb3@github.com>

> This PR contains a series of changes meant to integrate other ONNX providers.
> 
> * It contains generated Java bindings for CoreML provider: `oracle.code.onnx.foreign.coreml_provider_factory_h`
> * Configurations of providers are present in `oracle.code.onnx.provider`.
> * Provider configuration should occur via `SessionOptions` on inference session, prior to any work done by the native library. Hence why `OnnxRuntime` has an `executeWithOptions` method.
> * There is a small UI example that use [Facial Emotion Recognition](https://github.com/onnx/models/tree/main/validated/vision/body_analysis/emotion_ferplus) model on how use the CoreML provider under `test` directory, inside `oracle.code.onnx.fer`.
> * The `README.md` file contains instructions on how to run the example and how to regenerate the bindings for the provider. Instructions and scripts to regenerate the bindings were provided for macOS users only since CoreML is an Apple software.
> 
> Work done in this PR is done together with Lize Raes.

Ana Maria Mihalceanu has updated the pull request incrementally with one additional commit since the last revision:

  Update full name

-------------

Changes:
  - all: https://git.openjdk.org/babylon/pull/618/files
  - new: https://git.openjdk.org/babylon/pull/618/files/a76a9ef2..df32b1eb

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=babylon&pr=618&range=01
 - incr: https://webrevs.openjdk.org/?repo=babylon&pr=618&range=00-01

  Stats: 0 lines in 0 files changed: 0 ins; 0 del; 0 mod
  Patch: https://git.openjdk.org/babylon/pull/618.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/618/head:pull/618

PR: https://git.openjdk.org/babylon/pull/618

From asotona at openjdk.org  Fri Oct 17 13:04:37 2025
From: asotona at openjdk.org (Adam Sotona)
Date: Fri, 17 Oct 2025 13:04:37 GMT
Subject: [code-reflection] Integrated: OnnxProtoBuilder fixes
In-Reply-To: <WFdUPj68hK2RfeloCpVSTCOu_PurIelHv-EE5S1jW8c=.559ab9e8-d555-4759-b2ba-7e6ac26fce77@github.com>
References: <WFdUPj68hK2RfeloCpVSTCOu_PurIelHv-EE5S1jW8c=.559ab9e8-d555-4759-b2ba-7e6ac26fce77@github.com>
Message-ID: <e2evZNKtd23lsR1DV6BLi1QfNIjTi62s7A_22rn54w8=.c3934a76-d254-4c8d-b228-7e333323787b@github.com>

On Thu, 16 Oct 2025 13:37:13 GMT, Adam Sotona <asotona at openjdk.org> wrote:

> - fixes ONNX graph output name of returned tuple element
> - fixes String AttributeProto value

This pull request has now been integrated.

Changeset: 02d3c681
Author:    Adam Sotona <asotona at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/02d3c681cdc21976cdbe293be8f49aaea352d706
Stats:     7 lines in 1 file changed: 6 ins; 1 del; 0 mod

OnnxProtoBuilder fixes

-------------

PR: https://git.openjdk.org/babylon/pull/614

From asotona at openjdk.org  Fri Oct 17 13:06:34 2025
From: asotona at openjdk.org (Adam Sotona)
Date: Fri, 17 Oct 2025 13:06:34 GMT
Subject: git: openjdk/babylon: code-reflection: OnnxProtoBuilder fixes
Message-ID: <f026dd4d-8fb7-4b45-bf13-58a517348800@openjdk.org>

Changeset: 02d3c681
Branch: code-reflection
Author:    Adam Sotona <asotona at openjdk.org>
Date:      2025-10-17 13:01:55 +0000
URL:       https://git.openjdk.org/babylon/commit/02d3c681cdc21976cdbe293be8f49aaea352d706

OnnxProtoBuilder fixes

! cr-examples/onnx/src/main/java/oracle/code/onnx/OnnxProtoBuilder.java


From asotona at openjdk.org  Fri Oct 17 13:30:16 2025
From: asotona at openjdk.org (Adam Sotona)
Date: Fri, 17 Oct 2025 13:30:16 GMT
Subject: [code-reflection] RFR: Add support for other onnx providers
 (CoreML) [v2]
In-Reply-To: <zRusgia_NhzaeDkL6yHylgNoQqoXqW_UpujwB7slSo0=.280afd0b-199b-4bcd-bae4-c8d3da47ddb3@github.com>
References: <e7Q32XrF2lzb5NdbJDAbklKlodvEyxR9Ds38h_GSDvA=.96e74c03-8271-461d-bb73-638d83d07ea0@github.com>
 <zRusgia_NhzaeDkL6yHylgNoQqoXqW_UpujwB7slSo0=.280afd0b-199b-4bcd-bae4-c8d3da47ddb3@github.com>
Message-ID: <VTW-BTJXZ1pqsJVoUOd68IG545h-Oq3wQSX0Sgy55mI=.688d421b-aa91-4dd4-ad02-8883bcdeffdf@github.com>

On Fri, 17 Oct 2025 12:53:44 GMT, Ana Maria Mihalceanu <duke at openjdk.org> wrote:

>> This PR contains a series of changes meant to integrate other ONNX providers.
>> 
>> * It contains generated Java bindings for CoreML provider: `oracle.code.onnx.foreign.coreml_provider_factory_h`
>> * Configurations of providers are present in `oracle.code.onnx.provider`.
>> * Provider configuration should occur via `SessionOptions` on inference session, prior to any work done by the native library. Hence why `OnnxRuntime` has an `executeWithOptions` method.
>> * There is a small UI example that use [Facial Emotion Recognition](https://github.com/onnx/models/tree/main/validated/vision/body_analysis/emotion_ferplus) model on how use the CoreML provider under `test` directory, inside `oracle.code.onnx.fer`.
>> * The `README.md` file contains instructions on how to run the example and how to regenerate the bindings for the provider. Instructions and scripts to regenerate the bindings were provided for macOS users only since CoreML is an Apple software.
>> 
>> Work done in this PR is done together with Lize Raes.
>
> Ana Maria Mihalceanu has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Update full name

Generally it looks great, thank you for the contribution!

cr-examples/onnx/setup.sh line 1:

> 1: #!/bin/bash

Files related to (re)generation of the library could be better stored under opgen folder/project.

cr-examples/onnx/src/main/java/oracle/code/onnx/OnnxRuntime.java line 244:

> 242:     }
> 243: 
> 244:     public static <T> T executeWithOptions(Arena arena, MethodHandles.Lookup l, OnnxFunction<T> codeLambda, SessionOptions options) {

I would recommend to make this method as an override of execute.
Original execute duplicates the code, so it can delegate with no options instead.

cr-examples/onnx/src/test/resources/oracle/code/onnx/fer/emotion-ferplus-8.onnx.data line 1:

> 1: ?O:=;z?<?2I;F??i?=??<?&?~?;.s?E J?b2???Q=??=??/?3???FY?

We should double-check if it is OK to store 33M binary directly into the Git repo.

-------------

PR Review: https://git.openjdk.org/babylon/pull/618#pullrequestreview-3350323741
PR Review Comment: https://git.openjdk.org/babylon/pull/618#discussion_r2440025962
PR Review Comment: https://git.openjdk.org/babylon/pull/618#discussion_r2439993076
PR Review Comment: https://git.openjdk.org/babylon/pull/618#discussion_r2440029296

From psandoz at openjdk.org  Fri Oct 17 14:38:43 2025
From: psandoz at openjdk.org (Paul Sandoz)
Date: Fri, 17 Oct 2025 14:38:43 GMT
Subject: [code-reflection] RFR: Add support for other onnx providers
 (CoreML) [v2]
In-Reply-To: <VTW-BTJXZ1pqsJVoUOd68IG545h-Oq3wQSX0Sgy55mI=.688d421b-aa91-4dd4-ad02-8883bcdeffdf@github.com>
References: <e7Q32XrF2lzb5NdbJDAbklKlodvEyxR9Ds38h_GSDvA=.96e74c03-8271-461d-bb73-638d83d07ea0@github.com>
 <zRusgia_NhzaeDkL6yHylgNoQqoXqW_UpujwB7slSo0=.280afd0b-199b-4bcd-bae4-c8d3da47ddb3@github.com>
 <VTW-BTJXZ1pqsJVoUOd68IG545h-Oq3wQSX0Sgy55mI=.688d421b-aa91-4dd4-ad02-8883bcdeffdf@github.com>
Message-ID: <xr6MylZ8KC1-h3jxwAeXVnsgOgAOqQ9muIBNtftBk-M=.be4fce34-c2a7-4212-9021-06b6521819f6@github.com>

On Fri, 17 Oct 2025 13:26:17 GMT, Adam Sotona <asotona at openjdk.org> wrote:

>> Ana Maria Mihalceanu has updated the pull request incrementally with one additional commit since the last revision:
>> 
>>   Update full name
>
> cr-examples/onnx/src/test/resources/oracle/code/onnx/fer/emotion-ferplus-8.onnx.data line 1:
> 
>> 1: ?O:=;z?<?2I;F??i?=??<?&?~?;.s?E J?b2???Q=??=??/?3???FY?
> 
> We should double-check if it is OK to store 33M binary directly into the Git repo.

It's generally not ok. We should avoid committing any binary files, large or small, and instead provide instructions or a script to download such files. Especially, since this is in the Babylon repo, so everyone cloning this repo will be impacted.

-------------

PR Review Comment: https://git.openjdk.org/babylon/pull/618#discussion_r2440238252

From duke at openjdk.org  Fri Oct 17 14:58:35 2025
From: duke at openjdk.org (Ana Maria Mihalceanu)
Date: Fri, 17 Oct 2025 14:58:35 GMT
Subject: [code-reflection] RFR: Add support for other onnx providers
 (CoreML) [v2]
In-Reply-To: <xr6MylZ8KC1-h3jxwAeXVnsgOgAOqQ9muIBNtftBk-M=.be4fce34-c2a7-4212-9021-06b6521819f6@github.com>
References: <e7Q32XrF2lzb5NdbJDAbklKlodvEyxR9Ds38h_GSDvA=.96e74c03-8271-461d-bb73-638d83d07ea0@github.com>
 <zRusgia_NhzaeDkL6yHylgNoQqoXqW_UpujwB7slSo0=.280afd0b-199b-4bcd-bae4-c8d3da47ddb3@github.com>
 <VTW-BTJXZ1pqsJVoUOd68IG545h-Oq3wQSX0Sgy55mI=.688d421b-aa91-4dd4-ad02-8883bcdeffdf@github.com>
 <xr6MylZ8KC1-h3jxwAeXVnsgOgAOqQ9muIBNtftBk-M=.be4fce34-c2a7-4212-9021-06b6521819f6@github.com>
Message-ID: <YhS-Av3Q6eGNF3t9E2QCsIoKwzZfTgP-8_z9ya81LrA=.e028c30b-137c-437d-b371-1cf1233991ce@github.com>

On Fri, 17 Oct 2025 14:33:40 GMT, Paul Sandoz <psandoz at openjdk.org> wrote:

>> cr-examples/onnx/src/test/resources/oracle/code/onnx/fer/emotion-ferplus-8.onnx.data line 1:
>> 
>>> 1: ?O:=;z?<?2I;F??i?=??<?&?~?;.s?E J?b2???Q=??=??/?3???FY?
>> 
>> We should double-check if it is OK to store 33M binary directly into the Git repo.
>
> It's generally not ok. We should avoid committing any binary files, large or small, and instead provide instructions or a script to download such files. Especially, since this is in the Babylon repo, so everyone cloning this repo will be impacted.

I followed the example from: 
https://github.com/openjdk/babylon/tree/code-reflection/cr-examples/onnx/src/test/resources/oracle/code/onnx/mnist.

-------------

PR Review Comment: https://git.openjdk.org/babylon/pull/618#discussion_r2440300513

From duke at openjdk.org  Fri Oct 17 15:11:54 2025
From: duke at openjdk.org (Adam Pocock)
Date: Fri, 17 Oct 2025 15:11:54 GMT
Subject: [code-reflection] RFR: Add support for other onnx providers
 (CoreML) [v2]
In-Reply-To: <zRusgia_NhzaeDkL6yHylgNoQqoXqW_UpujwB7slSo0=.280afd0b-199b-4bcd-bae4-c8d3da47ddb3@github.com>
References: <e7Q32XrF2lzb5NdbJDAbklKlodvEyxR9Ds38h_GSDvA=.96e74c03-8271-461d-bb73-638d83d07ea0@github.com>
 <zRusgia_NhzaeDkL6yHylgNoQqoXqW_UpujwB7slSo0=.280afd0b-199b-4bcd-bae4-c8d3da47ddb3@github.com>
Message-ID: <X_FY8ADx0fJTTzqCtoFwMfHom80N7Nuvsb101JWNMVc=.c0cc21e9-9886-4c30-9540-57211cc7fdb4@github.com>

On Fri, 17 Oct 2025 12:53:44 GMT, Ana Maria Mihalceanu <duke at openjdk.org> wrote:

>> This PR contains a series of changes meant to integrate other ONNX providers.
>> 
>> * It contains generated Java bindings for CoreML provider: `oracle.code.onnx.foreign.coreml_provider_factory_h`
>> * Configurations of providers are present in `oracle.code.onnx.provider`.
>> * Provider configuration should occur via `SessionOptions` on inference session, prior to any work done by the native library. Hence why `OnnxRuntime` has an `executeWithOptions` method.
>> * There is a small UI example that use [Facial Emotion Recognition](https://github.com/onnx/models/tree/main/validated/vision/body_analysis/emotion_ferplus) model on how use the CoreML provider under `test` directory, inside `oracle.code.onnx.fer`.
>> * The `README.md` file contains instructions on how to run the example and how to regenerate the bindings for the provider. Instructions and scripts to regenerate the bindings were provided for macOS users only since CoreML is an Apple software.
>> 
>> Work done in this PR is done together with Lize Raes.
>
> Ana Maria Mihalceanu has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Update full name

cr-examples/onnx/src/main/java/oracle/code/onnx/provider/CoreMLProvider.java line 42:

> 40:         try {
> 41:             int coremlFlag = COREML_FLAG_USE_CPU_AND_GPU();
> 42:             var status = OrtSessionOptionsAppendExecutionProvider_CoreML(sessionOptionsAddress, coremlFlag);

The [ML_PROGRAM flag](https://github.com/microsoft/onnxruntime/blob/main/include/onnxruntime/core/providers/coreml/coreml_provider_factory.h#L32) is probably useful to turn on here as it expands the set of things that ORT can accelerate with CoreML (e.g. I can't get it to do anything to [sd4j](https://github.com/oracle/sd4j) without it, but it does help a bit with it, or at least it did when I tested it in 2023). ML programs are the more modern way of accelerating things with CoreML, but they might need slightly higher OS versions.

-------------

PR Review Comment: https://git.openjdk.org/babylon/pull/618#discussion_r2440337171

From asotona at openjdk.org  Fri Oct 17 15:21:39 2025
From: asotona at openjdk.org (Adam Sotona)
Date: Fri, 17 Oct 2025 15:21:39 GMT
Subject: [code-reflection] RFR: Add support for other onnx providers
 (CoreML) [v2]
In-Reply-To: <YhS-Av3Q6eGNF3t9E2QCsIoKwzZfTgP-8_z9ya81LrA=.e028c30b-137c-437d-b371-1cf1233991ce@github.com>
References: <e7Q32XrF2lzb5NdbJDAbklKlodvEyxR9Ds38h_GSDvA=.96e74c03-8271-461d-bb73-638d83d07ea0@github.com>
 <zRusgia_NhzaeDkL6yHylgNoQqoXqW_UpujwB7slSo0=.280afd0b-199b-4bcd-bae4-c8d3da47ddb3@github.com>
 <VTW-BTJXZ1pqsJVoUOd68IG545h-Oq3wQSX0Sgy55mI=.688d421b-aa91-4dd4-ad02-8883bcdeffdf@github.com>
 <xr6MylZ8KC1-h3jxwAeXVnsgOgAOqQ9muIBNtftBk-M=.be4fce34-c2a7-4212-9021-06b6521819f6@github.com>
 <YhS-Av3Q6eGNF3t9E2QCsIoKwzZfTgP-8_z9ya81LrA=.e028c30b-137c-437d-b371-1cf1233991ce@github.com>
Message-ID: <4uwwf8yd7LpVYO94At_q1B0s5z2biEfSbW1DFibMWM4=.f87ff0d8-44d7-489c-8151-8baf64d89c7b@github.com>

On Fri, 17 Oct 2025 14:55:36 GMT, Ana Maria Mihalceanu <duke at openjdk.org> wrote:

>> It's generally not ok. We should avoid committing any binary files, large or small, and instead provide instructions or a script to download such files. Especially, since this is in the Babylon repo, so everyone cloning this repo will be impacted.
>
> I followed the example from: 
> https://github.com/openjdk/babylon/tree/code-reflection/cr-examples/onnx/src/test/resources/oracle/code/onnx/mnist.

Right, we needed one model with weights to be a part of the build tests and MNIST weights are just ~170K.

-------------

PR Review Comment: https://git.openjdk.org/babylon/pull/618#discussion_r2440365413

From duke at openjdk.org  Fri Oct 17 17:06:12 2025
From: duke at openjdk.org (Ruby Chen)
Date: Fri, 17 Oct 2025 17:06:12 GMT
Subject: [code-reflection] RFR: Add kernel entrypoint to ModuleOp [v3]
In-Reply-To: <UkyuDt6Ei-D4V2Ly2ukTEV-Sx9qLIbN6ih3KSJ1BHYY=.45a19d55-ee90-46ec-b41c-28ec98117c5b@github.com>
References: <UkyuDt6Ei-D4V2Ly2ukTEV-Sx9qLIbN6ih3KSJ1BHYY=.45a19d55-ee90-46ec-b41c-28ec98117c5b@github.com>
Message-ID: <Q6DR9PorC9tLoEsR51EO4gH1CcMQXtV99GbbwoIsUjM=.595b2c92-d837-4490-ae34-7e06bbebc37f@github.com>

> Include the kernel (and compute) entrypoints in their respective ModuleOps.

Ruby Chen has updated the pull request with a new target base due to a merge or a rebase. The pull request now contains three commits:

 - Merge branch 'code-reflection' into cleanup
 - Merge branch 'code-reflection' into cleanup
 - Add kernel entrypoint to ModuleOp

-------------

Changes: https://git.openjdk.org/babylon/pull/605/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=605&range=02
  Stats: 11 lines in 2 files changed: 4 ins; 5 del; 2 mod
  Patch: https://git.openjdk.org/babylon/pull/605.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/605/head:pull/605

PR: https://git.openjdk.org/babylon/pull/605

From duke at openjdk.org  Fri Oct 17 17:10:22 2025
From: duke at openjdk.org (Ruby Chen)
Date: Fri, 17 Oct 2025 17:10:22 GMT
Subject: [code-reflection] RFR: Update buffer tagging with arrayview
 support [v2]
In-Reply-To: <Q9zdEEQTILTtqvItQQ26aC5TCXabT0YwpnlNFFY-fZg=.e0f9b877-1021-4d2f-b8e0-b687928a61dc@github.com>
References: <Q9zdEEQTILTtqvItQQ26aC5TCXabT0YwpnlNFFY-fZg=.e0f9b877-1021-4d2f-b8e0-b687928a61dc@github.com>
Message-ID: <1jdULGHfHB4nCUXUpVPryHcmGqIfOoAX82UivLaw8gc=.2010ead2-2d31-4acd-9aa7-338abe739c73@github.com>

> Add buffer tagging for array views, and also use `elements()` instead of `traverse()`.

Ruby Chen has updated the pull request with a new target base due to a merge or a rebase. The pull request now contains two commits:

 - Merge branch 'code-reflection' into module-op-support
 - Update buffer tagging with arrayview support

-------------

Changes: https://git.openjdk.org/babylon/pull/610/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=610&range=01
  Stats: 39 lines in 2 files changed: 17 ins; 3 del; 19 mod
  Patch: https://git.openjdk.org/babylon/pull/610.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/610/head:pull/610

PR: https://git.openjdk.org/babylon/pull/610

From gfrost at openjdk.org  Sun Oct 19 13:56:11 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Sun, 19 Oct 2025 13:56:11 GMT
Subject: git: openjdk/babylon: code-reflection: Added simple map transformer to
 OpTk + Config convenience methods
Message-ID: <126bb535-59a7-4fb0-8d2b-32d0f0516b67@openjdk.org>

Changeset: d509d3a0
Branch: code-reflection
Author:    Gary Frost <gfrost at openjdk.org>
Date:      2025-10-19 13:55:21 +0000
URL:       https://git.openjdk.org/babylon/commit/d509d3a0f0283680668f6ecb80a229ae6da5f24a

Added simple map transformer to OpTk + Config convenience methods

! hat/backends/ffi/cuda/src/main/java/hat/backend/ffi/CudaBackend.java
! hat/backends/ffi/opencl/src/main/java/hat/backend/ffi/OpenCLBackend.java
! hat/backends/ffi/shared/src/main/java/hat/backend/ffi/C99FFIBackend.java
! hat/backends/ffi/shared/src/main/java/hat/backend/ffi/FFIBackend.java
! hat/core/src/main/java/hat/Accelerator.java
! hat/core/src/main/java/hat/Config.java
! hat/core/src/main/java/hat/optools/OpTk.java
! hat/core/src/main/java/hat/phases/HATDialectifyBarrierPhase.java
! hat/core/src/main/java/hat/phases/HATDialectifyMemoryPhase.java
! hat/core/src/main/java/hat/phases/HATDialectifyThreadsPhase.java
! hat/core/src/main/java/hat/phases/HATDialectifyVSelectPhase.java
! hat/core/src/main/java/hat/phases/HATDialectifyVectorOpPhase.java
! hat/core/src/main/java/hat/phases/HATDialectifyVectorStorePhase.java
+ hat/examples/experiments/src/main/java/experiments/TransformState.java
! hat/examples/mandel/src/main/java/mandel/Main.java
! hat/examples/violajones/src/main/java/violajones/Main.java


From gfrost at openjdk.org  Sun Oct 19 13:59:10 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Sun, 19 Oct 2025 13:59:10 GMT
Subject: [code-reflection] Integrated: Added simple map transformer to OpTk +
 Config convenience methods
Message-ID: <9BWA-AV6SePiCK9E9KgbPf3aR0eYFoSYN5rAgnoJCaA=.e77b7589-f462-4077-affb-c3fb57c02d26@github.com>

Trying to minimize op.transform ceremony. 

Added a strawman simple mapping transformer to OpTk. 

Also added some convenience methods to Config.

-------------

Commit messages:
 - Added simple map transformer to OpTk + Config convenience methods

Changes: https://git.openjdk.org/babylon/pull/620/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=620&range=00
  Stats: 339 lines in 16 files changed: 288 ins; 10 del; 41 mod
  Patch: https://git.openjdk.org/babylon/pull/620.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/620/head:pull/620

PR: https://git.openjdk.org/babylon/pull/620

From gfrost at openjdk.org  Sun Oct 19 13:59:11 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Sun, 19 Oct 2025 13:59:11 GMT
Subject: [code-reflection] Integrated: Added simple map transformer to OpTk +
 Config convenience methods
In-Reply-To: <9BWA-AV6SePiCK9E9KgbPf3aR0eYFoSYN5rAgnoJCaA=.e77b7589-f462-4077-affb-c3fb57c02d26@github.com>
References: <9BWA-AV6SePiCK9E9KgbPf3aR0eYFoSYN5rAgnoJCaA=.e77b7589-f462-4077-affb-c3fb57c02d26@github.com>
Message-ID: <Iv4lWLqNFF1p65CqwAn-29mHbjkbn4xruspH2P_--nk=.24790e94-0ad4-4384-b519-c61044a62b2e@github.com>

On Sun, 19 Oct 2025 13:52:34 GMT, Gary Frost <gfrost at openjdk.org> wrote:

> Trying to minimize op.transform ceremony. 
> 
> Added a strawman simple mapping transformer to OpTk. 
> 
> Also added some convenience methods to Config.

This pull request has now been integrated.

Changeset: d509d3a0
Author:    Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/d509d3a0f0283680668f6ecb80a229ae6da5f24a
Stats:     339 lines in 16 files changed: 288 ins; 10 del; 41 mod

Added simple map transformer to OpTk + Config convenience methods

-------------

PR: https://git.openjdk.org/babylon/pull/620

From gfrost at openjdk.org  Mon Oct 20 14:10:16 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Mon, 20 Oct 2025 14:10:16 GMT
Subject: [code-reflection] Integrated: first pass at mapping HATOp phase
 Transformers to us  OpTk form
Message-ID: <E05m_15K5OSG-RXUE_RSIAigTstMuhpKLMs4S51BVAI=.e24b8ab8-5fe4-49e3-830f-835727ca658e@github.com>

After adding some helper classes (for simpletransforms and filtered transforms) I took a stab at refactoring the HatOp transformers. 

More work to do

-------------

Commit messages:
 - first pass at mapping HATOp phase Transformers to us  OpTk form

Changes: https://git.openjdk.org/babylon/pull/621/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=621&range=00
  Stats: 1580 lines in 34 files changed: 676 ins; 704 del; 200 mod
  Patch: https://git.openjdk.org/babylon/pull/621.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/621/head:pull/621

PR: https://git.openjdk.org/babylon/pull/621

From gfrost at openjdk.org  Mon Oct 20 14:10:18 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Mon, 20 Oct 2025 14:10:18 GMT
Subject: [code-reflection] Integrated: first pass at mapping HATOp phase
 Transformers to us  OpTk form
In-Reply-To: <E05m_15K5OSG-RXUE_RSIAigTstMuhpKLMs4S51BVAI=.e24b8ab8-5fe4-49e3-830f-835727ca658e@github.com>
References: <E05m_15K5OSG-RXUE_RSIAigTstMuhpKLMs4S51BVAI=.e24b8ab8-5fe4-49e3-830f-835727ca658e@github.com>
Message-ID: <DbCuOjbLnSFnfhyn8bnGBk86K3PB-ICgzSJ_GW6JmFU=.85824e03-6154-419d-86c1-76469d4bfeaa@github.com>

On Mon, 20 Oct 2025 14:04:37 GMT, Gary Frost <gfrost at openjdk.org> wrote:

> After adding some helper classes (for simpletransforms and filtered transforms) I took a stab at refactoring the HatOp transformers. 
> 
> More work to do

This pull request has now been integrated.

Changeset: 00e7c5c6
Author:    Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/00e7c5c6ad0417a580df7736d6086f15cf8cea9e
Stats:     1580 lines in 34 files changed: 676 ins; 704 del; 200 mod

first pass at mapping HATOp phase Transformers to us  OpTk form

-------------

PR: https://git.openjdk.org/babylon/pull/621

From gfrost at openjdk.org  Mon Oct 20 14:13:15 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Mon, 20 Oct 2025 14:13:15 GMT
Subject: git: openjdk/babylon: code-reflection: first pass at mapping HATOp
 phase Transformers to us  OpTk form
Message-ID: <6795c87d-22e4-45f1-b260-e3eac26e6d38@openjdk.org>

Changeset: 00e7c5c6
Branch: code-reflection
Author:    Gary Frost <gfrost at openjdk.org>
Date:      2025-10-20 14:07:10 +0000
URL:       https://git.openjdk.org/babylon/commit/00e7c5c6ad0417a580df7736d6086f15cf8cea9e

first pass at mapping HATOp phase Transformers to us  OpTk form

! hat/backends/ffi/cuda/src/main/java/hat/backend/ffi/CudaHATKernelBuilder.java
! hat/backends/ffi/opencl/src/main/java/hat/backend/ffi/OpenCLHATKernelBuilder.java
! hat/backends/ffi/shared/src/main/java/hat/backend/ffi/C99FFIBackend.java
! hat/backends/jextracted/opencl/src/main/java/hat/backend/jextracted/OpenCLHatKernelBuilder.java
! hat/core/src/main/java/hat/callgraph/KernelCallGraph.java
! hat/core/src/main/java/hat/codebuilders/BabylonOpBuilder.java
! hat/core/src/main/java/hat/codebuilders/HATCodeBuilderWithContext.java
! hat/core/src/main/java/hat/dialect/HATBarrierOp.java
! hat/core/src/main/java/hat/dialect/HATMemoryOp.java
! hat/core/src/main/java/hat/dialect/HATOp.java
- hat/core/src/main/java/hat/dialect/HATVSelectLoadOp.java
- hat/core/src/main/java/hat/dialect/HATVSelectStoreOp.java
! hat/core/src/main/java/hat/dialect/HATVectorBinaryOp.java
+ hat/core/src/main/java/hat/dialect/HATVectorSelectLoadOp.java
+ hat/core/src/main/java/hat/dialect/HATVectorSelectStoreOp.java
! hat/core/src/main/java/hat/dialect/HATVectorStoreView.java
! hat/core/src/main/java/hat/optools/OpTk.java
- hat/core/src/main/java/hat/phases/HATCompilationTier.java
+ hat/core/src/main/java/hat/phases/HATDialect.java
- hat/core/src/main/java/hat/phases/HATDialectAbstractPhase.java
! hat/core/src/main/java/hat/phases/HATDialectifyBarrierPhase.java
! hat/core/src/main/java/hat/phases/HATDialectifyMemoryPhase.java
+ hat/core/src/main/java/hat/phases/HATDialectifyMemoryPrivatePhase.java
+ hat/core/src/main/java/hat/phases/HATDialectifyMemorySharedPhase.java
- hat/core/src/main/java/hat/phases/HATDialectifyPhase.java
! hat/core/src/main/java/hat/phases/HATDialectifyThreadsPhase.java
! hat/core/src/main/java/hat/phases/HATDialectifyTier.java
- hat/core/src/main/java/hat/phases/HATDialectifyVSelectPhase.java
! hat/core/src/main/java/hat/phases/HATDialectifyVectorOpPhase.java
+ hat/core/src/main/java/hat/phases/HATDialectifyVectorSelectPhase.java
! hat/core/src/main/java/hat/phases/HATDialectifyVectorStorePhase.java
! hat/core/src/main/java/hat/phases/HATFinalDetectionPhase.java
- hat/core/src/main/java/hat/phases/HATPhase.java
! hat/tools/src/main/java/hat/tools/text/JavaHATCodeBuilder.java


From duke at openjdk.org  Mon Oct 20 14:36:40 2025
From: duke at openjdk.org (Ana Maria Mihalceanu)
Date: Mon, 20 Oct 2025 14:36:40 GMT
Subject: [code-reflection] RFR: Add support for other onnx providers
 (CoreML) [v3]
In-Reply-To: <e7Q32XrF2lzb5NdbJDAbklKlodvEyxR9Ds38h_GSDvA=.96e74c03-8271-461d-bb73-638d83d07ea0@github.com>
References: <e7Q32XrF2lzb5NdbJDAbklKlodvEyxR9Ds38h_GSDvA=.96e74c03-8271-461d-bb73-638d83d07ea0@github.com>
Message-ID: <fM4n8SZDFx90CBnXbZnUfvpFdn2o9crvKGWqRee_aZ8=.3e260241-25bf-4132-b30f-4f90106a7285@github.com>

> This PR contains a series of changes meant to integrate other ONNX providers.
> 
> * It contains generated Java bindings for CoreML provider: `oracle.code.onnx.foreign.coreml_provider_factory_h`
> * Configurations of providers are present in `oracle.code.onnx.provider`.
> * Provider configuration should occur via `SessionOptions` on inference session, prior to any work done by the native library. Hence why `OnnxRuntime` has an `executeWithOptions` method.
> * There is a small UI example that use [Facial Emotion Recognition](https://github.com/onnx/models/tree/main/validated/vision/body_analysis/emotion_ferplus) model on how use the CoreML provider under `test` directory, inside `oracle.code.onnx.fer`.
> * The `README.md` file contains instructions on how to run the example and how to regenerate the bindings for the provider. Instructions and scripts to regenerate the bindings were provided for macOS users only since CoreML is an Apple software.
> 
> Work done in this PR is done together with [Lize Raes](https://github.com/LizeRaes).

Ana Maria Mihalceanu has updated the pull request incrementally with two additional commits since the last revision:

 - Revert logging level to error.
 - Refactor OnnxRuntime and CoreMLProvider to address PR comments.

-------------

Changes:
  - all: https://git.openjdk.org/babylon/pull/618/files
  - new: https://git.openjdk.org/babylon/pull/618/files/df32b1eb..293d6e77

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=babylon&pr=618&range=02
 - incr: https://webrevs.openjdk.org/?repo=babylon&pr=618&range=01-02

  Stats: 129 lines in 10 files changed: 44 ins; 69 del; 16 mod
  Patch: https://git.openjdk.org/babylon/pull/618.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/618/head:pull/618

PR: https://git.openjdk.org/babylon/pull/618

From duke at openjdk.org  Mon Oct 20 14:36:43 2025
From: duke at openjdk.org (Ana Maria Mihalceanu)
Date: Mon, 20 Oct 2025 14:36:43 GMT
Subject: [code-reflection] RFR: Add support for other onnx providers
 (CoreML) [v2]
In-Reply-To: <VTW-BTJXZ1pqsJVoUOd68IG545h-Oq3wQSX0Sgy55mI=.688d421b-aa91-4dd4-ad02-8883bcdeffdf@github.com>
References: <e7Q32XrF2lzb5NdbJDAbklKlodvEyxR9Ds38h_GSDvA=.96e74c03-8271-461d-bb73-638d83d07ea0@github.com>
 <zRusgia_NhzaeDkL6yHylgNoQqoXqW_UpujwB7slSo0=.280afd0b-199b-4bcd-bae4-c8d3da47ddb3@github.com>
 <VTW-BTJXZ1pqsJVoUOd68IG545h-Oq3wQSX0Sgy55mI=.688d421b-aa91-4dd4-ad02-8883bcdeffdf@github.com>
Message-ID: <FJcT6LVqJ2iE3C4oePtKTDatHZpDdRxqFQFqjlgJPeY=.913803cc-b469-4133-b3cc-cdd8009f4309@github.com>

On Fri, 17 Oct 2025 13:25:07 GMT, Adam Sotona <asotona at openjdk.org> wrote:

>> Ana Maria Mihalceanu has updated the pull request incrementally with one additional commit since the last revision:
>> 
>>   Update full name
>
> cr-examples/onnx/opgen/setup.sh line 1:
> 
>> (failed to retrieve contents of file, check the PR for context)
> Files related to (re)generation of the library could be better stored under opgen folder/project.

I moved the files to `opgen` and changed the instructions accordingly.

> cr-examples/onnx/src/main/java/oracle/code/onnx/OnnxRuntime.java line 244:
> 
>> 242:     }
>> 243: 
>> 244:     public static <T> T executeWithOptions(Arena arena, MethodHandles.Lookup l, OnnxFunction<T> codeLambda, SessionOptions options) {
> 
> I would recommend to make this method as an override of execute.
> Original execute duplicates the code, so it can delegate with no options instead.

Adam, can you please check the refactoring done for execute?

-------------

PR Review Comment: https://git.openjdk.org/babylon/pull/618#discussion_r2445195367
PR Review Comment: https://git.openjdk.org/babylon/pull/618#discussion_r2445204160

From duke at openjdk.org  Mon Oct 20 14:58:06 2025
From: duke at openjdk.org (Ana Maria Mihalceanu)
Date: Mon, 20 Oct 2025 14:58:06 GMT
Subject: [code-reflection] RFR: Add support for other onnx providers
 (CoreML) [v4]
In-Reply-To: <e7Q32XrF2lzb5NdbJDAbklKlodvEyxR9Ds38h_GSDvA=.96e74c03-8271-461d-bb73-638d83d07ea0@github.com>
References: <e7Q32XrF2lzb5NdbJDAbklKlodvEyxR9Ds38h_GSDvA=.96e74c03-8271-461d-bb73-638d83d07ea0@github.com>
Message-ID: <Zw-wfkY5p5gynk8_ymQa6JKFJRNVKu_rXfKR6dN5xXw=.ef55f980-a300-4cad-8c98-27e542690162@github.com>

> This PR contains a series of changes meant to integrate other ONNX providers.
> 
> * It contains generated Java bindings for CoreML provider: `oracle.code.onnx.foreign.coreml_provider_factory_h`
> * Configurations of providers are present in `oracle.code.onnx.provider`.
> * Provider configuration should occur via `SessionOptions` on inference session, prior to any work done by the native library. Hence why `OnnxRuntime` has an `execute` that gets `SessionOptions` too.
> * There is a small UI example that use [Facial Emotion Recognition](https://github.com/onnx/models/tree/main/validated/vision/body_analysis/emotion_ferplus) model on how use the CoreML provider under `test` directory, inside `oracle.code.onnx.fer`.
> * The `README.md` file contains instructions on how to run the example and how to regenerate the bindings for the provider. Instructions and scripts to regenerate the bindings were provided for macOS users only since CoreML is an Apple software.
> 
> Work done in this PR is done together with [Lize Raes](https://github.com/LizeRaes).

Ana Maria Mihalceanu has updated the pull request incrementally with one additional commit since the last revision:

  Remove tabs and convert to whitespaces only.

-------------

Changes:
  - all: https://git.openjdk.org/babylon/pull/618/files
  - new: https://git.openjdk.org/babylon/pull/618/files/293d6e77..db73b054

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=babylon&pr=618&range=03
 - incr: https://webrevs.openjdk.org/?repo=babylon&pr=618&range=02-03

  Stats: 4 lines in 1 file changed: 0 ins; 0 del; 4 mod
  Patch: https://git.openjdk.org/babylon/pull/618.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/618/head:pull/618

PR: https://git.openjdk.org/babylon/pull/618

From duke at openjdk.org  Mon Oct 20 14:58:12 2025
From: duke at openjdk.org (Ana Maria Mihalceanu)
Date: Mon, 20 Oct 2025 14:58:12 GMT
Subject: [code-reflection] RFR: Add support for other onnx providers
 (CoreML) [v2]
In-Reply-To: <4uwwf8yd7LpVYO94At_q1B0s5z2biEfSbW1DFibMWM4=.f87ff0d8-44d7-489c-8151-8baf64d89c7b@github.com>
References: <e7Q32XrF2lzb5NdbJDAbklKlodvEyxR9Ds38h_GSDvA=.96e74c03-8271-461d-bb73-638d83d07ea0@github.com>
 <zRusgia_NhzaeDkL6yHylgNoQqoXqW_UpujwB7slSo0=.280afd0b-199b-4bcd-bae4-c8d3da47ddb3@github.com>
 <VTW-BTJXZ1pqsJVoUOd68IG545h-Oq3wQSX0Sgy55mI=.688d421b-aa91-4dd4-ad02-8883bcdeffdf@github.com>
 <xr6MylZ8KC1-h3jxwAeXVnsgOgAOqQ9muIBNtftBk-M=.be4fce34-c2a7-4212-9021-06b6521819f6@github.com>
 <YhS-Av3Q6eGNF3t9E2QCsIoKwzZfTgP-8_z9ya81LrA=.e028c30b-137c-437d-b371-1cf1233991ce@github.com>
 <4uwwf8yd7LpVYO94At_q1B0s5z2biEfSbW1DFibMWM4=.f87ff0d8-44d7-489c-8151-8baf64d89c7b@github.com>
Message-ID: <MRbznih0WRiBmNncTO7SrpslEF8sIyhKyLWvBAX4ETI=.25e65e44-3239-438f-8c00-d12f18961ab0@github.com>

On Fri, 17 Oct 2025 15:19:23 GMT, Adam Sotona <asotona at openjdk.org> wrote:

>> I followed the example from: 
>> https://github.com/openjdk/babylon/tree/code-reflection/cr-examples/onnx/src/test/resources/oracle/code/onnx/mnist.
>
> Right, we needed one model with weights to be a part of the build tests and MNIST weights are just ~170K.

I removed the file with weights and I updated the README instructions to have it downloaded from https://github.com/ammbra/fer-model-weights/raw/refs/heads/main/emotion-ferplus-8.onnx.data.

My other ideas involved having a ? script or publishing to Hugging Face.

-------------

PR Review Comment: https://git.openjdk.org/babylon/pull/618#discussion_r2445249351

From duke at openjdk.org  Mon Oct 20 14:58:10 2025
From: duke at openjdk.org (Ana Maria Mihalceanu)
Date: Mon, 20 Oct 2025 14:58:10 GMT
Subject: [code-reflection] RFR: Add support for other onnx providers
 (CoreML) [v2]
In-Reply-To: <X_FY8ADx0fJTTzqCtoFwMfHom80N7Nuvsb101JWNMVc=.c0cc21e9-9886-4c30-9540-57211cc7fdb4@github.com>
References: <e7Q32XrF2lzb5NdbJDAbklKlodvEyxR9Ds38h_GSDvA=.96e74c03-8271-461d-bb73-638d83d07ea0@github.com>
 <zRusgia_NhzaeDkL6yHylgNoQqoXqW_UpujwB7slSo0=.280afd0b-199b-4bcd-bae4-c8d3da47ddb3@github.com>
 <X_FY8ADx0fJTTzqCtoFwMfHom80N7Nuvsb101JWNMVc=.c0cc21e9-9886-4c30-9540-57211cc7fdb4@github.com>
Message-ID: <7lZDtQS9Ynr55gcMgi9siVuTgXj2xTbjYVbk8_49BWg=.afe181c5-bfe9-4b04-990c-5b0431972fc6@github.com>

On Fri, 17 Oct 2025 15:08:47 GMT, Adam Pocock <duke at openjdk.org> wrote:

>> Ana Maria Mihalceanu has updated the pull request incrementally with one additional commit since the last revision:
>> 
>>   Update full name
>
> cr-examples/onnx/src/main/java/oracle/code/onnx/provider/CoreMLProvider.java line 42:
> 
>> 40:         try {
>> 41:             int coremlFlag = COREML_FLAG_USE_CPU_AND_GPU();
>> 42:             var status = OrtSessionOptionsAppendExecutionProvider_CoreML(sessionOptionsAddress, coremlFlag);
> 
> The [ML_PROGRAM flag](https://github.com/microsoft/onnxruntime/blob/main/include/onnxruntime/core/providers/coreml/coreml_provider_factory.h#L32) is probably useful to turn on here as it expands the set of things that ORT can accelerate with CoreML (e.g. I can't get it to do anything to [sd4j](https://github.com/oracle/sd4j) without it, but it does help a bit with it, or at least it did when I tested it in 2023). ML programs are the more modern way of accelerating things with CoreML, but they might need slightly higher OS versions.

That's a very good point. The flag should not be hardcoded so I made it configurable.

-------------

PR Review Comment: https://git.openjdk.org/babylon/pull/618#discussion_r2445262034

From duke at openjdk.org  Mon Oct 20 15:27:38 2025
From: duke at openjdk.org (Adam Pocock)
Date: Mon, 20 Oct 2025 15:27:38 GMT
Subject: [code-reflection] RFR: Add support for other onnx providers
 (CoreML) [v4]
In-Reply-To: <Zw-wfkY5p5gynk8_ymQa6JKFJRNVKu_rXfKR6dN5xXw=.ef55f980-a300-4cad-8c98-27e542690162@github.com>
References: <e7Q32XrF2lzb5NdbJDAbklKlodvEyxR9Ds38h_GSDvA=.96e74c03-8271-461d-bb73-638d83d07ea0@github.com>
 <Zw-wfkY5p5gynk8_ymQa6JKFJRNVKu_rXfKR6dN5xXw=.ef55f980-a300-4cad-8c98-27e542690162@github.com>
Message-ID: <cZmDplYevL0j-a9r0oazH7Y58hajzfPnA4LXacNyfrE=.4194413e-0557-4ec8-ac87-5620e10f58a5@github.com>

On Mon, 20 Oct 2025 14:58:06 GMT, Ana Maria Mihalceanu <duke at openjdk.org> wrote:

>> This PR contains a series of changes meant to integrate other ONNX providers.
>> 
>> * It contains generated Java bindings for CoreML provider: `oracle.code.onnx.foreign.coreml_provider_factory_h`
>> * Configurations of providers are present in `oracle.code.onnx.provider`.
>> * Provider configuration should occur via `SessionOptions` on inference session, prior to any work done by the native library. Hence why `OnnxRuntime` has an `execute` that gets `SessionOptions` too.
>> * There is a small UI example that use [Facial Emotion Recognition](https://github.com/onnx/models/tree/main/validated/vision/body_analysis/emotion_ferplus) model on how use the CoreML provider under `test` directory, inside `oracle.code.onnx.fer`.
>> * The `README.md` file contains instructions on how to run the example and how to regenerate the bindings for the provider. Instructions and scripts to regenerate the bindings were provided for macOS users only since CoreML is an Apple software.
>> 
>> Work done in this PR is done together with [Lize Raes](https://github.com/LizeRaes).
>
> Ana Maria Mihalceanu has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Remove tabs and convert to whitespaces only.

cr-examples/onnx/src/main/java/oracle/code/onnx/provider/CoreMLProvider.java line 38:

> 36:     private int flag;
> 37: 
> 38:     public CoreMLProvider(int flag) {

This is a small thing, but I'd probably call this `flags` as it's really a set of bits, so you can or the flags together to get a set of things to turn on.

-------------

PR Review Comment: https://git.openjdk.org/babylon/pull/618#discussion_r2445349066

From asotona at openjdk.org  Mon Oct 20 15:37:40 2025
From: asotona at openjdk.org (Adam Sotona)
Date: Mon, 20 Oct 2025 15:37:40 GMT
Subject: [code-reflection] RFR: Add support for other onnx providers
 (CoreML) [v2]
In-Reply-To: <FJcT6LVqJ2iE3C4oePtKTDatHZpDdRxqFQFqjlgJPeY=.913803cc-b469-4133-b3cc-cdd8009f4309@github.com>
References: <e7Q32XrF2lzb5NdbJDAbklKlodvEyxR9Ds38h_GSDvA=.96e74c03-8271-461d-bb73-638d83d07ea0@github.com>
 <zRusgia_NhzaeDkL6yHylgNoQqoXqW_UpujwB7slSo0=.280afd0b-199b-4bcd-bae4-c8d3da47ddb3@github.com>
 <VTW-BTJXZ1pqsJVoUOd68IG545h-Oq3wQSX0Sgy55mI=.688d421b-aa91-4dd4-ad02-8883bcdeffdf@github.com>
 <FJcT6LVqJ2iE3C4oePtKTDatHZpDdRxqFQFqjlgJPeY=.913803cc-b469-4133-b3cc-cdd8009f4309@github.com>
Message-ID: <U8W-QmyDmbvdquJ7BaJKwAgwW8uazmanvh8hdTtvTi8=.d8dc4881-41ae-4813-a69d-e3b43ee43f22@github.com>

On Mon, 20 Oct 2025 14:33:03 GMT, Ana Maria Mihalceanu <duke at openjdk.org> wrote:

>> cr-examples/onnx/src/main/java/oracle/code/onnx/OnnxRuntime.java line 244:
>> 
>>> 242:     }
>>> 243: 
>>> 244:     public static <T> T executeWithOptions(Arena arena, MethodHandles.Lookup l, OnnxFunction<T> codeLambda, SessionOptions options) {
>> 
>> I would recommend to make this method as an override of execute.
>> Original execute duplicates the code, so it can delegate with no options instead.
>
> Adam, can you please check the refactoring done for execute?

Yes, it looks good now.

-------------

PR Review Comment: https://git.openjdk.org/babylon/pull/618#discussion_r2445375325

From gfrost at openjdk.org  Mon Oct 20 15:56:52 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Mon, 20 Oct 2025 15:56:52 GMT
Subject: [code-reflection] RFR: Quick hack to allow us to override codegen with
 annotation
Message-ID: <49d_yq_0ONjCprZiAmLduEK5_DVL-25HjSzAOHqQ9hY=.48050ba6-e5eb-47b9-9f4a-9ec15f4b77ad@github.com>

This allows us to override codegen (for experiments and to prototype cogegen fixes)

Add this to @CodeReflection annotated kernel entrypoints. 


@Kernel("""
     //kernel code
""")


If you need to add a new type (shared private mem) 

@TypeDef("""
     typedef struct SharedS32x256Array_s{
         int array[32];
     }SharedS32x256Array_t;
""" )


See PrefixSum as an example.

-------------

Commit messages:
 - Quick hack to allow us to override codegen with annotation

Changes: https://git.openjdk.org/babylon/pull/622/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=622&range=00
  Stats: 135 lines in 4 files changed: 68 ins; 32 del; 35 mod
  Patch: https://git.openjdk.org/babylon/pull/622.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/622/head:pull/622

PR: https://git.openjdk.org/babylon/pull/622

From gfrost at openjdk.org  Mon Oct 20 16:04:41 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Mon, 20 Oct 2025 16:04:41 GMT
Subject: [code-reflection] Integrated: Quick hack to allow us to override
 codegen with annotation
In-Reply-To: <49d_yq_0ONjCprZiAmLduEK5_DVL-25HjSzAOHqQ9hY=.48050ba6-e5eb-47b9-9f4a-9ec15f4b77ad@github.com>
References: <49d_yq_0ONjCprZiAmLduEK5_DVL-25HjSzAOHqQ9hY=.48050ba6-e5eb-47b9-9f4a-9ec15f4b77ad@github.com>
Message-ID: <NcQOLAFygU2-9wwERwryoaz_yHAKWKcllR873vLSgrk=.a682bac8-bf37-49ac-8952-9ec01a5dd0c0@github.com>

On Mon, 20 Oct 2025 15:51:58 GMT, Gary Frost <gfrost at openjdk.org> wrote:

> This allows us to override codegen (for experiments and to prototype cogegen fixes)
> 
> Add this to @CodeReflection annotated kernel entrypoints. 
> 
> 
> @Kernel("""
>      //kernel code
> """)
> 
> 
> If you need to add a new type (shared private mem) 
> 
> @TypeDef("""
>      typedef struct SharedS32x256Array_s{
>          int array[32];
>      }SharedS32x256Array_t;
> """ )
> 
> 
> See PrefixSum as an example.

This pull request has now been integrated.

Changeset: d6ec0a5f
Author:    Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/d6ec0a5f0a2ff334a58e2d2bfcaff6d658f5f5d4
Stats:     135 lines in 4 files changed: 68 ins; 32 del; 35 mod

Quick hack to allow us to override codegen with annotation

-------------

PR: https://git.openjdk.org/babylon/pull/622

From gfrost at openjdk.org  Mon Oct 20 16:05:25 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Mon, 20 Oct 2025 16:05:25 GMT
Subject: git: openjdk/babylon: code-reflection: Quick hack to allow us to
 override codegen with annotation
Message-ID: <f2e4a7f0-2316-42de-8a8b-59df2314a582@openjdk.org>

Changeset: d6ec0a5f
Branch: code-reflection
Author:    Gary Frost <gfrost at openjdk.org>
Date:      2025-10-20 16:02:08 +0000
URL:       https://git.openjdk.org/babylon/commit/d6ec0a5f0a2ff334a58e2d2bfcaff6d658f5f5d4

Quick hack to allow us to override codegen with annotation

! hat/backends/ffi/shared/src/main/java/hat/backend/ffi/C99FFIBackend.java
= hat/core/src/main/java/hat/annotations/TypeDef.java
! hat/core/src/main/java/hat/codebuilders/TextBuilder.java
! hat/examples/experiments/src/main/java/experiments/PrefixSum.java


From duke at openjdk.org  Mon Oct 20 16:12:30 2025
From: duke at openjdk.org (Ana Maria Mihalceanu)
Date: Mon, 20 Oct 2025 16:12:30 GMT
Subject: [code-reflection] RFR: Add support for other onnx providers
 (CoreML) [v5]
In-Reply-To: <e7Q32XrF2lzb5NdbJDAbklKlodvEyxR9Ds38h_GSDvA=.96e74c03-8271-461d-bb73-638d83d07ea0@github.com>
References: <e7Q32XrF2lzb5NdbJDAbklKlodvEyxR9Ds38h_GSDvA=.96e74c03-8271-461d-bb73-638d83d07ea0@github.com>
Message-ID: <CFT6BKX3OYeUCCM7j6O5JpBx4MeR7ymfWsjXNJ_jXfs=.6e9ecfc4-49ac-46ba-a558-c6dad8692535@github.com>

> This PR contains a series of changes meant to integrate other ONNX providers.
> 
> * It contains generated Java bindings for CoreML provider: `oracle.code.onnx.foreign.coreml_provider_factory_h`
> * Configurations of providers are present in `oracle.code.onnx.provider`.
> * Provider configuration should occur via `SessionOptions` on inference session, prior to any work done by the native library. Hence why `OnnxRuntime` has an `execute` that gets `SessionOptions` too.
> * There is a small UI example that use [Facial Emotion Recognition](https://github.com/onnx/models/tree/main/validated/vision/body_analysis/emotion_ferplus) model on how use the CoreML provider under `test` directory, inside `oracle.code.onnx.fer`.
> * The `README.md` file contains instructions on how to run the example and how to regenerate the bindings for the provider. Instructions and scripts to regenerate the bindings were provided for macOS users only since CoreML is an Apple software.
> 
> Work done in this PR is done together with [Lize Raes](https://github.com/LizeRaes).

Ana Maria Mihalceanu has updated the pull request incrementally with one additional commit since the last revision:

  Rename field to flags for consistency with the Native API.

-------------

Changes:
  - all: https://git.openjdk.org/babylon/pull/618/files
  - new: https://git.openjdk.org/babylon/pull/618/files/db73b054..d2802dde

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=babylon&pr=618&range=04
 - incr: https://webrevs.openjdk.org/?repo=babylon&pr=618&range=03-04

  Stats: 4 lines in 1 file changed: 0 ins; 0 del; 4 mod
  Patch: https://git.openjdk.org/babylon/pull/618.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/618/head:pull/618

PR: https://git.openjdk.org/babylon/pull/618

From duke at openjdk.org  Mon Oct 20 16:12:33 2025
From: duke at openjdk.org (Ana Maria Mihalceanu)
Date: Mon, 20 Oct 2025 16:12:33 GMT
Subject: [code-reflection] RFR: Add support for other onnx providers
 (CoreML) [v4]
In-Reply-To: <cZmDplYevL0j-a9r0oazH7Y58hajzfPnA4LXacNyfrE=.4194413e-0557-4ec8-ac87-5620e10f58a5@github.com>
References: <e7Q32XrF2lzb5NdbJDAbklKlodvEyxR9Ds38h_GSDvA=.96e74c03-8271-461d-bb73-638d83d07ea0@github.com>
 <Zw-wfkY5p5gynk8_ymQa6JKFJRNVKu_rXfKR6dN5xXw=.ef55f980-a300-4cad-8c98-27e542690162@github.com>
 <cZmDplYevL0j-a9r0oazH7Y58hajzfPnA4LXacNyfrE=.4194413e-0557-4ec8-ac87-5620e10f58a5@github.com>
Message-ID: <X3YOICmiY4zR7bEzjhrF-MdhHk3lHNA59bVnb7DD9qo=.bd89c6ba-a603-43d3-bd64-b4849d237e43@github.com>

On Mon, 20 Oct 2025 15:25:23 GMT, Adam Pocock <duke at openjdk.org> wrote:

>> Ana Maria Mihalceanu has updated the pull request incrementally with one additional commit since the last revision:
>> 
>>   Remove tabs and convert to whitespaces only.
>
> cr-examples/onnx/src/main/java/oracle/code/onnx/provider/CoreMLProvider.java line 38:
> 
>> 36:     private int flag;
>> 37: 
>> 38:     public CoreMLProvider(int flag) {
> 
> This is a small thing, but I'd probably call this `flags` as it's really a set of bits, so you can or the flags together to get a set of things to turn on.

Done ?

-------------

PR Review Comment: https://git.openjdk.org/babylon/pull/618#discussion_r2445467905

From asotona at openjdk.org  Mon Oct 20 16:45:27 2025
From: asotona at openjdk.org (Adam Sotona)
Date: Mon, 20 Oct 2025 16:45:27 GMT
Subject: [code-reflection] RFR: BytecodeLift and related tests move under tests
 bytecode/lift
Message-ID: <VmyTQ-sae84tiL1UYtbJ1JKhjj5JjHnkqNFqu_WHoug=.c5841e83-f263-4c7c-92bc-b044dde5f81e@github.com>

`BytecodeLift`, all related code and tests moved under `test/jdk/java/lang/reflect/code/bytecode/lift`

-------------

Commit messages:
 - BytecodeLift and related tests move under test/jdk/java/lang/reflect/code/bytecode/lift

Changes: https://git.openjdk.org/babylon/pull/623/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=623&range=00
  Stats: 810 lines in 14 files changed: 710 ins; 85 del; 15 mod
  Patch: https://git.openjdk.org/babylon/pull/623.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/623/head:pull/623

PR: https://git.openjdk.org/babylon/pull/623

From asotona at openjdk.org  Mon Oct 20 18:14:32 2025
From: asotona at openjdk.org (Adam Sotona)
Date: Mon, 20 Oct 2025 18:14:32 GMT
Subject: [code-reflection] RFR: Add support for other onnx providers
 (CoreML) [v5]
In-Reply-To: <CFT6BKX3OYeUCCM7j6O5JpBx4MeR7ymfWsjXNJ_jXfs=.6e9ecfc4-49ac-46ba-a558-c6dad8692535@github.com>
References: <e7Q32XrF2lzb5NdbJDAbklKlodvEyxR9Ds38h_GSDvA=.96e74c03-8271-461d-bb73-638d83d07ea0@github.com>
 <CFT6BKX3OYeUCCM7j6O5JpBx4MeR7ymfWsjXNJ_jXfs=.6e9ecfc4-49ac-46ba-a558-c6dad8692535@github.com>
Message-ID: <Dd_ndirHzojNjjVSQ3l3Tbfa1CLjv2tvo4jwRk9sK08=.18755d85-9c03-402b-80c5-5ef1d2f65724@github.com>

On Mon, 20 Oct 2025 16:12:30 GMT, Ana Maria Mihalceanu <duke at openjdk.org> wrote:

>> This PR contains a series of changes meant to integrate other ONNX providers.
>> 
>> * It contains generated Java bindings for CoreML provider: `oracle.code.onnx.foreign.coreml_provider_factory_h`
>> * Configurations of providers are present in `oracle.code.onnx.provider`.
>> * Provider configuration should occur via `SessionOptions` on inference session, prior to any work done by the native library. Hence why `OnnxRuntime` has an `execute` that gets `SessionOptions` too.
>> * There is a small UI example that use [Facial Emotion Recognition](https://github.com/onnx/models/tree/main/validated/vision/body_analysis/emotion_ferplus) model on how use the CoreML provider under `test` directory, inside `oracle.code.onnx.fer`.
>> * The `README.md` file contains instructions on how to run the example and how to regenerate the bindings for the provider. Instructions and scripts to regenerate the bindings were provided for macOS users only since CoreML is an Apple software.
>> 
>> Work done in this PR is done together with [Lize Raes](https://github.com/LizeRaes).
>
> Ana Maria Mihalceanu has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Rename field to flags for consistency with the Native API.

Marked as reviewed by asotona (Reviewer).

-------------

PR Review: https://git.openjdk.org/babylon/pull/618#pullrequestreview-3357458900

From duke at openjdk.org  Mon Oct 20 18:26:33 2025
From: duke at openjdk.org (duke)
Date: Mon, 20 Oct 2025 18:26:33 GMT
Subject: [code-reflection] RFR: Add support for other onnx providers
 (CoreML) [v5]
In-Reply-To: <CFT6BKX3OYeUCCM7j6O5JpBx4MeR7ymfWsjXNJ_jXfs=.6e9ecfc4-49ac-46ba-a558-c6dad8692535@github.com>
References: <e7Q32XrF2lzb5NdbJDAbklKlodvEyxR9Ds38h_GSDvA=.96e74c03-8271-461d-bb73-638d83d07ea0@github.com>
 <CFT6BKX3OYeUCCM7j6O5JpBx4MeR7ymfWsjXNJ_jXfs=.6e9ecfc4-49ac-46ba-a558-c6dad8692535@github.com>
Message-ID: <9SDxTvGQe07bFcPwc-C3uvVO8-FruYS-9APFcBEV8So=.684d7f45-36e2-4a3e-bb6a-0100bb205d76@github.com>

On Mon, 20 Oct 2025 16:12:30 GMT, Ana Maria Mihalceanu <duke at openjdk.org> wrote:

>> This PR contains a series of changes meant to integrate other ONNX providers.
>> 
>> * It contains generated Java bindings for CoreML provider: `oracle.code.onnx.foreign.coreml_provider_factory_h`
>> * Configurations of providers are present in `oracle.code.onnx.provider`.
>> * Provider configuration should occur via `SessionOptions` on inference session, prior to any work done by the native library. Hence why `OnnxRuntime` has an `execute` that gets `SessionOptions` too.
>> * There is a small UI example that use [Facial Emotion Recognition](https://github.com/onnx/models/tree/main/validated/vision/body_analysis/emotion_ferplus) model on how use the CoreML provider under `test` directory, inside `oracle.code.onnx.fer`.
>> * The `README.md` file contains instructions on how to run the example and how to regenerate the bindings for the provider. Instructions and scripts to regenerate the bindings were provided for macOS users only since CoreML is an Apple software.
>> 
>> Work done in this PR is done together with [Lize Raes](https://github.com/LizeRaes).
>
> Ana Maria Mihalceanu has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Rename field to flags for consistency with the Native API.

@ammbra 
Your change (at version d2802dde059851124dd6caf22d81cf2deaecf37a) is now ready to be sponsored by a Committer.

-------------

PR Comment: https://git.openjdk.org/babylon/pull/618#issuecomment-3423259040

From psandoz at openjdk.org  Mon Oct 20 19:53:41 2025
From: psandoz at openjdk.org (Paul Sandoz)
Date: Mon, 20 Oct 2025 19:53:41 GMT
Subject: [code-reflection] RFR: BytecodeLift and related tests move under
 tests bytecode/lift
In-Reply-To: <VmyTQ-sae84tiL1UYtbJ1JKhjj5JjHnkqNFqu_WHoug=.c5841e83-f263-4c7c-92bc-b044dde5f81e@github.com>
References: <VmyTQ-sae84tiL1UYtbJ1JKhjj5JjHnkqNFqu_WHoug=.c5841e83-f263-4c7c-92bc-b044dde5f81e@github.com>
Message-ID: <k0U3Ev014YJu2fMwMB4tKUyShbz9UksEgxMh9OiReNM=.0bac1779-108f-46e5-9af7-9638a07f284b@github.com>

On Mon, 20 Oct 2025 16:38:28 GMT, Adam Sotona <asotona at openjdk.org> wrote:

> `BytecodeLift`, all related code and tests moved under `test/jdk/java/lang/reflect/code/bytecode/lift`

Looks good. Note to other folks that might be observing, this is a tactical move in preparation for incubation, reducing the pressure on code review, maintenance, and testing. We think the lifter is strategic feature, and longer term it will likely move back into the incubating module in a more robust form.

-------------

Marked as reviewed by psandoz (Lead).

PR Review: https://git.openjdk.org/babylon/pull/623#pullrequestreview-3357721040

From asotona at openjdk.org  Tue Oct 21 07:41:38 2025
From: asotona at openjdk.org (Adam Sotona)
Date: Tue, 21 Oct 2025 07:41:38 GMT
Subject: git: openjdk/babylon: code-reflection: Add support for other onnx
 providers (CoreML)
Message-ID: <e53a84f4-039c-463e-b20c-902268569409@openjdk.org>

Changeset: a43beadb
Branch: code-reflection
Author:    ammbra <mihalceanuana at gmail.com>
Committer: Adam Sotona <asotona at openjdk.org>
Date:      2025-10-21 07:38:50 +0000
URL:       https://git.openjdk.org/babylon/commit/a43beadb9347f7ce03d3d7f9b27e83fef5b7cf3f

Add support for other onnx providers (CoreML)

Reviewed-by: asotona

! cr-examples/onnx/README.md
+ cr-examples/onnx/opgen/setup.sh
+ cr-examples/onnx/opgen/symbols
! cr-examples/onnx/src/main/java/oracle/code/onnx/OnnxProtoBuilder.java
! cr-examples/onnx/src/main/java/oracle/code/onnx/OnnxRuntime.java
+ cr-examples/onnx/src/main/java/oracle/code/onnx/foreign/coreml_provider_factory_h.java
+ cr-examples/onnx/src/main/java/oracle/code/onnx/provider/CoreMLProvider.java
+ cr-examples/onnx/src/main/java/oracle/code/onnx/provider/OnnxProvider.java
+ cr-examples/onnx/src/test/java/oracle/code/onnx/fer/FERCoreMLDemo.java
+ cr-examples/onnx/src/test/java/oracle/code/onnx/fer/FERInference.java
+ cr-examples/onnx/src/test/java/oracle/code/onnx/fer/FERModel.java
+ cr-examples/onnx/src/test/resources/oracle/code/onnx/fer/happy.png
+ cr-examples/onnx/src/test/resources/oracle/code/onnx/fer/sad.png


From duke at openjdk.org  Tue Oct 21 07:41:46 2025
From: duke at openjdk.org (Ana Maria Mihalceanu)
Date: Tue, 21 Oct 2025 07:41:46 GMT
Subject: [code-reflection] Integrated: Add support for other onnx providers
 (CoreML)
In-Reply-To: <e7Q32XrF2lzb5NdbJDAbklKlodvEyxR9Ds38h_GSDvA=.96e74c03-8271-461d-bb73-638d83d07ea0@github.com>
References: <e7Q32XrF2lzb5NdbJDAbklKlodvEyxR9Ds38h_GSDvA=.96e74c03-8271-461d-bb73-638d83d07ea0@github.com>
Message-ID: <RSOiGUW0ZCd5RpDbJCdpvRRNUl60WtD7JoKSmrfKOe8=.31d3f9e6-61fb-47d1-a769-a655f5d06ae3@github.com>

On Fri, 17 Oct 2025 11:45:24 GMT, Ana Maria Mihalceanu <duke at openjdk.org> wrote:

> This PR contains a series of changes meant to integrate other ONNX providers.
> 
> * It contains generated Java bindings for CoreML provider: `oracle.code.onnx.foreign.coreml_provider_factory_h`
> * Configurations of providers are present in `oracle.code.onnx.provider`.
> * Provider configuration should occur via `SessionOptions` on inference session, prior to any work done by the native library. Hence why `OnnxRuntime` has an `execute` that gets `SessionOptions` too.
> * There is a small UI example that use [Facial Emotion Recognition](https://github.com/onnx/models/tree/main/validated/vision/body_analysis/emotion_ferplus) model on how use the CoreML provider under `test` directory, inside `oracle.code.onnx.fer`.
> * The `README.md` file contains instructions on how to run the example and how to regenerate the bindings for the provider. Instructions and scripts to regenerate the bindings were provided for macOS users only since CoreML is an Apple software.
> 
> Work done in this PR is done together with [Lize Raes](https://github.com/LizeRaes).

This pull request has now been integrated.

Changeset: a43beadb
Author:    ammbra <mihalceanuana at gmail.com>
Committer: Adam Sotona <asotona at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/a43beadb9347f7ce03d3d7f9b27e83fef5b7cf3f
Stats:     1067 lines in 13 files changed: 1050 ins; 3 del; 14 mod

Add support for other onnx providers (CoreML)

Reviewed-by: asotona

-------------

PR: https://git.openjdk.org/babylon/pull/618

From asotona at openjdk.org  Tue Oct 21 07:45:33 2025
From: asotona at openjdk.org (Adam Sotona)
Date: Tue, 21 Oct 2025 07:45:33 GMT
Subject: git: openjdk/babylon: code-reflection: BytecodeLift and related tests
 move under tests bytecode/lift
Message-ID: <d860e62f-6d98-4547-b09f-bcb0804245de@openjdk.org>

Changeset: 987eed59
Branch: code-reflection
Author:    Adam Sotona <asotona at openjdk.org>
Date:      2025-10-21 07:44:46 +0000
URL:       https://git.openjdk.org/babylon/commit/987eed59b3bf38c19e70410c35c0759e3125d251

BytecodeLift and related tests move under tests bytecode/lift

Reviewed-by: psandoz

! src/jdk.incubator.code/share/classes/jdk/incubator/code/bytecode/impl/BranchCompactor.java
- src/jdk.incubator.code/share/classes/jdk/incubator/code/bytecode/impl/BytecodeHelpers.java
! test/jdk/java/lang/reflect/code/bytecode/TestBytecode.java
= test/jdk/java/lang/reflect/code/bytecode/lift/BytecodeLift.java
= test/jdk/java/lang/reflect/code/bytecode/lift/SlotOp.java
= test/jdk/java/lang/reflect/code/bytecode/lift/SlotToVarTransformer.java
= test/jdk/java/lang/reflect/code/bytecode/lift/TEST.properties
+ test/jdk/java/lang/reflect/code/bytecode/lift/TestBytecodeLift.java
= test/jdk/java/lang/reflect/code/bytecode/lift/TestLiftCustomBytecode.java
= test/jdk/java/lang/reflect/code/bytecode/lift/TestLiftExample.java
= test/jdk/java/lang/reflect/code/bytecode/lift/TestSmallCorpus.java
= test/jdk/java/lang/reflect/code/bytecode/lift/UnresolvedType.java
= test/jdk/java/lang/reflect/code/bytecode/lift/UnresolvedTypesTransformer.java
= test/jdk/java/lang/reflect/code/bytecode/lift/Verifier.java


From asotona at openjdk.org  Tue Oct 21 07:47:44 2025
From: asotona at openjdk.org (Adam Sotona)
Date: Tue, 21 Oct 2025 07:47:44 GMT
Subject: [code-reflection] RFR: BytecodeLift and related tests move under
 tests bytecode/lift
In-Reply-To: <VmyTQ-sae84tiL1UYtbJ1JKhjj5JjHnkqNFqu_WHoug=.c5841e83-f263-4c7c-92bc-b044dde5f81e@github.com>
References: <VmyTQ-sae84tiL1UYtbJ1JKhjj5JjHnkqNFqu_WHoug=.c5841e83-f263-4c7c-92bc-b044dde5f81e@github.com>
Message-ID: <TJStpdwRmjt4Apfpymtn_n02_IIWokpDDDxgnXCYcE8=.637f0dc9-4662-4512-8bb3-a955aae7f155@github.com>

On Mon, 20 Oct 2025 16:38:28 GMT, Adam Sotona <asotona at openjdk.org> wrote:

> `BytecodeLift`, all related code and tests moved under `test/jdk/java/lang/reflect/code/bytecode/lift`

Thank you!

-------------

PR Comment: https://git.openjdk.org/babylon/pull/623#issuecomment-3425220130

From asotona at openjdk.org  Tue Oct 21 07:47:47 2025
From: asotona at openjdk.org (Adam Sotona)
Date: Tue, 21 Oct 2025 07:47:47 GMT
Subject: [code-reflection] Integrated: BytecodeLift and related tests move
 under tests bytecode/lift
In-Reply-To: <VmyTQ-sae84tiL1UYtbJ1JKhjj5JjHnkqNFqu_WHoug=.c5841e83-f263-4c7c-92bc-b044dde5f81e@github.com>
References: <VmyTQ-sae84tiL1UYtbJ1JKhjj5JjHnkqNFqu_WHoug=.c5841e83-f263-4c7c-92bc-b044dde5f81e@github.com>
Message-ID: <SM5NC9GUn-fy8iK8Xqb-TWhOyq0xCIrk6tWunFDnBCM=.da92fd6f-6110-4026-ad6f-72cd55c89596@github.com>

On Mon, 20 Oct 2025 16:38:28 GMT, Adam Sotona <asotona at openjdk.org> wrote:

> `BytecodeLift`, all related code and tests moved under `test/jdk/java/lang/reflect/code/bytecode/lift`

This pull request has now been integrated.

Changeset: 987eed59
Author:    Adam Sotona <asotona at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/987eed59b3bf38c19e70410c35c0759e3125d251
Stats:     810 lines in 14 files changed: 710 ins; 85 del; 15 mod

BytecodeLift and related tests move under tests bytecode/lift

Reviewed-by: psandoz

-------------

PR: https://git.openjdk.org/babylon/pull/623

From jfumero at openjdk.org  Tue Oct 21 09:06:13 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Tue, 21 Oct 2025 09:06:13 GMT
Subject: [code-reflection] RFR: [hat] Initial support for 16-bits float types
 (half)
Message-ID: <goCuOJSq59gOqKqzpYCNeNkk1d_rpT2gQA1jMxDY57k=.59722e0e-a6d4-420f-a3f7-8a3db64ff0f0@github.com>

This PR extends HAT with support for OpenCL and CUDA `half` types. 

>From the Java side:


F16Array.F16 ha = a.array(kernelContext.gix);
F16Array.F16 hb = b.array(kernelContext.gix);
hb.value(ha.value());


We can also perform some common operations with half types:


F16Array.F16 ha = a.array(kernelContext.gix);
F16Array.F16 hb = b.array(kernelContext.gix);

F16Array.F16 r1 = F16.mul(ha, hb);
F16Array.F16 r2 = F16.div(ha, hb);
F16Array.F16 r3 = F16.sub(ha, hb);
F16Array.F16 r4 = F16.add(r1, r2);
F16Array.F16 r5 = F16.mul(r4, r3);
F16Array.F16 hC = c.array(kernelContext.gix);


To run with FP16, HAT needs to run GPUs with HALF support for both OpenCL and CUDA backends. For example, for OpenCL, it needs to run on devices that support the `cl_khr_fp16` macro. 

In near future, we will extend with more operations. 

How to test (OpenCL):


HAT=SHOW_CODE,SHOW_COMPILATION_PHASES java @hat/otest ffi-opencl hat.test.TestFP16Type


CUDA: 


HAT=SHOW_CODE,SHOW_COMPILATION_PHASES java @hat/otest ffi-cuda hat.test.TestFP16Type

-------------

Commit messages:
 - [hat] hatVectorVarLoadOp moved to C99 builder
 - [hat] cleanup f16phase
 - [HAT] CUDA backend F16 initial support
 - [hat] missing header license added
 - Merge branch 'code-reflection' into hat/types/fp16
 - [hat] HalfType unittest fixed
 - [hat] HalfType for OpenCL and common operations
 - [hat] Initial implementation for FP16 and OpenCL backend

Changes: https://git.openjdk.org/babylon/pull/624/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=624&range=00
  Stats: 986 lines in 27 files changed: 929 ins; 44 del; 13 mod
  Patch: https://git.openjdk.org/babylon/pull/624.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/624/head:pull/624

PR: https://git.openjdk.org/babylon/pull/624

From jfumero at openjdk.org  Tue Oct 21 09:31:48 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Tue, 21 Oct 2025 09:31:48 GMT
Subject: git: openjdk/babylon: code-reflection: [hat] Initial support for
 16-bits float types (half)
Message-ID: <cf6614a5-d1be-4c78-96d8-8a62d5b84b61@openjdk.org>

Changeset: 5128c70b
Branch: code-reflection
Author:    Juan Fumero <jfumero at openjdk.org>
Date:      2025-10-21 09:29:11 +0000
URL:       https://git.openjdk.org/babylon/commit/5128c70b87c209d781307cf583826a2f5b8a1a8d

[hat] Initial support for 16-bits float types (half)

! hat/backends/ffi/cuda/src/main/java/hat/backend/ffi/CudaHATKernelBuilder.java
! hat/backends/ffi/opencl/src/main/java/hat/backend/ffi/OpenCLHATKernelBuilder.java
! hat/core/src/main/java/hat/buffer/ArgArray.java
! hat/core/src/main/java/hat/buffer/BufferAllocator.java
+ hat/core/src/main/java/hat/buffer/F16Array.java
! hat/core/src/main/java/hat/buffer/F32Array2D.java
! hat/core/src/main/java/hat/buffer/S08x3RGBImage.java
! hat/core/src/main/java/hat/buffer/S32Array.java
! hat/core/src/main/java/hat/buffer/S32Array2D.java
! hat/core/src/main/java/hat/buffer/S32RGBAImage.java
! hat/core/src/main/java/hat/buffer/U16GreyImage.java
! hat/core/src/main/java/hat/codebuilders/BabylonOpBuilder.java
! hat/core/src/main/java/hat/codebuilders/C99HATKernelBuilder.java
! hat/core/src/main/java/hat/codebuilders/HATCodeBuilder.java
! hat/core/src/main/java/hat/codebuilders/HATCodeBuilderWithContext.java
! hat/core/src/main/java/hat/codebuilders/ScopedCodeBuilderContext.java
+ hat/core/src/main/java/hat/dialect/HATF16BinaryOp.java
+ hat/core/src/main/java/hat/dialect/HATF16Op.java
+ hat/core/src/main/java/hat/dialect/HATF16VarLoadOp.java
+ hat/core/src/main/java/hat/dialect/HATF16VarOp.java
! hat/core/src/main/java/hat/optools/OpTk.java
+ hat/core/src/main/java/hat/phases/HATDialectifyFP16Phase.java
! hat/core/src/main/java/hat/phases/HATDialectifyTier.java
! hat/core/src/main/java/hat/phases/HATDialectifyVectorSelectPhase.java
! hat/hat/test.java
+ hat/tests/src/main/java/hat/test/TestFP16Type.java
! hat/tools/src/main/java/hat/tools/text/JavaHATCodeBuilder.java


From jfumero at openjdk.org  Tue Oct 21 09:33:04 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Tue, 21 Oct 2025 09:33:04 GMT
Subject: [code-reflection] Integrated: [hat] Initial support for 16-bits float
 types (half)
In-Reply-To: <goCuOJSq59gOqKqzpYCNeNkk1d_rpT2gQA1jMxDY57k=.59722e0e-a6d4-420f-a3f7-8a3db64ff0f0@github.com>
References: <goCuOJSq59gOqKqzpYCNeNkk1d_rpT2gQA1jMxDY57k=.59722e0e-a6d4-420f-a3f7-8a3db64ff0f0@github.com>
Message-ID: <tg6ThjIENwcWj97XDEw0DZZZnWN7iSgCub2jxlkytL8=.a26cd7c9-5280-4e8a-91ad-9509490d21de@github.com>

On Tue, 21 Oct 2025 08:40:11 GMT, Juan Fumero <jfumero at openjdk.org> wrote:

> This PR extends HAT with support for OpenCL and CUDA `half` types. 
> 
> From the Java side:
> 
> 
> F16Array.F16 ha = a.array(kernelContext.gix);
> F16Array.F16 hb = b.array(kernelContext.gix);
> hb.value(ha.value());
> 
> 
> We can also perform some common operations with half types:
> 
> 
> F16Array.F16 ha = a.array(kernelContext.gix);
> F16Array.F16 hb = b.array(kernelContext.gix);
> 
> F16Array.F16 r1 = F16.mul(ha, hb);
> F16Array.F16 r2 = F16.div(ha, hb);
> F16Array.F16 r3 = F16.sub(ha, hb);
> F16Array.F16 r4 = F16.add(r1, r2);
> F16Array.F16 r5 = F16.mul(r4, r3);
> F16Array.F16 hC = c.array(kernelContext.gix);
> 
> 
> To run with FP16, HAT needs to run GPUs with HALF support for both OpenCL and CUDA backends. For example, for OpenCL, it needs to run on devices that support the `cl_khr_fp16` macro. 
> 
> In near future, we will extend with more operations. 
> 
> How to test (OpenCL):
> 
> 
> HAT=SHOW_CODE,SHOW_COMPILATION_PHASES java @hat/otest ffi-opencl hat.test.TestFP16Type
> 
> 
> CUDA: 
> 
> 
> HAT=SHOW_CODE,SHOW_COMPILATION_PHASES java @hat/otest ffi-cuda hat.test.TestFP16Type

This pull request has now been integrated.

Changeset: 5128c70b
Author:    Juan Fumero <jfumero at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/5128c70b87c209d781307cf583826a2f5b8a1a8d
Stats:     986 lines in 27 files changed: 929 ins; 44 del; 13 mod

[hat] Initial support for 16-bits float types (half)

-------------

PR: https://git.openjdk.org/babylon/pull/624

From jfumero at openjdk.org  Tue Oct 21 10:05:34 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Tue, 21 Oct 2025 10:05:34 GMT
Subject: [code-reflection] RFR: [hat] clean-up non-necessary methods for vector
 types
Message-ID: <9CzZKR5kDA6FrSQwrqFVW7A-_D6FrAeQ3SCuu9stOhY=.3f34a1a1-614a-4ecb-8891-b694a37f97f1@github.com>

[hat] clean-up non-necessary methods for vector types

-------------

Commit messages:
 - Merge branch 'code-reflection' into hat/cleanup/codegen
 - [hat] clean-up non-necessary methods for vector types

Changes: https://git.openjdk.org/babylon/pull/625/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=625&range=00
  Stats: 84 lines in 5 files changed: 1 ins; 73 del; 10 mod
  Patch: https://git.openjdk.org/babylon/pull/625.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/625/head:pull/625

PR: https://git.openjdk.org/babylon/pull/625

From jfumero at openjdk.org  Tue Oct 21 10:13:54 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Tue, 21 Oct 2025 10:13:54 GMT
Subject: [code-reflection] Integrated: [hat] clean-up non-necessary methods for
 vector types
In-Reply-To: <9CzZKR5kDA6FrSQwrqFVW7A-_D6FrAeQ3SCuu9stOhY=.3f34a1a1-614a-4ecb-8891-b694a37f97f1@github.com>
References: <9CzZKR5kDA6FrSQwrqFVW7A-_D6FrAeQ3SCuu9stOhY=.3f34a1a1-614a-4ecb-8891-b694a37f97f1@github.com>
Message-ID: <rl1R7wUel68PKTwDFrHCNJ3cE_M0NraN_kux2TPV_Ys=.d474e3f7-d49a-41d4-a2c4-6249c2f6cd97@github.com>

On Tue, 21 Oct 2025 09:53:55 GMT, Juan Fumero <jfumero at openjdk.org> wrote:

> [hat] clean-up non-necessary methods for vector types

This pull request has now been integrated.

Changeset: 1a3011f0
Author:    Juan Fumero <jfumero at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/1a3011f07ff182b1d5940c48ed0901aafdaaa5ad
Stats:     84 lines in 5 files changed: 1 ins; 73 del; 10 mod

[hat] clean-up non-necessary methods for vector types

-------------

PR: https://git.openjdk.org/babylon/pull/625

From jfumero at openjdk.org  Tue Oct 21 10:14:19 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Tue, 21 Oct 2025 10:14:19 GMT
Subject: git: openjdk/babylon: code-reflection: [hat] clean-up non-necessary
 methods for vector types
Message-ID: <50b0a264-49f6-496f-b199-058575c36baa@openjdk.org>

Changeset: 1a3011f0
Branch: code-reflection
Author:    Juan Fumero <jfumero at openjdk.org>
Date:      2025-10-21 10:10:47 +0000
URL:       https://git.openjdk.org/babylon/commit/1a3011f07ff182b1d5940c48ed0901aafdaaa5ad

[hat] clean-up non-necessary methods for vector types

! hat/backends/ffi/cuda/src/main/java/hat/backend/ffi/CudaHATKernelBuilder.java
! hat/backends/ffi/opencl/src/main/java/hat/backend/ffi/OpenCLHATKernelBuilder.java
! hat/core/src/main/java/hat/codebuilders/C99HATKernelBuilder.java
! hat/core/src/main/java/hat/codebuilders/HATCodeBuilderWithContext.java
! hat/tools/src/main/java/hat/tools/text/JavaHATCodeBuilder.java


From gfrost at openjdk.org  Tue Oct 21 12:03:21 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Tue, 21 Oct 2025 12:03:21 GMT
Subject: git: openjdk/babylon: code-reflection: Refactored MemoryPhases to
 innerclasses
Message-ID: <19f86656-59ad-436d-a062-90184c5a8482@openjdk.org>

Changeset: 9baa108f
Branch: code-reflection
Author:    Gary Frost <gfrost at openjdk.org>
Date:      2025-10-21 12:01:55 +0000
URL:       https://git.openjdk.org/babylon/commit/9baa108f44555bfa817dfcbac05b3e34f0151bbe

Refactored MemoryPhases to innerclasses

! hat/core/src/main/java/hat/phases/HATDialect.java
! hat/core/src/main/java/hat/phases/HATDialectifyBarrierPhase.java
! hat/core/src/main/java/hat/phases/HATDialectifyMemoryPhase.java
- hat/core/src/main/java/hat/phases/HATDialectifyMemoryPrivatePhase.java
- hat/core/src/main/java/hat/phases/HATDialectifyMemorySharedPhase.java
! hat/core/src/main/java/hat/phases/HATDialectifyThreadsPhase.java
! hat/core/src/main/java/hat/phases/HATDialectifyTier.java
! hat/core/src/main/java/hat/phases/HATDialectifyVectorOpPhase.java
! hat/core/src/main/java/hat/phases/HATDialectifyVectorSelectPhase.java
! hat/core/src/main/java/hat/phases/HATDialectifyVectorStorePhase.java


From gfrost at openjdk.org  Tue Oct 21 12:04:54 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Tue, 21 Oct 2025 12:04:54 GMT
Subject: [code-reflection] Integrated: Refactored MemoryPhases to innerclasses
Message-ID: <LfZ_QPoQfIU4mqjx34PU0OrHv8LheMHw9boZnfU_IFA=.5a7201c1-bfa0-4b4c-906a-0186d7c8db91@github.com>

More phase transform refactoring. 

Re-establishing a pattern where we lean on switches over instance patterns (and inheritance) for phases and ops.

-------------

Commit messages:
 - Refactored MemoryPhases to innerclasses

Changes: https://git.openjdk.org/babylon/pull/626/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=626&range=00
  Stats: 426 lines in 10 files changed: 129 ins; 229 del; 68 mod
  Patch: https://git.openjdk.org/babylon/pull/626.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/626/head:pull/626

PR: https://git.openjdk.org/babylon/pull/626

From gfrost at openjdk.org  Tue Oct 21 12:04:55 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Tue, 21 Oct 2025 12:04:55 GMT
Subject: [code-reflection] Integrated: Refactored MemoryPhases to innerclasses
In-Reply-To: <LfZ_QPoQfIU4mqjx34PU0OrHv8LheMHw9boZnfU_IFA=.5a7201c1-bfa0-4b4c-906a-0186d7c8db91@github.com>
References: <LfZ_QPoQfIU4mqjx34PU0OrHv8LheMHw9boZnfU_IFA=.5a7201c1-bfa0-4b4c-906a-0186d7c8db91@github.com>
Message-ID: <hDuYrorzfsoyKR78CNaH1KLQzI4hI4hS7E_UD-JiyoM=.a6f5a3e6-aa72-4428-93ee-d8bc22e71fd9@github.com>

On Tue, 21 Oct 2025 11:58:04 GMT, Gary Frost <gfrost at openjdk.org> wrote:

> More phase transform refactoring. 
> 
> Re-establishing a pattern where we lean on switches over instance patterns (and inheritance) for phases and ops.

This pull request has now been integrated.

Changeset: 9baa108f
Author:    Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/9baa108f44555bfa817dfcbac05b3e34f0151bbe
Stats:     426 lines in 10 files changed: 129 ins; 229 del; 68 mod

Refactored MemoryPhases to innerclasses

-------------

PR: https://git.openjdk.org/babylon/pull/626

From gfrost at openjdk.org  Tue Oct 21 15:07:32 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Tue, 21 Oct 2025 15:07:32 GMT
Subject: [code-reflection] Integrated: more phase transform refctoring and a
 fix to hat.java for props
Message-ID: <VZYz9FcukKEsZcoT5UQLiXFDHb2xpq6vSEpP2tNKq38=.f2040211-ad02-43a4-a60c-5d664898a3d4@github.com>

More refactoring of transform phases. 

Also a fix to allow `java @hat/test-suite` ro accept properties. 

ala

java @hat/test-suite ffi-opencl -DHAT=SHOW_COMPILATION_PHASES -DTRACE_CALLSITES=true

-------------

Commit messages:
 - more phase transform refctoring and a fix to hat.java for props

Changes: https://git.openjdk.org/babylon/pull/627/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=627&range=00
  Stats: 60 lines in 4 files changed: 47 ins; 4 del; 9 mod
  Patch: https://git.openjdk.org/babylon/pull/627.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/627/head:pull/627

PR: https://git.openjdk.org/babylon/pull/627

From gfrost at openjdk.org  Tue Oct 21 15:11:45 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Tue, 21 Oct 2025 15:11:45 GMT
Subject: git: openjdk/babylon: code-reflection: more phase transform refctoring
 and a fix to hat.java for props
Message-ID: <10297a87-96f0-4414-b0ae-5c0f4a8d9af3@openjdk.org>

Changeset: 768bf4c5
Branch: code-reflection
Author:    Gary Frost <gfrost at openjdk.org>
Date:      2025-10-21 15:04:14 +0000
URL:       https://git.openjdk.org/babylon/commit/768bf4c59b60ce6b12cc23ac44c99ee6a42ba749

more phase transform refctoring and a fix to hat.java for props

! hat/core/src/main/java/hat/phases/HATDialectifyTier.java
! hat/core/src/main/java/hat/phases/HATDialectifyVectorOpPhase.java
! hat/core/src/main/java/hat/phases/HATDialectifyVectorStorePhase.java
! hat/hat.java


From gfrost at openjdk.org  Tue Oct 21 15:07:33 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Tue, 21 Oct 2025 15:07:33 GMT
Subject: [code-reflection] Integrated: more phase transform refctoring and a
 fix to hat.java for props
In-Reply-To: <VZYz9FcukKEsZcoT5UQLiXFDHb2xpq6vSEpP2tNKq38=.f2040211-ad02-43a4-a60c-5d664898a3d4@github.com>
References: <VZYz9FcukKEsZcoT5UQLiXFDHb2xpq6vSEpP2tNKq38=.f2040211-ad02-43a4-a60c-5d664898a3d4@github.com>
Message-ID: <ApUpjnPgyFp80k2TjRKj5DCzYltbLNYkpd74A3wAySI=.424aec7f-7ead-46b5-ae3b-f4c7a30d6aa0@github.com>

On Tue, 21 Oct 2025 14:58:54 GMT, Gary Frost <gfrost at openjdk.org> wrote:

> More refactoring of transform phases. 
> 
> Also a fix to allow `java @hat/test-suite` ro accept properties. 
> 
> ala
> 
> java @hat/test-suite ffi-opencl -DHAT=SHOW_COMPILATION_PHASES -DTRACE_CALLSITES=true

This pull request has now been integrated.

Changeset: 768bf4c5
Author:    Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/768bf4c59b60ce6b12cc23ac44c99ee6a42ba749
Stats:     60 lines in 4 files changed: 47 ins; 4 del; 9 mod

more phase transform refctoring and a fix to hat.java for props

-------------

PR: https://git.openjdk.org/babylon/pull/627

From gfrost at openjdk.org  Tue Oct 21 15:50:12 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Tue, 21 Oct 2025 15:50:12 GMT
Subject: git: openjdk/babylon: code-reflection: Added opencl-jextracted-backend
 to bld
Message-ID: <44279b22-b4f7-48f1-b205-e0144ae25d5a@openjdk.org>

Changeset: 20e9f4a7
Branch: code-reflection
Author:    Gary Frost <gfrost at openjdk.org>
Date:      2025-10-21 15:47:52 +0000
URL:       https://git.openjdk.org/babylon/commit/20e9f4a750113bcdd4236080b2a39660268c62bd

Added  opencl-jextracted-backend to bld

! hat/backends/jextracted/opencl/src/main/java/hat/backend/jextracted/OpenCLHatKernelBuilder.java
! hat/hat.java


From gfrost at openjdk.org  Tue Oct 21 15:50:52 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Tue, 21 Oct 2025 15:50:52 GMT
Subject: [code-reflection] Integrated: Added opencl-jextracted-backend to bld
Message-ID: <qug2stpTfPxRrXNhgOMe7hEt9b8zZPjyk44j2lNfI94=.1838e1ce-1869-4452-99ee-0509a94aebb8@github.com>

I noticed that opencl jextracted backend stub (not working but should be built) was not included in the bld 

So added it and the backends-jextracted-shared.

-------------

Commit messages:
 - Added  opencl-jextracted-backend to bld

Changes: https://git.openjdk.org/babylon/pull/628/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=628&range=00
  Stats: 15 lines in 2 files changed: 3 ins; 0 del; 12 mod
  Patch: https://git.openjdk.org/babylon/pull/628.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/628/head:pull/628

PR: https://git.openjdk.org/babylon/pull/628

From gfrost at openjdk.org  Tue Oct 21 15:50:54 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Tue, 21 Oct 2025 15:50:54 GMT
Subject: [code-reflection] Integrated: Added opencl-jextracted-backend to bld
In-Reply-To: <qug2stpTfPxRrXNhgOMe7hEt9b8zZPjyk44j2lNfI94=.1838e1ce-1869-4452-99ee-0509a94aebb8@github.com>
References: <qug2stpTfPxRrXNhgOMe7hEt9b8zZPjyk44j2lNfI94=.1838e1ce-1869-4452-99ee-0509a94aebb8@github.com>
Message-ID: <UklCp-z04tNhMhUkt_FdDpwYXeOhvkhFTsyAnygeYqU=.dca30387-a276-4951-98c3-3130ec771114@github.com>

On Tue, 21 Oct 2025 15:43:48 GMT, Gary Frost <gfrost at openjdk.org> wrote:

> I noticed that opencl jextracted backend stub (not working but should be built) was not included in the bld 
> 
> So added it and the backends-jextracted-shared.

This pull request has now been integrated.

Changeset: 20e9f4a7
Author:    Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/20e9f4a750113bcdd4236080b2a39660268c62bd
Stats:     15 lines in 2 files changed: 3 ins; 0 del; 12 mod

Added  opencl-jextracted-backend to bld

-------------

PR: https://git.openjdk.org/babylon/pull/628

From jfumero at openjdk.org  Wed Oct 22 08:40:02 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Wed, 22 Oct 2025 08:40:02 GMT
Subject: [code-reflection] RFR: [hat] F16 object creation within device code
 via F16.of
Message-ID: <_DLnk1p4Bn4-msvb3eHQclsqPlu3UdvllXdOwJYTHOc=.4d45567b-e64f-497b-bdac-fc9348a8d6ff@github.com>

FP16 matmul supported. 

Example:


F16Array.F16 ha = a.array(kernelContext.gix);
F16Array.F16 initVal = F16.of( 2.1f);
// store new half value
ha.value(initVal.value());


How to test:


java @hat/test ffi-opencl hat.test.TestFP16Type -DHAT=SHOW_COMPILATION_PHASES -DTRACE_CALLSITES=true

-------------

Commit messages:
 - [hat] Include vmoptions in the hat single test
 - [hat] jextracted OpenCLHatKernelBuilder fixed
 - Merge branch 'code-reflection' into hat/fp16/init
 - [hat] New test for float initialization
 - [hat] F16 dialect phase refactored
 - Merge branch 'code-reflection' into hat/fp16/init
 - [hat] minor cleanup
 - [hat] Rename F16.init to F16.of
 - [hat] F16 init values within the device supported

Changes: https://git.openjdk.org/babylon/pull/629/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=629&range=00
  Stats: 840 lines in 19 files changed: 685 ins; 81 del; 74 mod
  Patch: https://git.openjdk.org/babylon/pull/629.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/629/head:pull/629

PR: https://git.openjdk.org/babylon/pull/629

From jfumero at openjdk.org  Wed Oct 22 08:45:12 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Wed, 22 Oct 2025 08:45:12 GMT
Subject: [code-reflection] RFR: [hat] F16 object creation within device
 code via F16.of [v2]
In-Reply-To: <_DLnk1p4Bn4-msvb3eHQclsqPlu3UdvllXdOwJYTHOc=.4d45567b-e64f-497b-bdac-fc9348a8d6ff@github.com>
References: <_DLnk1p4Bn4-msvb3eHQclsqPlu3UdvllXdOwJYTHOc=.4d45567b-e64f-497b-bdac-fc9348a8d6ff@github.com>
Message-ID: <_dLHAiiouPoVwFDbGniBATC2iTDaYl8Ak0wgmqPgKbs=.b683dbd5-cd50-4999-941d-37e68c7099c8@github.com>

> FP16 matmul supported. 
> 
> Example:
> 
> 
> F16Array.F16 ha = a.array(kernelContext.gix);
> F16Array.F16 initVal = F16.of( 2.1f);
> // store new half value
> ha.value(initVal.value());
> 
> 
> How to test:
> 
> 
> java @hat/test ffi-opencl hat.test.TestFP16Type -DHAT=SHOW_COMPILATION_PHASES -DTRACE_CALLSITES=true

Juan Fumero has updated the pull request incrementally with one additional commit since the last revision:

  [hat] Refactor for F16 instrinsics

-------------

Changes:
  - all: https://git.openjdk.org/babylon/pull/629/files
  - new: https://git.openjdk.org/babylon/pull/629/files/3cc61d26..a30fc0d2

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=babylon&pr=629&range=01
 - incr: https://webrevs.openjdk.org/?repo=babylon&pr=629&range=00-01

  Stats: 30 lines in 4 files changed: 11 ins; 2 del; 17 mod
  Patch: https://git.openjdk.org/babylon/pull/629.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/629/head:pull/629

PR: https://git.openjdk.org/babylon/pull/629

From jfumero at openjdk.org  Wed Oct 22 09:18:09 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Wed, 22 Oct 2025 09:18:09 GMT
Subject: git: openjdk/babylon: code-reflection: [hat] F16 object creation
 within device code via F16.of
Message-ID: <3006a273-f954-4cf8-bafe-c4939dcd87ec@openjdk.org>

Changeset: ae059ba1
Branch: code-reflection
Author:    Juan Fumero <jfumero at openjdk.org>
Date:      2025-10-22 09:16:43 +0000
URL:       https://git.openjdk.org/babylon/commit/ae059ba17322b9d615cc932e3a26522ba73aa916

[hat] F16 object creation within device code via F16.of

! hat/backends/ffi/cuda/src/main/java/hat/backend/ffi/CudaHATKernelBuilder.java
! hat/backends/ffi/opencl/src/main/java/hat/backend/ffi/OpenCLHATKernelBuilder.java
! hat/backends/jextracted/opencl/src/main/java/hat/backend/jextracted/OpenCLHatKernelBuilder.java
! hat/core/src/main/java/hat/buffer/F16Array.java
! hat/core/src/main/java/hat/codebuilders/BabylonOpBuilder.java
! hat/core/src/main/java/hat/codebuilders/HATCodeBuilderWithContext.java
+ hat/core/src/main/java/hat/dialect/HATF16AddOp.java
! hat/core/src/main/java/hat/dialect/HATF16BinaryOp.java
+ hat/core/src/main/java/hat/dialect/HATF16ConvOp.java
+ hat/core/src/main/java/hat/dialect/HATF16DivOp.java
+ hat/core/src/main/java/hat/dialect/HATF16MulOp.java
+ hat/core/src/main/java/hat/dialect/HATF16SubOp.java
! hat/core/src/main/java/hat/optools/OpTk.java
! hat/core/src/main/java/hat/phases/HATDialect.java
! hat/core/src/main/java/hat/phases/HATDialectifyFP16Phase.java
! hat/hat.java
! hat/tests/src/main/java/hat/test/TestFP16Type.java
! hat/tests/src/main/java/hat/test/TestMatMul.java
! hat/tools/src/main/java/hat/tools/text/JavaHATCodeBuilder.java


From jfumero at openjdk.org  Wed Oct 22 09:22:28 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Wed, 22 Oct 2025 09:22:28 GMT
Subject: [code-reflection] Integrated: [hat] F16 object creation within device
 code via F16.of
In-Reply-To: <_DLnk1p4Bn4-msvb3eHQclsqPlu3UdvllXdOwJYTHOc=.4d45567b-e64f-497b-bdac-fc9348a8d6ff@github.com>
References: <_DLnk1p4Bn4-msvb3eHQclsqPlu3UdvllXdOwJYTHOc=.4d45567b-e64f-497b-bdac-fc9348a8d6ff@github.com>
Message-ID: <vcDIdweNThsiFbHABAAQdKqRNbc0e5lXSqjr8qlgIZ0=.f38000a6-0e91-485e-931b-50d9012c582a@github.com>

On Wed, 22 Oct 2025 08:33:20 GMT, Juan Fumero <jfumero at openjdk.org> wrote:

> FP16 matmul supported. 
> 
> Example:
> 
> 
> F16Array.F16 ha = a.array(kernelContext.gix);
> F16Array.F16 initVal = F16.of( 2.1f);
> // store new half value
> ha.value(initVal.value());
> 
> 
> How to test:
> 
> 
> java @hat/test ffi-opencl hat.test.TestFP16Type -DHAT=SHOW_COMPILATION_PHASES -DTRACE_CALLSITES=true

This pull request has now been integrated.

Changeset: ae059ba1
Author:    Juan Fumero <jfumero at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/ae059ba17322b9d615cc932e3a26522ba73aa916
Stats:     862 lines in 19 files changed: 696 ins; 83 del; 83 mod

[hat] F16 object creation within device code via F16.of

-------------

PR: https://git.openjdk.org/babylon/pull/629

From jfumero at openjdk.org  Wed Oct 22 09:52:04 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Wed, 22 Oct 2025 09:52:04 GMT
Subject: [code-reflection] RFR: [hat] CPU Implementation for F16
Message-ID: <cL2BVuWVELiVMNDOuIGd8bhr5z7OIm_iShlizblM0-s=.ac1be67c-6d9a-47ba-b569-380d41597280@github.com>

[hat] CPU Implementation for F16

-------------

Commit messages:
 - [hat] CPU Impl for F16 added

Changes: https://git.openjdk.org/babylon/pull/630/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=630&range=00
  Stats: 643 lines in 5 files changed: 330 ins; 304 del; 9 mod
  Patch: https://git.openjdk.org/babylon/pull/630.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/630/head:pull/630

PR: https://git.openjdk.org/babylon/pull/630

From jfumero at openjdk.org  Wed Oct 22 09:56:12 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Wed, 22 Oct 2025 09:56:12 GMT
Subject: git: openjdk/babylon: code-reflection: [hat] CPU Implementation for
 F16
Message-ID: <26fafe26-586d-45cb-9375-72ec950a606e@openjdk.org>

Changeset: 4731bacd
Branch: code-reflection
Author:    Juan Fumero <jfumero at openjdk.org>
Date:      2025-10-22 09:55:13 +0000
URL:       https://git.openjdk.org/babylon/commit/4731bacdbf1779b503fa4f9c31bf092fcfa8517f

[hat] CPU Implementation for F16

! hat/core/src/main/java/hat/buffer/F16Array.java
! hat/hat/test.java
+ hat/tests/src/main/java/hat/test/TestF16Type.java
- hat/tests/src/main/java/hat/test/TestFP16Type.java
! hat/tests/src/main/java/hat/test/TestMatMul.java


From jfumero at openjdk.org  Wed Oct 22 09:58:27 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Wed, 22 Oct 2025 09:58:27 GMT
Subject: [code-reflection] Integrated: [hat] CPU Implementation for F16
In-Reply-To: <cL2BVuWVELiVMNDOuIGd8bhr5z7OIm_iShlizblM0-s=.ac1be67c-6d9a-47ba-b569-380d41597280@github.com>
References: <cL2BVuWVELiVMNDOuIGd8bhr5z7OIm_iShlizblM0-s=.ac1be67c-6d9a-47ba-b569-380d41597280@github.com>
Message-ID: <s0A9diWfI42WxfrHtv1I3WArsKlVyOS2M8eqaDdwUik=.3a0a1f53-45ae-44d2-a79a-e96c8bee5c67@github.com>

On Wed, 22 Oct 2025 09:45:31 GMT, Juan Fumero <jfumero at openjdk.org> wrote:

> [hat] CPU Implementation for F16

This pull request has now been integrated.

Changeset: 4731bacd
Author:    Juan Fumero <jfumero at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/4731bacdbf1779b503fa4f9c31bf092fcfa8517f
Stats:     643 lines in 5 files changed: 330 ins; 304 del; 9 mod

[hat] CPU Implementation for F16

-------------

PR: https://git.openjdk.org/babylon/pull/630

From gfrost at openjdk.org  Wed Oct 22 13:13:21 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Wed, 22 Oct 2025 13:13:21 GMT
Subject: git: openjdk/babylon: code-reflection: Phase tranform refactoring and
 fixed before/after CallSite tracking
Message-ID: <8e532667-6c87-4c66-b0a7-095aff09d63f@openjdk.org>

Changeset: edb3718a
Branch: code-reflection
Author:    Gary Frost <gfrost at openjdk.org>
Date:      2025-10-22 13:12:07 +0000
URL:       https://git.openjdk.org/babylon/commit/edb3718ae70df775e7bc6edcd146eb7956f68666

Phase tranform refactoring and fixed before/after CallSite tracking

! hat/core/src/main/java/hat/buffer/F32Array.java
! hat/core/src/main/java/hat/buffer/F32ArrayPadded.java
! hat/core/src/main/java/hat/optools/OpTk.java
! hat/core/src/main/java/hat/phases/HATDialectifyMemoryPhase.java


From gfrost at openjdk.org  Wed Oct 22 13:16:03 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Wed, 22 Oct 2025 13:16:03 GMT
Subject: [code-reflection] Integrated: Phase tranform refactoring and fixed
 before/after CallSite tracking
Message-ID: <qKwOqwSI0xpXYNp9tM0GGQoOmA3KRRTWLqpoAhow02E=.84859642-2ca5-4187-b029-edef14b90cac@github.com>

Continued refactoring phase transforms 

Also fixed issue with CallSite.  So nowe can turn on SHOW_COMPILATION_PHASES without CALLSITE_TRACING

-------------

Commit messages:
 - More phase tranform refactoring Also fixed before/after CallSite tracking issue

Changes: https://git.openjdk.org/babylon/pull/631/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=631&range=00
  Stats: 87 lines in 4 files changed: 33 ins; 23 del; 31 mod
  Patch: https://git.openjdk.org/babylon/pull/631.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/631/head:pull/631

PR: https://git.openjdk.org/babylon/pull/631

From gfrost at openjdk.org  Wed Oct 22 13:16:05 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Wed, 22 Oct 2025 13:16:05 GMT
Subject: [code-reflection] Integrated: Phase tranform refactoring and fixed
 before/after CallSite tracking
In-Reply-To: <qKwOqwSI0xpXYNp9tM0GGQoOmA3KRRTWLqpoAhow02E=.84859642-2ca5-4187-b029-edef14b90cac@github.com>
References: <qKwOqwSI0xpXYNp9tM0GGQoOmA3KRRTWLqpoAhow02E=.84859642-2ca5-4187-b029-edef14b90cac@github.com>
Message-ID: <AivX2JDfeyPtOIvcUojogT0LZzb8NJSoH2CdCq_AeXU=.72927fd9-211e-40a8-8d8c-adbfe5494f8a@github.com>

On Wed, 22 Oct 2025 13:09:40 GMT, Gary Frost <gfrost at openjdk.org> wrote:

> Continued refactoring phase transforms 
> 
> Also fixed issue with CallSite.  So nowe can turn on SHOW_COMPILATION_PHASES without CALLSITE_TRACING

This pull request has now been integrated.

Changeset: edb3718a
Author:    Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/edb3718ae70df775e7bc6edcd146eb7956f68666
Stats:     87 lines in 4 files changed: 33 ins; 23 del; 31 mod

Phase tranform refactoring and fixed before/after CallSite tracking

-------------

PR: https://git.openjdk.org/babylon/pull/631

From gfrost at openjdk.org  Thu Oct 23 09:44:50 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Thu, 23 Oct 2025 09:44:50 GMT
Subject: git: openjdk/babylon: code-reflection: Added non-hatified view 3d
 perspective example
Message-ID: <59b3bba9-fcd0-462a-8b6b-589ce75ba697@openjdk.org>

Changeset: adcb1ea0
Branch: code-reflection
Author:    Gary Frost <gfrost at openjdk.org>
Date:      2025-10-23 09:40:54 +0000
URL:       https://git.openjdk.org/babylon/commit/adcb1ea06f5a2c0b04e2d944b59e4da1e8226d71

Added non-hatified view 3d perspective example

! hat/examples/pom.xml
= hat/examples/view/.gitignore
+ hat/examples/view/pom.xml
+ hat/examples/view/src/main/java/view/EliteMeshReader.java
+ hat/examples/view/src/main/java/view/Main.java
+ hat/examples/view/src/main/java/view/Rasterizer.java
+ hat/examples/view/src/main/java/view/View.java
+ hat/examples/view/src/main/java/view/ViewFrame.java
+ hat/examples/view/src/main/java/view/ZPos.java
+ hat/examples/view/src/main/java/view/f32/F32Line3D.java
+ hat/examples/view/src/main/java/view/f32/F32Mat4.java
+ hat/examples/view/src/main/java/view/f32/F32Mesh3D.java
+ hat/examples/view/src/main/java/view/f32/F32Triangle3D.java
+ hat/examples/view/src/main/java/view/f32/F32Vec3.java
+ hat/examples/view/src/main/java/view/f32/mat4.java
+ hat/examples/view/src/main/java/view/f32/projectionMat4.java
+ hat/examples/view/src/main/java/view/f32/rotationMat4.java
+ hat/examples/view/src/main/java/view/f32/scaleMat4.java
+ hat/examples/view/src/main/java/view/f32/translateMat4.java
+ hat/examples/view/src/main/java/view/f32/tri.java
+ hat/examples/view/src/main/java/view/f32/vec3.java
+ hat/examples/view/src/main/java/view/i32/I32Line2D.java
+ hat/examples/view/src/main/java/view/i32/I32Triangle2D.java
+ hat/examples/view/src/main/java/view/i32/I32Vec2.java
+ hat/examples/view/src/main/resources/meshes/Elite.txt
! hat/hat.java
! hat/intellij/.idea/modules.xml


From gfrost at openjdk.org  Thu Oct 23 09:45:16 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Thu, 23 Oct 2025 09:45:16 GMT
Subject: [code-reflection] Integrated: Added non-hatified view 3d perspective
 example
Message-ID: <aUdxaP-x1376Koo_2nvf5ge5CJJLT-MvqPtrbU-1aOE=.49543b2d-e631-41b1-b29d-610e49f2f954@github.com>

I have had this code for a few years. Was waiting for HAT to get vector support.  

I think its time to add as an example. 


java @hat/bld
java @hat/run ffi-opencl view


For UK men/women of a certain age

java @hat/run ffi-opencl view BOA

-------------

Commit messages:
 - Added non-hatified view 3d perspective example

Changes: https://git.openjdk.org/babylon/pull/632/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=632&range=00
  Stats: 4145 lines in 27 files changed: 4145 ins; 0 del; 0 mod
  Patch: https://git.openjdk.org/babylon/pull/632.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/632/head:pull/632

PR: https://git.openjdk.org/babylon/pull/632

From gfrost at openjdk.org  Thu Oct 23 09:45:16 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Thu, 23 Oct 2025 09:45:16 GMT
Subject: [code-reflection] Integrated: Added non-hatified view 3d perspective
 example
In-Reply-To: <aUdxaP-x1376Koo_2nvf5ge5CJJLT-MvqPtrbU-1aOE=.49543b2d-e631-41b1-b29d-610e49f2f954@github.com>
References: <aUdxaP-x1376Koo_2nvf5ge5CJJLT-MvqPtrbU-1aOE=.49543b2d-e631-41b1-b29d-610e49f2f954@github.com>
Message-ID: <DMR8pNK-OC1TSZiA95GtT5BvJXxL8Ymje1vZJQVTwsA=.9ccbcf7e-f67c-417b-9048-0db6c689a494@github.com>

On Thu, 23 Oct 2025 09:36:57 GMT, Gary Frost <gfrost at openjdk.org> wrote:

> I have had this code for a few years. Was waiting for HAT to get vector support.  
> 
> I think its time to add as an example. 
> 
> 
> java @hat/bld
> java @hat/run ffi-opencl view
> 
> 
> For UK men/women of a certain age
> 
> java @hat/run ffi-opencl view BOA

This pull request has now been integrated.

Changeset: adcb1ea0
Author:    Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/adcb1ea06f5a2c0b04e2d944b59e4da1e8226d71
Stats:     4145 lines in 27 files changed: 4145 ins; 0 del; 0 mod

Added non-hatified view 3d perspective example

-------------

PR: https://git.openjdk.org/babylon/pull/632

From gfrost at openjdk.org  Thu Oct 23 10:34:34 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Thu, 23 Oct 2025 10:34:34 GMT
Subject: git: openjdk/babylon: code-reflection: Allow arbritrary text to be
 injected using annotations
Message-ID: <dee5c44c-4a24-48bb-a985-0f6c7d405560@openjdk.org>

Changeset: 5fa15c4c
Branch: code-reflection
Author:    Gary Frost <gfrost at openjdk.org>
Date:      2025-10-23 10:33:06 +0000
URL:       https://git.openjdk.org/babylon/commit/5fa15c4c70ba7a17df306b81875dfddc596eaae6

Allow arbritrary text to be injected using annotations

! hat/backends/ffi/shared/src/main/java/hat/backend/ffi/C99FFIBackend.java
= hat/core/src/main/java/hat/annotations/Preformatted.java
! hat/core/src/main/java/hat/annotations/TypeDef.java
! hat/examples/experiments/src/main/java/experiments/PrefixSum.java


From gfrost at openjdk.org  Thu Oct 23 10:36:10 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Thu, 23 Oct 2025 10:36:10 GMT
Subject: [code-reflection] Integrated: Allow arbritrary text to be injected
 using annotations
Message-ID: <EFRQGQ6wquvB7LNBtM149ufb0oUMPrRCKgeliznTbCQ=.ae04e726-796c-49bf-bd4c-bdffcc419879@github.com>

We previously allow typdefs and kernel bodies to be injected using annotations on kernels>

Sometimes we want to inject arbitrary chunks of code.  

We can do this with `@Preformatted(" //My text")`

-------------

Commit messages:
 - Allow arbritrary text to be injected using annotations

Changes: https://git.openjdk.org/babylon/pull/633/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=633&range=00
  Stats: 21 lines in 4 files changed: 12 ins; 2 del; 7 mod
  Patch: https://git.openjdk.org/babylon/pull/633.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/633/head:pull/633

PR: https://git.openjdk.org/babylon/pull/633

From gfrost at openjdk.org  Thu Oct 23 10:36:11 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Thu, 23 Oct 2025 10:36:11 GMT
Subject: [code-reflection] Integrated: Allow arbritrary text to be injected
 using annotations
In-Reply-To: <EFRQGQ6wquvB7LNBtM149ufb0oUMPrRCKgeliznTbCQ=.ae04e726-796c-49bf-bd4c-bdffcc419879@github.com>
References: <EFRQGQ6wquvB7LNBtM149ufb0oUMPrRCKgeliznTbCQ=.ae04e726-796c-49bf-bd4c-bdffcc419879@github.com>
Message-ID: <xC0NxIlk842uoPvCqKCLfPc4kM_shH_h2KBB_GfxcfU=.f0e5c986-cb2e-471b-98e7-a1c0e88ccd29@github.com>

On Thu, 23 Oct 2025 10:29:29 GMT, Gary Frost <gfrost at openjdk.org> wrote:

> We previously allow typdefs and kernel bodies to be injected using annotations on kernels>
> 
> Sometimes we want to inject arbitrary chunks of code.  
> 
> We can do this with `@Preformatted(" //My text")`

This pull request has now been integrated.

Changeset: 5fa15c4c
Author:    Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/5fa15c4c70ba7a17df306b81875dfddc596eaae6
Stats:     21 lines in 4 files changed: 12 ins; 2 del; 7 mod

Allow arbritrary text to be injected using annotations

-------------

PR: https://git.openjdk.org/babylon/pull/633

From duke at openjdk.org  Thu Oct 23 12:35:12 2025
From: duke at openjdk.org (Ana Maria Mihalceanu)
Date: Thu, 23 Oct 2025 12:35:12 GMT
Subject: [code-reflection] RFR: Update jdk version in opgen project
Message-ID: <aJmXFUwqZjVSRPbpLU46PjgySQPpuat2Y74Ap1fEn2s=.fe775e1c-e9d4-43de-82cb-d7a49ca0a339@github.com>

This PR fixes the JDK version in `opgen` project. 
As Babylon is based now on JDK 26, the project setup should be based on JDK 26.

In a terminal window, running `mvn clean verify` from `cr-examples/onnx/pom.xml` failed with  


[ERROR] Failed to execute goal org.apache.maven.plugins:maven-compiler-plugin:3.11.0:compile (default-compile) on project onnx-opgen: Compilation failure
[ERROR] invalid source release 25 with --enable-preview
[ERROR]   (preview language features are only supported for release 26)

-------------

Commit messages:
 - Update jdk version in opgen project.

Changes: https://git.openjdk.org/babylon/pull/634/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=634&range=00
  Stats: 2 lines in 1 file changed: 0 ins; 0 del; 2 mod
  Patch: https://git.openjdk.org/babylon/pull/634.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/634/head:pull/634

PR: https://git.openjdk.org/babylon/pull/634

From duke at openjdk.org  Thu Oct 23 12:44:19 2025
From: duke at openjdk.org (Ana Maria Mihalceanu)
Date: Thu, 23 Oct 2025 12:44:19 GMT
Subject: [code-reflection] RFR: Update jdk version in opgen project [v2]
In-Reply-To: <aJmXFUwqZjVSRPbpLU46PjgySQPpuat2Y74Ap1fEn2s=.fe775e1c-e9d4-43de-82cb-d7a49ca0a339@github.com>
References: <aJmXFUwqZjVSRPbpLU46PjgySQPpuat2Y74Ap1fEn2s=.fe775e1c-e9d4-43de-82cb-d7a49ca0a339@github.com>
Message-ID: <zUIbOOACHFD87SPyaCJODnIf-JxtKKJDKAwh_sAGbc4=.afb86c2f-85f2-401d-9bc2-8b42af73d501@github.com>

> This PR fixes the JDK version in `opgen` project. 
> As Babylon is based now on JDK 26, the project setup should be based on JDK 26.
> 
> In a terminal window, running `mvn clean verify` from `cr-examples/onnx/pom.xml` failed with  
> 
> 
> [ERROR] Failed to execute goal org.apache.maven.plugins:maven-compiler-plugin:3.11.0:compile (default-compile) on project onnx-opgen: Compilation failure
> [ERROR] invalid source release 25 with --enable-preview
> [ERROR]   (preview language features are only supported for release 26)

Ana Maria Mihalceanu has updated the pull request with a new target base due to a merge or a rebase. The pull request now contains one commit:

  Allow arbritrary text to be injected using annotations

-------------

Changes: https://git.openjdk.org/babylon/pull/634/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=634&range=01
  Stats: 23 lines in 5 files changed: 12 ins; 2 del; 9 mod
  Patch: https://git.openjdk.org/babylon/pull/634.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/634/head:pull/634

PR: https://git.openjdk.org/babylon/pull/634

From duke at openjdk.org  Thu Oct 23 12:44:21 2025
From: duke at openjdk.org (Ana Maria Mihalceanu)
Date: Thu, 23 Oct 2025 12:44:21 GMT
Subject: [code-reflection] Withdrawn: Update jdk version in opgen project
In-Reply-To: <aJmXFUwqZjVSRPbpLU46PjgySQPpuat2Y74Ap1fEn2s=.fe775e1c-e9d4-43de-82cb-d7a49ca0a339@github.com>
References: <aJmXFUwqZjVSRPbpLU46PjgySQPpuat2Y74Ap1fEn2s=.fe775e1c-e9d4-43de-82cb-d7a49ca0a339@github.com>
Message-ID: <R00udfA1Fvml0UQ3TkoRu7NsfOV-2XI5GuyBnci-_v4=.72a26c98-3ddf-4488-aaef-7405e8ff1aef@github.com>

On Thu, 23 Oct 2025 12:28:50 GMT, Ana Maria Mihalceanu <duke at openjdk.org> wrote:

> This PR fixes the JDK version in `opgen` project. 
> As Babylon is based now on JDK 26, the project setup should be based on JDK 26.
> 
> In a terminal window, running `mvn clean verify` from `cr-examples/onnx/pom.xml` failed with  
> 
> 
> [ERROR] Failed to execute goal org.apache.maven.plugins:maven-compiler-plugin:3.11.0:compile (default-compile) on project onnx-opgen: Compilation failure
> [ERROR] invalid source release 25 with --enable-preview
> [ERROR]   (preview language features are only supported for release 26)

This pull request has been closed without being integrated.

-------------

PR: https://git.openjdk.org/babylon/pull/634

From duke at openjdk.org  Thu Oct 23 12:53:38 2025
From: duke at openjdk.org (Ana Maria Mihalceanu)
Date: Thu, 23 Oct 2025 12:53:38 GMT
Subject: [code-reflection] RFR: Update JDK version in opgen project
Message-ID: <Z4EmTRtMZ0iW0deLZ9ziyDyQ2zwu521zrmS1NvicqvA=.a8a0ccf4-c914-425d-affa-68ae0df390fe@github.com>

This PR fixes the JDK version in `opgen` project. 
As Babylon is based now on JDK 26, the project setup should be based on JDK 26.

In a terminal window, running `mvn clean verify` from `cr-examples/onnx/pom.xml` failed with  


[ERROR] Failed to execute goal org.apache.maven.plugins:maven-compiler-plugin:3.11.0:compile (default-compile) on project onnx-opgen: Compilation failure
[ERROR] invalid source release 25 with --enable-preview
[ERROR]   (preview language features are only supported for release 26)

-------------

Commit messages:
 - Update jdk version in opgen project.

Changes: https://git.openjdk.org/babylon/pull/635/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=635&range=00
  Stats: 2 lines in 1 file changed: 0 ins; 0 del; 2 mod
  Patch: https://git.openjdk.org/babylon/pull/635.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/635/head:pull/635

PR: https://git.openjdk.org/babylon/pull/635

From duke at openjdk.org  Thu Oct 23 12:58:49 2025
From: duke at openjdk.org (Ana Maria Mihalceanu)
Date: Thu, 23 Oct 2025 12:58:49 GMT
Subject: [code-reflection] RFR: Update JDK version in opgen project [v2]
In-Reply-To: <Z4EmTRtMZ0iW0deLZ9ziyDyQ2zwu521zrmS1NvicqvA=.a8a0ccf4-c914-425d-affa-68ae0df390fe@github.com>
References: <Z4EmTRtMZ0iW0deLZ9ziyDyQ2zwu521zrmS1NvicqvA=.a8a0ccf4-c914-425d-affa-68ae0df390fe@github.com>
Message-ID: <U6F6FLmKSsmHvN7iwXNyETR8TqWQHZrg8Vyk-8s3zc0=.18059e04-f761-4078-9d49-959e17fab38b@github.com>

> This PR fixes the JDK version in `opgen` project. 
> As Babylon is based now on JDK 26, the project setup should be based on JDK 26.
> 
> In a terminal window, running `mvn clean verify` from `cr-examples/onnx/pom.xml` failed with  
> 
> 
> [ERROR] Failed to execute goal org.apache.maven.plugins:maven-compiler-plugin:3.11.0:compile (default-compile) on project onnx-opgen: Compilation failure
> [ERROR] invalid source release 25 with --enable-preview
> [ERROR]   (preview language features are only supported for release 26)

Ana Maria Mihalceanu has updated the pull request incrementally with one additional commit since the last revision:

  Update full name

-------------

Changes:
  - all: https://git.openjdk.org/babylon/pull/635/files
  - new: https://git.openjdk.org/babylon/pull/635/files/f8ca1dc6..3c3e93fa

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=babylon&pr=635&range=01
 - incr: https://webrevs.openjdk.org/?repo=babylon&pr=635&range=00-01

  Stats: 0 lines in 0 files changed: 0 ins; 0 del; 0 mod
  Patch: https://git.openjdk.org/babylon/pull/635.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/635/head:pull/635

PR: https://git.openjdk.org/babylon/pull/635

From duke at openjdk.org  Thu Oct 23 13:07:34 2025
From: duke at openjdk.org (Ana Maria Mihalceanu)
Date: Thu, 23 Oct 2025 13:07:34 GMT
Subject: [code-reflection] RFR: Update JDK version in opgen project [v3]
In-Reply-To: <Z4EmTRtMZ0iW0deLZ9ziyDyQ2zwu521zrmS1NvicqvA=.a8a0ccf4-c914-425d-affa-68ae0df390fe@github.com>
References: <Z4EmTRtMZ0iW0deLZ9ziyDyQ2zwu521zrmS1NvicqvA=.a8a0ccf4-c914-425d-affa-68ae0df390fe@github.com>
Message-ID: <33HEZArwzz3WdyArUUvMYNcqD6H60lLUXcMyiFda4Pk=.dd7c74ad-cf8f-4676-a7c4-0cb946669496@github.com>

> This PR fixes the JDK version in `opgen` project. 
> It also removes a duplicate branch in a `switch` of `OnnxProtoBuilder.java`.
> 
> As Babylon is based now on JDK 26, the project setup should be based on JDK 26.
> In a terminal window, running `mvn clean verify` from `cr-examples/onnx/opgen/pom.xml` failed with  
> 
> 
> [ERROR] Failed to execute goal org.apache.maven.plugins:maven-compiler-plugin:3.11.0:compile (default-compile) on project onnx-opgen: Compilation failure
> [ERROR] invalid source release 25 with --enable-preview
> [ERROR]   (preview language features are only supported for release 26)

Ana Maria Mihalceanu has updated the pull request incrementally with one additional commit since the last revision:

  Fix switch compilation error after merge.
  
  Signed-off-by: ammbra <mihalceanuana at gmail.com>

-------------

Changes:
  - all: https://git.openjdk.org/babylon/pull/635/files
  - new: https://git.openjdk.org/babylon/pull/635/files/3c3e93fa..30c70097

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=babylon&pr=635&range=02
 - incr: https://webrevs.openjdk.org/?repo=babylon&pr=635&range=01-02

  Stats: 4 lines in 1 file changed: 0 ins; 4 del; 0 mod
  Patch: https://git.openjdk.org/babylon/pull/635.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/635/head:pull/635

PR: https://git.openjdk.org/babylon/pull/635

From gfrost at openjdk.org  Thu Oct 23 14:38:58 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Thu, 23 Oct 2025 14:38:58 GMT
Subject: git: openjdk/babylon: code-reflection: Simplify parsing of view data
Message-ID: <183d2c09-87ab-4cb8-9722-e250dd5f87c2@openjdk.org>

Changeset: ae7306c7
Branch: code-reflection
Author:    Gary Frost <gfrost at openjdk.org>
Date:      2025-10-23 14:36:05 +0000
URL:       https://git.openjdk.org/babylon/commit/ae7306c7426311e9583a366f8467ba5f05a3a384

Simplify parsing of view data

! hat/examples/view/src/main/java/view/EliteMeshReader.java
! hat/examples/view/src/main/java/view/Main.java
+ hat/examples/view/src/main/java/view/Regex.java
+ hat/intellij/example_view.iml


From gfrost at openjdk.org  Thu Oct 23 14:39:46 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Thu, 23 Oct 2025 14:39:46 GMT
Subject: [code-reflection] Integrated: Simplify parsing of view data
In-Reply-To: <5Y5hlQz_g6n3ysbzRO3JbqzXwCRoHV_-WbF3_DwzajQ=.80ff86fd-b4c6-45dc-a514-f386a92872ed@github.com>
References: <5Y5hlQz_g6n3ysbzRO3JbqzXwCRoHV_-WbF3_DwzajQ=.80ff86fd-b4c6-45dc-a514-f386a92872ed@github.com>
Message-ID: <wTGmwTkGE04RqH2kbdrVXAesCmvS4hL4qyRM95k3jrA=.6177acb7-e8cd-4cd1-829d-aa76699dc9fb@github.com>

On Thu, 23 Oct 2025 14:31:44 GMT, Gary Frost <gfrost at openjdk.org> wrote:

> Added example-view .iml file 
> 
> Also simplified parsing of meshes.

This pull request has now been integrated.

Changeset: ae7306c7
Author:    Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/ae7306c7426311e9583a366f8467ba5f05a3a384
Stats:     224 lines in 4 files changed: 106 ins; 64 del; 54 mod

Simplify parsing of view data

-------------

PR: https://git.openjdk.org/babylon/pull/636

From gfrost at openjdk.org  Thu Oct 23 14:39:45 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Thu, 23 Oct 2025 14:39:45 GMT
Subject: [code-reflection] Integrated: Simplify parsing of view data
Message-ID: <5Y5hlQz_g6n3ysbzRO3JbqzXwCRoHV_-WbF3_DwzajQ=.80ff86fd-b4c6-45dc-a514-f386a92872ed@github.com>

Added example-view .iml file 

Also simplified parsing of meshes.

-------------

Commit messages:
 - Simplify parsing of view data

Changes: https://git.openjdk.org/babylon/pull/636/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=636&range=00
  Stats: 224 lines in 4 files changed: 106 ins; 64 del; 54 mod
  Patch: https://git.openjdk.org/babylon/pull/636.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/636/head:pull/636

PR: https://git.openjdk.org/babylon/pull/636

From duke at openjdk.org  Thu Oct 23 15:21:13 2025
From: duke at openjdk.org (Ana Maria Mihalceanu)
Date: Thu, 23 Oct 2025 15:21:13 GMT
Subject: [code-reflection] Integrated: Update JDK version in opgen project
In-Reply-To: <Z4EmTRtMZ0iW0deLZ9ziyDyQ2zwu521zrmS1NvicqvA=.a8a0ccf4-c914-425d-affa-68ae0df390fe@github.com>
References: <Z4EmTRtMZ0iW0deLZ9ziyDyQ2zwu521zrmS1NvicqvA=.a8a0ccf4-c914-425d-affa-68ae0df390fe@github.com>
Message-ID: <MyzyL2g4iYvW8O2b5qHWFxh7V7qON_8xLHPcDbrMsko=.33e76073-458e-4ca4-bd33-98df6ef9fec9@github.com>

On Thu, 23 Oct 2025 12:47:03 GMT, Ana Maria Mihalceanu <duke at openjdk.org> wrote:

> This PR fixes the JDK version in `opgen` project. 
> It also removes a duplicate branch in a `switch` of `OnnxProtoBuilder.java`.
> 
> As Babylon is based now on JDK 26, the project setup should be based on JDK 26.
> In a terminal window, running `mvn clean verify` from `cr-examples/onnx/opgen/pom.xml` failed with  
> 
> 
> [ERROR] Failed to execute goal org.apache.maven.plugins:maven-compiler-plugin:3.11.0:compile (default-compile) on project onnx-opgen: Compilation failure
> [ERROR] invalid source release 25 with --enable-preview
> [ERROR]   (preview language features are only supported for release 26)

This pull request has now been integrated.

Changeset: cbc0ed44
Author:    ammbra <mihalceanuana at gmail.com>
Committer: Adam Sotona <asotona at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/cbc0ed446c7eb669b051e2aa4bb9dfc88c612606
Stats:     6 lines in 2 files changed: 0 ins; 4 del; 2 mod

Update JDK version in opgen project

Reviewed-by: asotona

-------------

PR: https://git.openjdk.org/babylon/pull/635

From asotona at openjdk.org  Thu Oct 23 15:22:22 2025
From: asotona at openjdk.org (Adam Sotona)
Date: Thu, 23 Oct 2025 15:22:22 GMT
Subject: git: openjdk/babylon: code-reflection: Update JDK version in opgen
 project
Message-ID: <f84ae7b0-36e3-49d7-998d-e66f04574cb1@openjdk.org>

Changeset: cbc0ed44
Branch: code-reflection
Author:    ammbra <mihalceanuana at gmail.com>
Committer: Adam Sotona <asotona at openjdk.org>
Date:      2025-10-23 15:18:32 +0000
URL:       https://git.openjdk.org/babylon/commit/cbc0ed446c7eb669b051e2aa4bb9dfc88c612606

Update JDK version in opgen project

Reviewed-by: asotona

! cr-examples/onnx/opgen/pom.xml
! cr-examples/onnx/src/main/java/oracle/code/onnx/OnnxProtoBuilder.java


From asotona at openjdk.org  Thu Oct 23 15:10:38 2025
From: asotona at openjdk.org (Adam Sotona)
Date: Thu, 23 Oct 2025 15:10:38 GMT
Subject: [code-reflection] RFR: Update JDK version in opgen project [v3]
In-Reply-To: <33HEZArwzz3WdyArUUvMYNcqD6H60lLUXcMyiFda4Pk=.dd7c74ad-cf8f-4676-a7c4-0cb946669496@github.com>
References: <Z4EmTRtMZ0iW0deLZ9ziyDyQ2zwu521zrmS1NvicqvA=.a8a0ccf4-c914-425d-affa-68ae0df390fe@github.com>
 <33HEZArwzz3WdyArUUvMYNcqD6H60lLUXcMyiFda4Pk=.dd7c74ad-cf8f-4676-a7c4-0cb946669496@github.com>
Message-ID: <jh-IMaVAKYnu3gCPEQKvwFcyJkJG3ri0344NZi4eI5g=.7d2b3f95-5a59-415a-b1f6-b82807199f97@github.com>

On Thu, 23 Oct 2025 13:07:34 GMT, Ana Maria Mihalceanu <duke at openjdk.org> wrote:

>> This PR fixes the JDK version in `opgen` project. 
>> It also removes a duplicate branch in a `switch` of `OnnxProtoBuilder.java`.
>> 
>> As Babylon is based now on JDK 26, the project setup should be based on JDK 26.
>> In a terminal window, running `mvn clean verify` from `cr-examples/onnx/opgen/pom.xml` failed with  
>> 
>> 
>> [ERROR] Failed to execute goal org.apache.maven.plugins:maven-compiler-plugin:3.11.0:compile (default-compile) on project onnx-opgen: Compilation failure
>> [ERROR] invalid source release 25 with --enable-preview
>> [ERROR]   (preview language features are only supported for release 26)
>
> Ana Maria Mihalceanu has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Fix switch compilation error after merge.
>   
>   Signed-off-by: ammbra <mihalceanuana at gmail.com>

Looks good to me.

-------------

Marked as reviewed by asotona (Reviewer).

PR Review: https://git.openjdk.org/babylon/pull/635#pullrequestreview-3370486768

From duke at openjdk.org  Thu Oct 23 15:14:47 2025
From: duke at openjdk.org (duke)
Date: Thu, 23 Oct 2025 15:14:47 GMT
Subject: [code-reflection] RFR: Update JDK version in opgen project [v3]
In-Reply-To: <33HEZArwzz3WdyArUUvMYNcqD6H60lLUXcMyiFda4Pk=.dd7c74ad-cf8f-4676-a7c4-0cb946669496@github.com>
References: <Z4EmTRtMZ0iW0deLZ9ziyDyQ2zwu521zrmS1NvicqvA=.a8a0ccf4-c914-425d-affa-68ae0df390fe@github.com>
 <33HEZArwzz3WdyArUUvMYNcqD6H60lLUXcMyiFda4Pk=.dd7c74ad-cf8f-4676-a7c4-0cb946669496@github.com>
Message-ID: <mdmibqe--FNxgrkgdoUaeujAcuuZK33BUPE-HL64rZY=.a6ebe7c8-d8e6-48a2-b456-9fd0088fd274@github.com>

On Thu, 23 Oct 2025 13:07:34 GMT, Ana Maria Mihalceanu <duke at openjdk.org> wrote:

>> This PR fixes the JDK version in `opgen` project. 
>> It also removes a duplicate branch in a `switch` of `OnnxProtoBuilder.java`.
>> 
>> As Babylon is based now on JDK 26, the project setup should be based on JDK 26.
>> In a terminal window, running `mvn clean verify` from `cr-examples/onnx/opgen/pom.xml` failed with  
>> 
>> 
>> [ERROR] Failed to execute goal org.apache.maven.plugins:maven-compiler-plugin:3.11.0:compile (default-compile) on project onnx-opgen: Compilation failure
>> [ERROR] invalid source release 25 with --enable-preview
>> [ERROR]   (preview language features are only supported for release 26)
>
> Ana Maria Mihalceanu has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Fix switch compilation error after merge.
>   
>   Signed-off-by: ammbra <mihalceanuana at gmail.com>

@ammbra 
Your change (at version 30c70097da04f3cefa190f49ee65f6cc22c4c17b) is now ready to be sponsored by a Committer.

-------------

PR Comment: https://git.openjdk.org/babylon/pull/635#issuecomment-3437552378

From duke at openjdk.org  Fri Oct 24 06:16:37 2025
From: duke at openjdk.org (Ruby Chen)
Date: Fri, 24 Oct 2025 06:16:37 GMT
Subject: [code-reflection] Withdrawn: Add kernel entrypoint to ModuleOp
In-Reply-To: <UkyuDt6Ei-D4V2Ly2ukTEV-Sx9qLIbN6ih3KSJ1BHYY=.45a19d55-ee90-46ec-b41c-28ec98117c5b@github.com>
References: <UkyuDt6Ei-D4V2Ly2ukTEV-Sx9qLIbN6ih3KSJ1BHYY=.45a19d55-ee90-46ec-b41c-28ec98117c5b@github.com>
Message-ID: <Jp1re7C004DM-MpujGHdv_RafwMiFl51SHaOsm2IQ7M=.5b5e7542-70f2-46c7-a66c-88d7fb0cb660@github.com>

On Tue, 7 Oct 2025 21:45:20 GMT, Ruby Chen <duke at openjdk.org> wrote:

> Include the kernel (and compute) entrypoints in their respective ModuleOps.

This pull request has been closed without being integrated.

-------------

PR: https://git.openjdk.org/babylon/pull/605

From duke at openjdk.org  Fri Oct 24 06:18:37 2025
From: duke at openjdk.org (duke)
Date: Fri, 24 Oct 2025 06:18:37 GMT
Subject: [code-reflection] RFR: Update buffer tagging with arrayview
 support [v2]
In-Reply-To: <1jdULGHfHB4nCUXUpVPryHcmGqIfOoAX82UivLaw8gc=.2010ead2-2d31-4acd-9aa7-338abe739c73@github.com>
References: <Q9zdEEQTILTtqvItQQ26aC5TCXabT0YwpnlNFFY-fZg=.e0f9b877-1021-4d2f-b8e0-b687928a61dc@github.com>
 <1jdULGHfHB4nCUXUpVPryHcmGqIfOoAX82UivLaw8gc=.2010ead2-2d31-4acd-9aa7-338abe739c73@github.com>
Message-ID: <xEg29YCklbSyfDaw4QDtbpgv4wEhCcErjZxUXVQYuAI=.0e4c49b1-c991-4110-9e83-5457e5b416df@github.com>

On Fri, 17 Oct 2025 17:10:22 GMT, Ruby Chen <duke at openjdk.org> wrote:

>> Add buffer tagging for array views, and also use `elements()` instead of `traverse()`.
>
> Ruby Chen has updated the pull request with a new target base due to a merge or a rebase. The pull request now contains two commits:
> 
>  - Merge branch 'code-reflection' into module-op-support
>  - Update buffer tagging with arrayview support

@rbrchen 
Your change (at version bc4776b4d2b1ec838bfcbd26c87c1af7adfed5d3) is now ready to be sponsored by a Committer.

-------------

PR Comment: https://git.openjdk.org/babylon/pull/610#issuecomment-3441251195

From jfumero at openjdk.org  Fri Oct 24 08:56:57 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Fri, 24 Oct 2025 08:56:57 GMT
Subject: [code-reflection] RFR: [hat] TestMul 1D thread-block and test f16
 improved
Message-ID: <3CtT0RmoDCipCTlwM_7moBy1r4CrFnRYrHZ2nLTd4B4=.f3bb5553-bf14-4c9c-b9f0-c6be4a2371f5@github.com>

TestMul 1D thread-block and test f16 improved

-------------

Commit messages:
 - [hat] TestMul 1D thread-block and test f16 improved

Changes: https://git.openjdk.org/babylon/pull/637/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=637&range=00
  Stats: 21 lines in 3 files changed: 9 ins; 7 del; 5 mod
  Patch: https://git.openjdk.org/babylon/pull/637.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/637/head:pull/637

PR: https://git.openjdk.org/babylon/pull/637

From asotona at openjdk.org  Fri Oct 24 09:24:50 2025
From: asotona at openjdk.org (Adam Sotona)
Date: Fri, 24 Oct 2025 09:24:50 GMT
Subject: [code-reflection] RFR: OnnxLift prototype
Message-ID: <dFnT1a6oJdYYzER2-cOVWCIgNce9uh0mBloQ3T52EDc=.612309ef-e96a-4a48-9223-3cafbda9663d@github.com>

This is a prototype of OnnxLift tool.
It lifts ONNX model binary to ONNX code reflection model, extracts weights, and generates Java model source.

-------------

Commit messages:
 - OnnxLift prototype

Changes: https://git.openjdk.org/babylon/pull/638/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=638&range=00
  Stats: 729 lines in 3 files changed: 727 ins; 1 del; 1 mod
  Patch: https://git.openjdk.org/babylon/pull/638.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/638/head:pull/638

PR: https://git.openjdk.org/babylon/pull/638

From jfumero at openjdk.org  Fri Oct 24 10:17:14 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Fri, 24 Oct 2025 10:17:14 GMT
Subject: [code-reflection] Integrated: [hat] TestMul 1D thread-block and test
 f16 improved
In-Reply-To: <3CtT0RmoDCipCTlwM_7moBy1r4CrFnRYrHZ2nLTd4B4=.f3bb5553-bf14-4c9c-b9f0-c6be4a2371f5@github.com>
References: <3CtT0RmoDCipCTlwM_7moBy1r4CrFnRYrHZ2nLTd4B4=.f3bb5553-bf14-4c9c-b9f0-c6be4a2371f5@github.com>
Message-ID: <GUW3Wi6TP6znTaC_8zxMG9-eiD9wGe2rylBKj_Otyt0=.cb2cd771-3a4c-41f1-8a68-5030c089f431@github.com>

On Fri, 24 Oct 2025 08:51:34 GMT, Juan Fumero <jfumero at openjdk.org> wrote:

> TestMul 1D thread-block and test f16 improved

This pull request has now been integrated.

Changeset: 66109287
Author:    Juan Fumero <jfumero at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/66109287d4fbb21622b13bc9ba2293d512000e80
Stats:     21 lines in 3 files changed: 9 ins; 7 del; 5 mod

[hat] TestMul 1D thread-block and test f16 improved

-------------

PR: https://git.openjdk.org/babylon/pull/637

From jfumero at openjdk.org  Fri Oct 24 10:19:56 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Fri, 24 Oct 2025 10:19:56 GMT
Subject: git: openjdk/babylon: code-reflection: [hat] TestMul 1D thread-block
 and test f16 improved
Message-ID: <21834153-8fb6-46c6-ba71-0108fa5eba18@openjdk.org>

Changeset: 66109287
Branch: code-reflection
Author:    Juan Fumero <jfumero at openjdk.org>
Date:      2025-10-24 10:14:25 +0000
URL:       https://git.openjdk.org/babylon/commit/66109287d4fbb21622b13bc9ba2293d512000e80

[hat] TestMul 1D thread-block and test f16 improved

! hat/examples/matmul/src/main/java/matmul/Main.java
! hat/tests/src/main/java/hat/test/TestF16Type.java
! hat/tests/src/main/java/hat/test/TestMatMul.java


From jfumero at openjdk.org  Fri Oct 24 10:31:02 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Fri, 24 Oct 2025 10:31:02 GMT
Subject: [code-reflection] RFR: [hat] Enable PROFILE_CUDA_KERNEL option
Message-ID: <p9mO9qpCgbSXbrHBQFxvX-FQaeihzJ-ovO6W9PMmqg4=.64c5a55c-bc06-4f3c-9f7b-8f00a26ea7da@github.com>

This PR adds a new option to profiler CUDA kernels (`PROFILE_CUDA_KERNEL`).

How to use it?


HAT=PROFILE_CUDA_KERNEL java -cp hat/job.jar hat.java run ffi-cuda ....

-------------

Commit messages:
 - [hat] Enable PROFILE_CUDA_KERNEL option

Changes: https://git.openjdk.org/babylon/pull/639/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=639&range=00
  Stats: 48 lines in 4 files changed: 30 ins; 2 del; 16 mod
  Patch: https://git.openjdk.org/babylon/pull/639.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/639/head:pull/639

PR: https://git.openjdk.org/babylon/pull/639

From jfumero at openjdk.org  Fri Oct 24 10:38:31 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Fri, 24 Oct 2025 10:38:31 GMT
Subject: [code-reflection] Integrated: [hat] Enable PROFILE_CUDA_KERNEL option
In-Reply-To: <p9mO9qpCgbSXbrHBQFxvX-FQaeihzJ-ovO6W9PMmqg4=.64c5a55c-bc06-4f3c-9f7b-8f00a26ea7da@github.com>
References: <p9mO9qpCgbSXbrHBQFxvX-FQaeihzJ-ovO6W9PMmqg4=.64c5a55c-bc06-4f3c-9f7b-8f00a26ea7da@github.com>
Message-ID: <_1xzpR6jUlh2WFoxc15H916P93RFxwdke2Uioi-XGpo=.7fba1981-3d0d-4a34-8274-d16ac96c40fe@github.com>

On Fri, 24 Oct 2025 10:24:59 GMT, Juan Fumero <jfumero at openjdk.org> wrote:

> This PR adds a new option to profiler CUDA kernels (`PROFILE_CUDA_KERNEL`).
> 
> How to use it?
> 
> 
> HAT=PROFILE_CUDA_KERNEL java -cp hat/job.jar hat.java run ffi-cuda ....

This pull request has now been integrated.

Changeset: 70545418
Author:    Juan Fumero <jfumero at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/70545418e164ff9c6acbe95317ff5649c996400f
Stats:     48 lines in 4 files changed: 30 ins; 2 del; 16 mod

[hat] Enable PROFILE_CUDA_KERNEL option

-------------

PR: https://git.openjdk.org/babylon/pull/639

From jfumero at openjdk.org  Fri Oct 24 10:37:34 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Fri, 24 Oct 2025 10:37:34 GMT
Subject: git: openjdk/babylon: code-reflection: [hat] Enable
 PROFILE_CUDA_KERNEL option
Message-ID: <7fd4ee5f-bc7d-454b-af50-c5efe78fa408@openjdk.org>

Changeset: 70545418
Branch: code-reflection
Author:    Juan Fumero <jfumero at openjdk.org>
Date:      2025-10-24 10:35:12 +0000
URL:       https://git.openjdk.org/babylon/commit/70545418e164ff9c6acbe95317ff5649c996400f

[hat] Enable PROFILE_CUDA_KERNEL option

! hat/backends/ffi/cuda/src/main/native/cpp/cuda_backend.cpp
! hat/backends/ffi/cuda/src/main/native/include/cuda_backend.h
! hat/backends/ffi/shared/src/main/native/include/config.h
! hat/core/src/main/java/hat/Config.java


From jfumero at openjdk.org  Fri Oct 24 11:43:52 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Fri, 24 Oct 2025 11:43:52 GMT
Subject: [code-reflection] Integrated: [hat] config.h in sync with the
 auto-generator
Message-ID: <KN15jgoNuc6_n7sWxDosC8g8kIXQQ8mjywLWVXzUcH8=.f7e922ea-161c-46d2-a6b9-206234e2d8fb@github.com>

`config.h` in sync using the `hat.FFIConfigCreator`

-------------

Commit messages:
 - [hat] config.h in sync with the auto-generator

Changes: https://git.openjdk.org/babylon/pull/640/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=640&range=00
  Stats: 5 lines in 1 file changed: 2 ins; 1 del; 2 mod
  Patch: https://git.openjdk.org/babylon/pull/640.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/640/head:pull/640

PR: https://git.openjdk.org/babylon/pull/640

From jfumero at openjdk.org  Fri Oct 24 11:43:53 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Fri, 24 Oct 2025 11:43:53 GMT
Subject: [code-reflection] Integrated: [hat] config.h in sync with the
 auto-generator
In-Reply-To: <KN15jgoNuc6_n7sWxDosC8g8kIXQQ8mjywLWVXzUcH8=.f7e922ea-161c-46d2-a6b9-206234e2d8fb@github.com>
References: <KN15jgoNuc6_n7sWxDosC8g8kIXQQ8mjywLWVXzUcH8=.f7e922ea-161c-46d2-a6b9-206234e2d8fb@github.com>
Message-ID: <FmbhrxcS9WMTfJFxAdx1oGNp3hW0BE6s6k2My35ed04=.51ff15cb-0ea5-41c5-a99e-a99e9d1ad791@github.com>

On Fri, 24 Oct 2025 11:30:43 GMT, Juan Fumero <jfumero at openjdk.org> wrote:

> `config.h` in sync using the `hat.FFIConfigCreator`

This pull request has now been integrated.

Changeset: e0fe71a4
Author:    Juan Fumero <jfumero at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/e0fe71a411d7036aea23e97abfb24b28f405d922
Stats:     5 lines in 1 file changed: 2 ins; 1 del; 2 mod

[hat] config.h in sync with the auto-generator

-------------

PR: https://git.openjdk.org/babylon/pull/640

From jfumero at openjdk.org  Fri Oct 24 11:44:17 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Fri, 24 Oct 2025 11:44:17 GMT
Subject: git: openjdk/babylon: code-reflection: [hat] config.h in sync with the
 auto-generator
Message-ID: <fc5ed0a4-7bf3-4389-aa06-5c0ada0fdf24@openjdk.org>

Changeset: e0fe71a4
Branch: code-reflection
Author:    Juan Fumero <jfumero at openjdk.org>
Date:      2025-10-24 11:40:40 +0000
URL:       https://git.openjdk.org/babylon/commit/e0fe71a411d7036aea23e97abfb24b28f405d922

[hat] config.h in sync with the auto-generator

! hat/backends/ffi/shared/src/main/native/include/config.h


From duke at openjdk.org  Fri Oct 24 11:58:36 2025
From: duke at openjdk.org (Ruby Chen)
Date: Fri, 24 Oct 2025 11:58:36 GMT
Subject: [code-reflection] Integrated: Update buffer tagging with arrayview
 support
In-Reply-To: <Q9zdEEQTILTtqvItQQ26aC5TCXabT0YwpnlNFFY-fZg=.e0f9b877-1021-4d2f-b8e0-b687928a61dc@github.com>
References: <Q9zdEEQTILTtqvItQQ26aC5TCXabT0YwpnlNFFY-fZg=.e0f9b877-1021-4d2f-b8e0-b687928a61dc@github.com>
Message-ID: <MbfcK9t9nDNQhvRcenALnuEGfsSe59QZUFg4KtryDx8=.ad6ead91-5239-4f30-aabf-801deee98b96@github.com>

On Wed, 15 Oct 2025 18:50:16 GMT, Ruby Chen <duke at openjdk.org> wrote:

> Add buffer tagging for array views, and also use `elements()` instead of `traverse()`.

This pull request has now been integrated.

Changeset: 255b62fb
Author:    Ruby Chen <ruby.r.chen at gmail.com>
Committer: Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/255b62fb74a18c2a79abef31f2051909dee1cbac
Stats:     39 lines in 2 files changed: 17 ins; 3 del; 19 mod

Update buffer tagging with arrayview support

-------------

PR: https://git.openjdk.org/babylon/pull/610

From gfrost at openjdk.org  Fri Oct 24 12:02:45 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Fri, 24 Oct 2025 12:02:45 GMT
Subject: git: openjdk/babylon: code-reflection: Update buffer tagging with
 arrayview support
Message-ID: <b9d32b8e-60af-423c-ab3b-d7b8c02b2b3c@openjdk.org>

Changeset: 255b62fb
Branch: code-reflection
Author:    Ruby Chen <ruby.r.chen at gmail.com>
Committer: Gary Frost <gfrost at openjdk.org>
Date:      2025-10-24 11:56:08 +0000
URL:       https://git.openjdk.org/babylon/commit/255b62fb74a18c2a79abef31f2051909dee1cbac

Update buffer tagging with arrayview support

! hat/core/src/main/java/hat/BufferTagger.java
! hat/examples/arrayview/src/main/java/arrayview/Main.java


From asotona at openjdk.org  Fri Oct 24 13:44:05 2025
From: asotona at openjdk.org (Adam Sotona)
Date: Fri, 24 Oct 2025 13:44:05 GMT
Subject: [code-reflection] RFR: OnnxLift prototype [v2]
In-Reply-To: <dFnT1a6oJdYYzER2-cOVWCIgNce9uh0mBloQ3T52EDc=.612309ef-e96a-4a48-9223-3cafbda9663d@github.com>
References: <dFnT1a6oJdYYzER2-cOVWCIgNce9uh0mBloQ3T52EDc=.612309ef-e96a-4a48-9223-3cafbda9663d@github.com>
Message-ID: <CpW2Uj-UpWUvVqiFjoawVzRvSCInN00wrHM57IblllM=.f95cf533-e6e0-43de-9a20-e2dac98fb249@github.com>

> This is a prototype of OnnxLift tool.
> It lifts ONNX model binary to ONNX code reflection model, extracts weights, and generates Java model source.

Adam Sotona has updated the pull request incrementally with one additional commit since the last revision:

  missing annotation

-------------

Changes:
  - all: https://git.openjdk.org/babylon/pull/638/files
  - new: https://git.openjdk.org/babylon/pull/638/files/43bb1088..003e0d5d

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=babylon&pr=638&range=01
 - incr: https://webrevs.openjdk.org/?repo=babylon&pr=638&range=00-01

  Stats: 2 lines in 1 file changed: 2 ins; 0 del; 0 mod
  Patch: https://git.openjdk.org/babylon/pull/638.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/638/head:pull/638

PR: https://git.openjdk.org/babylon/pull/638

From gfrost at openjdk.org  Fri Oct 24 15:30:34 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Fri, 24 Oct 2025 15:30:34 GMT
Subject: git: openjdk/babylon: code-reflection: Hat util regex
Message-ID: <832f5cc1-4b6f-4e86-84e0-f98716f77626@openjdk.org>

Changeset: 801d7368
Branch: code-reflection
Author:    Gary Frost <gfrost at openjdk.org>
Date:      2025-10-24 15:28:45 +0000
URL:       https://git.openjdk.org/babylon/commit/801d7368218e6e17c072c64c31da4a17ae3210dc

Hat util regex

+ hat/core/src/main/java/hat/util/Regex.java
! hat/core/src/main/java/hat/util/StreamMutable.java
! hat/examples/experiments/src/main/java/experiments/PrefixSum.java
! hat/examples/view/src/main/java/view/EliteMeshReader.java
! hat/examples/view/src/main/java/view/Main.java
- hat/examples/view/src/main/java/view/Regex.java
+ hat/hat/ffisync


From gfrost at openjdk.org  Fri Oct 24 15:31:38 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Fri, 24 Oct 2025 15:31:38 GMT
Subject: [code-reflection] Integrated: Hat util regex
Message-ID: <RYQOEBe5iGw6hojrMQ18vV6X793tFO8VAK37D_Um4gI=.30763312-1ed0-403d-ab8a-cbc2f6c9a671@github.com>

Added a wrapper around java Pattern regex matcher which is easier to use (instanceof patterns) 

Result of matcher generates distict classes (FAIL vs PASS) so we can pattren switch more easily. 

Also added hat/ffisync

-------------

Commit messages:
 - Add hat.util.Regex and hat/ffisync
 - Add hat.util.Regex and hat/ffisync

Changes: https://git.openjdk.org/babylon/pull/641/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=641&range=00
  Stats: 309 lines in 7 files changed: 153 ins; 115 del; 41 mod
  Patch: https://git.openjdk.org/babylon/pull/641.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/641/head:pull/641

PR: https://git.openjdk.org/babylon/pull/641

From gfrost at openjdk.org  Fri Oct 24 15:31:39 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Fri, 24 Oct 2025 15:31:39 GMT
Subject: [code-reflection] Integrated: Hat util regex
In-Reply-To: <RYQOEBe5iGw6hojrMQ18vV6X793tFO8VAK37D_Um4gI=.30763312-1ed0-403d-ab8a-cbc2f6c9a671@github.com>
References: <RYQOEBe5iGw6hojrMQ18vV6X793tFO8VAK37D_Um4gI=.30763312-1ed0-403d-ab8a-cbc2f6c9a671@github.com>
Message-ID: <9o2NXfH5Ubiotk0fJtTD6i7QY2D9Y04P_9SAAct_3k8=.afb352d6-ea21-4657-8672-1e2d20e80d18@github.com>

On Fri, 24 Oct 2025 15:25:23 GMT, Gary Frost <gfrost at openjdk.org> wrote:

> Added a wrapper around java Pattern regex matcher which is easier to use (instanceof patterns) 
> 
> Result of matcher generates distict classes (FAIL vs PASS) so we can pattren switch more easily. 
> 
> Also added hat/ffisync

This pull request has now been integrated.

Changeset: 801d7368
Author:    Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/801d7368218e6e17c072c64c31da4a17ae3210dc
Stats:     309 lines in 7 files changed: 153 ins; 115 del; 41 mod

Hat util regex

-------------

PR: https://git.openjdk.org/babylon/pull/641

From gfrost at openjdk.org  Fri Oct 24 17:22:17 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Fri, 24 Oct 2025 17:22:17 GMT
Subject: [code-reflection] Integrated: Coverted enum base view parser to iface
 based
Message-ID: <GqdsAjkWoFqFJA7l0sJ2zIQhy6aFc87mreC5rXo3IlU=.2b735ca5-c8a0-4e5e-92a8-192632676a5a@github.com>

Simplified view parser.  

Also increased number of vertices in view demo

-------------

Commit messages:
 - Coverted enum base view parser to iface based

Changes: https://git.openjdk.org/babylon/pull/643/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=643&range=00
  Stats: 74 lines in 6 files changed: 20 ins; 21 del; 33 mod
  Patch: https://git.openjdk.org/babylon/pull/643.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/643/head:pull/643

PR: https://git.openjdk.org/babylon/pull/643

From gfrost at openjdk.org  Fri Oct 24 17:22:22 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Fri, 24 Oct 2025 17:22:22 GMT
Subject: git: openjdk/babylon: code-reflection: Coverted enum base view parser
 to iface based
Message-ID: <30e349fa-8aca-444a-b04f-4fd44f7dea63@openjdk.org>

Changeset: 873d7965
Branch: code-reflection
Author:    Gary Frost <gfrost at openjdk.org>
Date:      2025-10-24 17:18:19 +0000
URL:       https://git.openjdk.org/babylon/commit/873d79657c907cdd7993dd226f87e150444c06b3

Coverted enum base view parser to iface based

! hat/examples/view/src/main/java/view/EliteMeshReader.java
! hat/examples/view/src/main/java/view/Main.java
! hat/examples/view/src/main/java/view/f32/F32Triangle3D.java
! hat/examples/view/src/main/java/view/f32/F32Vec3.java
! hat/examples/view/src/main/java/view/i32/I32Triangle2D.java
! hat/examples/view/src/main/java/view/i32/I32Vec2.java


From gfrost at openjdk.org  Fri Oct 24 17:22:17 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Fri, 24 Oct 2025 17:22:17 GMT
Subject: [code-reflection] Integrated: Coverted enum base view parser to iface
 based
In-Reply-To: <GqdsAjkWoFqFJA7l0sJ2zIQhy6aFc87mreC5rXo3IlU=.2b735ca5-c8a0-4e5e-92a8-192632676a5a@github.com>
References: <GqdsAjkWoFqFJA7l0sJ2zIQhy6aFc87mreC5rXo3IlU=.2b735ca5-c8a0-4e5e-92a8-192632676a5a@github.com>
Message-ID: <WjswZNOAxwxZg1xQGp2mM8xRHPpzzb2DoLPjyVWkUBg=.95790d6f-02fe-4685-af65-c69f685d04e8@github.com>

On Fri, 24 Oct 2025 16:58:38 GMT, Gary Frost <gfrost at openjdk.org> wrote:

> Simplified view parser.  
> 
> Also increased number of vertices in view demo

This pull request has now been integrated.

Changeset: 873d7965
Author:    Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/873d79657c907cdd7993dd226f87e150444c06b3
Stats:     74 lines in 6 files changed: 20 ins; 21 del; 33 mod

Coverted enum base view parser to iface based

-------------

PR: https://git.openjdk.org/babylon/pull/643

From jfumero at openjdk.org  Mon Oct 27 10:10:18 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Mon, 27 Oct 2025 10:10:18 GMT
Subject: [code-reflection] RFR: Proposal for supporting Float4.of for CPU host
 code and GPU device code
Message-ID: <paNHesfhOTHU4U8SWowRhNWUroeVKz9GA4SJVTaimnI=.23d8e3d6-26a5-4ff4-901e-200306e5ad7e@github.com>

This PR adds support for creating float4 objects within the device code:


Float4 f = Float4.of(x, y, z ,w)


In addition, it provides a CPU implementation for this type a new method to operate in `lanewise`. This operation is CPU only at the moment:


Float4 vA ... 
Float4 vB ... 
vA.linewise(vB, (a, b) -> Float::sum);

-------------

Commit messages:
 - [hat] minor change
 - [hat] makeFloat4 from immutable supported on GPU
 - [hat][wip] float4 implementation with mutable/immutable variants
 - [hat] Float4 views docs
 - Merge branch 'code-reflection' into hat/float4/of
 - [hat] CUDA backend Float4.of supported
 - [hat] fix dialectify Float.of phase
 - [hat] Float4.of supported for OpenCL

Changes: https://git.openjdk.org/babylon/pull/642/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=642&range=00
  Stats: 604 lines in 26 files changed: 525 ins; 15 del; 64 mod
  Patch: https://git.openjdk.org/babylon/pull/642.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/642/head:pull/642

PR: https://git.openjdk.org/babylon/pull/642

From jfumero at openjdk.org  Mon Oct 27 10:36:00 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Mon, 27 Oct 2025 10:36:00 GMT
Subject: [code-reflection] RFR: Proposal for supporting Float4.of for CPU
 host code and GPU device code [v2]
In-Reply-To: <paNHesfhOTHU4U8SWowRhNWUroeVKz9GA4SJVTaimnI=.23d8e3d6-26a5-4ff4-901e-200306e5ad7e@github.com>
References: <paNHesfhOTHU4U8SWowRhNWUroeVKz9GA4SJVTaimnI=.23d8e3d6-26a5-4ff4-901e-200306e5ad7e@github.com>
Message-ID: <cEHwJWibH8lS8i8TOIhATCeQMJ0SuAEX9wcJ0H1LB2k=.f07b2218-1bd2-4479-abf0-fc0552986dde@github.com>

> This PR adds support for creating float4 objects within the device code:
> 
> 
> Float4 f = Float4.of(x, y, z ,w)
> 
> 
> In addition, it provides a CPU implementation for this type a new method to operate in `lanewise`. This operation is CPU only at the moment:
> 
> 
> Float4 vA ... 
> Float4 vB ... 
> vA.linewise(vB, (a, b) -> Float::sum);

Juan Fumero has updated the pull request incrementally with one additional commit since the last revision:

  [hat] Move vectorOf codegen method to common C99 codegen

-------------

Changes:
  - all: https://git.openjdk.org/babylon/pull/642/files
  - new: https://git.openjdk.org/babylon/pull/642/files/31e0db21..2f954e0b

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=babylon&pr=642&range=01
 - incr: https://webrevs.openjdk.org/?repo=babylon&pr=642&range=00-01

  Stats: 80 lines in 4 files changed: 24 ins; 50 del; 6 mod
  Patch: https://git.openjdk.org/babylon/pull/642.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/642/head:pull/642

PR: https://git.openjdk.org/babylon/pull/642

From jfumero at openjdk.org  Mon Oct 27 10:39:35 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Mon, 27 Oct 2025 10:39:35 GMT
Subject: [code-reflection] RFR: Proposal for supporting Float4.of for CPU
 host code and GPU device code [v3]
In-Reply-To: <paNHesfhOTHU4U8SWowRhNWUroeVKz9GA4SJVTaimnI=.23d8e3d6-26a5-4ff4-901e-200306e5ad7e@github.com>
References: <paNHesfhOTHU4U8SWowRhNWUroeVKz9GA4SJVTaimnI=.23d8e3d6-26a5-4ff4-901e-200306e5ad7e@github.com>
Message-ID: <0I0b7tFu9M8W_Q2Vk1yBKYXYp6yeS_U8G-jQjJSLiPs=.3a185c30-c0a3-49e2-89be-e14d4ed7fb83@github.com>

> This PR adds support for creating float4 objects within the device code:
> 
> 
> Float4 f = Float4.of(x, y, z ,w)
> 
> 
> In addition, it provides a CPU implementation for this type a new method to operate in `lanewise`. This operation is CPU only at the moment:
> 
> 
> Float4 vA ... 
> Float4 vB ... 
> vA.linewise(vB, (a, b) -> Float::sum);

Juan Fumero has updated the pull request incrementally with one additional commit since the last revision:

  [hat] codegen - cleanup old OpenCL typecasts

-------------

Changes:
  - all: https://git.openjdk.org/babylon/pull/642/files
  - new: https://git.openjdk.org/babylon/pull/642/files/2f954e0b..e9864a77

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=babylon&pr=642&range=02
 - incr: https://webrevs.openjdk.org/?repo=babylon&pr=642&range=01-02

  Stats: 5 lines in 2 files changed: 0 ins; 5 del; 0 mod
  Patch: https://git.openjdk.org/babylon/pull/642.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/642/head:pull/642

PR: https://git.openjdk.org/babylon/pull/642

From jfumero at openjdk.org  Mon Oct 27 11:17:30 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Mon, 27 Oct 2025 11:17:30 GMT
Subject: [code-reflection] RFR: Proposal for supporting Float4.of for CPU
 host code and GPU device code [v4]
In-Reply-To: <paNHesfhOTHU4U8SWowRhNWUroeVKz9GA4SJVTaimnI=.23d8e3d6-26a5-4ff4-901e-200306e5ad7e@github.com>
References: <paNHesfhOTHU4U8SWowRhNWUroeVKz9GA4SJVTaimnI=.23d8e3d6-26a5-4ff4-901e-200306e5ad7e@github.com>
Message-ID: <VbJFsRvSKv56SZzoXzElavPoe0VdmpCwF1af2R7IAcc=.1b5b27f9-a470-4653-9ec2-f17ef69a8dbb@github.com>

> This PR adds support for creating float4 objects within the device code:
> 
> 
> Float4 f = Float4.of(x, y, z ,w)
> 
> 
> In addition, it provides a CPU implementation for this type a new method to operate in `lanewise`. This operation is CPU only at the moment:
> 
> 
> Float4 vA ... 
> Float4 vB ... 
> vA.linewise(vB, (a, b) -> Float::sum);

Juan Fumero has updated the pull request incrementally with one additional commit since the last revision:

  [hat] minor refactor ocl-codegen

-------------

Changes:
  - all: https://git.openjdk.org/babylon/pull/642/files
  - new: https://git.openjdk.org/babylon/pull/642/files/e9864a77..fd9c6d3c

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=babylon&pr=642&range=03
 - incr: https://webrevs.openjdk.org/?repo=babylon&pr=642&range=02-03

  Stats: 11 lines in 4 files changed: 7 ins; 2 del; 2 mod
  Patch: https://git.openjdk.org/babylon/pull/642.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/642/head:pull/642

PR: https://git.openjdk.org/babylon/pull/642

From jfumero at openjdk.org  Mon Oct 27 11:38:56 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Mon, 27 Oct 2025 11:38:56 GMT
Subject: [code-reflection] RFR: Proposal for supporting Float4.of for CPU
 host code and GPU device code [v5]
In-Reply-To: <paNHesfhOTHU4U8SWowRhNWUroeVKz9GA4SJVTaimnI=.23d8e3d6-26a5-4ff4-901e-200306e5ad7e@github.com>
References: <paNHesfhOTHU4U8SWowRhNWUroeVKz9GA4SJVTaimnI=.23d8e3d6-26a5-4ff4-901e-200306e5ad7e@github.com>
Message-ID: <IP5E92y2ywegzrIziXoHaS-DQlSRPxqjlNe--llqqbc=.dca9563d-041a-44d3-9274-5609e583908a@github.com>

> This PR adds support for creating float4 objects within the device code:
> 
> 
> Float4 f = Float4.of(x, y, z ,w)
> 
> 
> In addition, it provides a CPU implementation for this type a new method to operate in `lanewise`. This operation is CPU only at the moment:
> 
> 
> Float4 vA ... 
> Float4 vB ... 
> vA.linewise(vB, (a, b) -> Float::sum);

Juan Fumero has updated the pull request incrementally with one additional commit since the last revision:

  [hat] Dialect for handling vector types simplified

-------------

Changes:
  - all: https://git.openjdk.org/babylon/pull/642/files
  - new: https://git.openjdk.org/babylon/pull/642/files/fd9c6d3c..3c45324c

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=babylon&pr=642&range=04
 - incr: https://webrevs.openjdk.org/?repo=babylon&pr=642&range=03-04

  Stats: 109 lines in 12 files changed: 23 ins; 69 del; 17 mod
  Patch: https://git.openjdk.org/babylon/pull/642.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/642/head:pull/642

PR: https://git.openjdk.org/babylon/pull/642

From jfumero at openjdk.org  Mon Oct 27 12:07:16 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Mon, 27 Oct 2025 12:07:16 GMT
Subject: [code-reflection] RFR: Proposal for supporting Float4.of for CPU
 host code and GPU device code [v6]
In-Reply-To: <paNHesfhOTHU4U8SWowRhNWUroeVKz9GA4SJVTaimnI=.23d8e3d6-26a5-4ff4-901e-200306e5ad7e@github.com>
References: <paNHesfhOTHU4U8SWowRhNWUroeVKz9GA4SJVTaimnI=.23d8e3d6-26a5-4ff4-901e-200306e5ad7e@github.com>
Message-ID: <8kf36Fz0oyJOMfXQ9pPUpHdvgG9WtDy0V5vZCM-NHXE=.7ad01b44-d0ef-4fc3-96e9-772bfcb27062@github.com>

> This PR adds support for creating float4 objects within the device code:
> 
> 
> Float4 f = Float4.of(x, y, z ,w)
> 
> 
> In addition, it provides a CPU implementation for this type a new method to operate in `lanewise`. This operation is CPU only at the moment:
> 
> 
> Float4 vA ... 
> Float4 vB ... 
> vA.linewise(vB, (a, b) -> Float::sum);

Juan Fumero has updated the pull request incrementally with one additional commit since the last revision:

  [hat] Log Config bits for OpenCL moved to OpenCL backend instantiation

-------------

Changes:
  - all: https://git.openjdk.org/babylon/pull/642/files
  - new: https://git.openjdk.org/babylon/pull/642/files/3c45324c..251f1a52

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=babylon&pr=642&range=05
 - incr: https://webrevs.openjdk.org/?repo=babylon&pr=642&range=04-05

  Stats: 9 lines in 1 file changed: 5 ins; 4 del; 0 mod
  Patch: https://git.openjdk.org/babylon/pull/642.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/642/head:pull/642

PR: https://git.openjdk.org/babylon/pull/642

From asotona at openjdk.org  Mon Oct 27 16:19:16 2025
From: asotona at openjdk.org (Adam Sotona)
Date: Mon, 27 Oct 2025 16:19:16 GMT
Subject: [code-reflection] RFR: OnnxLift prototype [v3]
In-Reply-To: <dFnT1a6oJdYYzER2-cOVWCIgNce9uh0mBloQ3T52EDc=.612309ef-e96a-4a48-9223-3cafbda9663d@github.com>
References: <dFnT1a6oJdYYzER2-cOVWCIgNce9uh0mBloQ3T52EDc=.612309ef-e96a-4a48-9223-3cafbda9663d@github.com>
Message-ID: <bT5yxqlvPZbK2Q3VBmFPju5EXF8Y9qvqpslqfSz1scg=.1da5bc96-12d7-4929-b430-59d378145a73@github.com>

> This is a prototype of OnnxLift tool.
> It lifts ONNX model binary to ONNX code reflection model, extracts weights, and generates Java model source.

Adam Sotona has updated the pull request incrementally with one additional commit since the last revision:

  refactoring split to JavaTemplate

-------------

Changes:
  - all: https://git.openjdk.org/babylon/pull/638/files
  - new: https://git.openjdk.org/babylon/pull/638/files/003e0d5d..4aa3ce64

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=babylon&pr=638&range=02
 - incr: https://webrevs.openjdk.org/?repo=babylon&pr=638&range=01-02

  Stats: 718 lines in 2 files changed: 381 ins; 323 del; 14 mod
  Patch: https://git.openjdk.org/babylon/pull/638.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/638/head:pull/638

PR: https://git.openjdk.org/babylon/pull/638

From asotona at openjdk.org  Mon Oct 27 16:27:26 2025
From: asotona at openjdk.org (Adam Sotona)
Date: Mon, 27 Oct 2025 16:27:26 GMT
Subject: [code-reflection] RFR: OnnxLift prototype [v4]
In-Reply-To: <dFnT1a6oJdYYzER2-cOVWCIgNce9uh0mBloQ3T52EDc=.612309ef-e96a-4a48-9223-3cafbda9663d@github.com>
References: <dFnT1a6oJdYYzER2-cOVWCIgNce9uh0mBloQ3T52EDc=.612309ef-e96a-4a48-9223-3cafbda9663d@github.com>
Message-ID: <NYESeEQ2aHnJbSin1eLWz9BkA369dzYTrApDrR6fHeo=.17e084e4-55ae-4d7c-922f-de6f122b279b@github.com>

> This is a prototype of OnnxLift tool.
> It lifts ONNX model binary to ONNX code reflection model, extracts weights, and generates Java model source.

Adam Sotona has updated the pull request incrementally with one additional commit since the last revision:

  added class name cmd option

-------------

Changes:
  - all: https://git.openjdk.org/babylon/pull/638/files
  - new: https://git.openjdk.org/babylon/pull/638/files/4aa3ce64..bf54e820

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=babylon&pr=638&range=03
 - incr: https://webrevs.openjdk.org/?repo=babylon&pr=638&range=02-03

  Stats: 9 lines in 3 files changed: 2 ins; 0 del; 7 mod
  Patch: https://git.openjdk.org/babylon/pull/638.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/638/head:pull/638

PR: https://git.openjdk.org/babylon/pull/638

From jnett96 at gmail.com  Mon Oct 27 18:47:17 2025
From: jnett96 at gmail.com (Ryan Nett)
Date: Mon, 27 Oct 2025 11:47:17 -0700
Subject: Babylon and MLIR
Message-ID: <CABrFf6jgFE7_oZWxoU3aAwP6Jsn1CyuCfk+=xkfuwkuPwVLHSw@mail.gmail.com>

Hi folks,

I've seen Babylon's code model's similarity to MLIR be mentioned in a few
conference talks, and was wondering if there were any plans to support MLIR
as a "destination" via a 1st party library or something like HAT?  There
are several existing projects that take MLIR as input and compile it to
optimized device code (e.g. IREE, XLA, this list
<https://mlir.llvm.org/users/>), so if Java code could easily be converted
to MLIR it could open up many integrations and support for additional
devices without requiring any work on the Java side.

Thanks,
Ryan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.org/pipermail/babylon-dev/attachments/20251027/708cd652/attachment.htm>

From duke at openjdk.org  Mon Oct 27 21:00:26 2025
From: duke at openjdk.org (Ana Maria Mihalceanu)
Date: Mon, 27 Oct 2025 21:00:26 GMT
Subject: [code-reflection] RFR: Upgrade ONNX runtime to use API version 1.23.1
Message-ID: <ogjLjVBV2D868E0DIzs1f9fP4806ecVBz4nhwtDN5Jg=.86eabea2-9283-453f-956a-e01a3014df01@github.com>

This PR contains the following:

- [x] Upgrade to version 1.23.1 ONNX runtime API.
- [x] Refactoring of the python script `AttributeProto` is now in `onnx` (I used version `1.19.1` which is the latest)
- [x] Regeneration of tensors and tensor operations.
- [x] Regeneration of the runtime API based on `onnxruntime_c_api.h`.
- [x] Appends execution provider options to runtime, as shown in the new API,
- [x] Allows users to append execution providers like `WebGpuExecutionProvider` and `CpuExecutionProvider` through `OnnxRuntime.appendExecutionProvider_V2` method.

-------------

Commit messages:
 - Update full name
 - Attempt to fix the the trailing whitespaces from generation.
 - Merge remote-tracking branch 'origin/upgrade-onnx-runtime' into upgrade-onnx-runtime
 - Merge branch 'openjdk:code-reflection' into upgrade-onnx-runtime
 - Add implementation for custom Execution Provider libraries.
 - Fix compilation errors and minimize friction with changes in OnnxRuntime.
 - Upgrade ONNX prototype to use version 1.23.1 of the ONNX runtime C library.

Changes: https://git.openjdk.org/babylon/pull/645/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=645&range=00
  Stats: 49849 lines in 41 files changed: 42947 ins; 3317 del; 3585 mod
  Patch: https://git.openjdk.org/babylon/pull/645.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/645/head:pull/645

PR: https://git.openjdk.org/babylon/pull/645

From duke at openjdk.org  Mon Oct 27 21:49:18 2025
From: duke at openjdk.org (Ana Maria Mihalceanu)
Date: Mon, 27 Oct 2025 21:49:18 GMT
Subject: [code-reflection] RFR: Upgrade ONNX runtime to use API version
 1.23.1 [v2]
In-Reply-To: <ogjLjVBV2D868E0DIzs1f9fP4806ecVBz4nhwtDN5Jg=.86eabea2-9283-453f-956a-e01a3014df01@github.com>
References: <ogjLjVBV2D868E0DIzs1f9fP4806ecVBz4nhwtDN5Jg=.86eabea2-9283-453f-956a-e01a3014df01@github.com>
Message-ID: <jDPR5m6_wYKJ67BgwDGKeunBpvhTtTk89dpw569RmBk=.a0efb365-3942-4fa5-bd2f-402df045800d@github.com>

> This PR contains the following:
> 
> - [x] Upgrade to version 1.23.1 ONNX runtime API.
> - [x] Refactoring of the python script as `AttributeProto` is now in `onnx` (I used version `1.19.1` which is the latest)
> - [x] Regeneration of tensors and tensor operations.
> - [x] Regeneration of the runtime API based on `onnxruntime_c_api.h`.
> - [x] Appends execution provider options to runtime, as shown in the `[new API](https://onnxruntime.ai/docs/execution-providers/CoreML-ExecutionProvider.html)`. 
> - [x] Allows users to append execution providers like `WebGpuExecutionProvider` and `CpuExecutionProvider` through `OnnxRuntime.appendExecutionProvider_V2` method.

Ana Maria Mihalceanu has updated the pull request incrementally with one additional commit since the last revision:

  Add enable on subgraphs for more speed.
  
  Signed-off-by: ammbra <mihalceanuana at gmail.com>

-------------

Changes:
  - all: https://git.openjdk.org/babylon/pull/645/files
  - new: https://git.openjdk.org/babylon/pull/645/files/b4c00fc4..fbe9869b

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=babylon&pr=645&range=01
 - incr: https://webrevs.openjdk.org/?repo=babylon&pr=645&range=00-01

  Stats: 1 line in 1 file changed: 0 ins; 0 del; 1 mod
  Patch: https://git.openjdk.org/babylon/pull/645.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/645/head:pull/645

PR: https://git.openjdk.org/babylon/pull/645

From duke at openjdk.org  Mon Oct 27 21:59:52 2025
From: duke at openjdk.org (Ana Maria Mihalceanu)
Date: Mon, 27 Oct 2025 21:59:52 GMT
Subject: [code-reflection] RFR: Upgrade ONNX runtime to use API version
 1.23.1 [v3]
In-Reply-To: <ogjLjVBV2D868E0DIzs1f9fP4806ecVBz4nhwtDN5Jg=.86eabea2-9283-453f-956a-e01a3014df01@github.com>
References: <ogjLjVBV2D868E0DIzs1f9fP4806ecVBz4nhwtDN5Jg=.86eabea2-9283-453f-956a-e01a3014df01@github.com>
Message-ID: <zE7HVzNTY6U9gDdl4NmvUZWBks7W4VJMBcCqLYj4XfY=.ab2c502c-a8ad-4b92-a558-19d46eef62bb@github.com>

> This PR contains the following:
> 
> - [x] Upgrade to version 1.23.1 ONNX runtime API.
> - [x] Refactoring of the python script as `AttributeProto` is now in `onnx` (I used version `1.19.1` which is the latest)
> - [x] Regeneration of tensors and tensor operations.
> - [x] Regeneration of the runtime API based on `onnxruntime_c_api.h`.
> - [x] Appends execution provider options to runtime, as shown in the `[new API](https://onnxruntime.ai/docs/execution-providers/CoreML-ExecutionProvider.html)`. 
> - [x] Allows users to append execution providers like `WebGpuExecutionProvider` and `CpuExecutionProvider` through `OnnxRuntime.appendExecutionProvider_V2` method.

Ana Maria Mihalceanu has updated the pull request incrementally with one additional commit since the last revision:

  Allow low-precision accumulation on a GPU.
  
  Signed-off-by: ammbra <mihalceanuana at gmail.com>

-------------

Changes:
  - all: https://git.openjdk.org/babylon/pull/645/files
  - new: https://git.openjdk.org/babylon/pull/645/files/fbe9869b..1d67ef6a

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=babylon&pr=645&range=02
 - incr: https://webrevs.openjdk.org/?repo=babylon&pr=645&range=01-02

  Stats: 4 lines in 1 file changed: 2 ins; 1 del; 1 mod
  Patch: https://git.openjdk.org/babylon/pull/645.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/645/head:pull/645

PR: https://git.openjdk.org/babylon/pull/645

From paul.sandoz at oracle.com  Mon Oct 27 22:02:22 2025
From: paul.sandoz at oracle.com (Paul Sandoz)
Date: Mon, 27 Oct 2025 22:02:22 +0000
Subject: Babylon and MLIR
In-Reply-To: <CABrFf6jgFE7_oZWxoU3aAwP6Jsn1CyuCfk+=xkfuwkuPwVLHSw@mail.gmail.com>
References: <CABrFf6jgFE7_oZWxoU3aAwP6Jsn1CyuCfk+=xkfuwkuPwVLHSw@mail.gmail.com>
Message-ID: <725880CC-2AD5-4320-9A8A-BFD9BB8E19FA@oracle.com>

Hi Ryan,

We have been inspired by the design of MLIR, partly (in theory) to make it easier to translate to MLIR and therefore easier to leverage native MLIR compiler toolchains, just like you indicate.

We did an experiment translating Java code to the Triton MLIR dialect and inputing into Intel?s SPIRV compiler pipeline to generate GPU kernels. That?s as far as we have explored things. It needs to be proved out more, but we don?t have any plans to do more at the moment as we (as in us folks at Oracle) are maxed out focusing on other areas. I was hoping some folks in the community might be interested in further experiments.

HAT could be a very an interesting area to focus on e.g., a HAT accelerator that generates MLIR kernels from Java code in the HAT programming model.

Paul.



On Oct 27, 2025, at 11:47?AM, Ryan Nett <jnett96 at gmail.com> wrote:

Hi folks,

I've seen Babylon's code model's similarity to MLIR be mentioned in a few conference talks, and was wondering if there were any plans to support MLIR as a "destination" via a 1st party library or something like HAT?  There are several existing projects that take MLIR as input and compile it to optimized device code (e.g. IREE, XLA, this list<https://mlir.llvm.org/users/>), so if Java code could easily be converted to MLIR it could open up many integrations and support for additional devices without requiring any work on the Java side.

Thanks,
Ryan

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.org/pipermail/babylon-dev/attachments/20251027/880db96f/attachment.htm>

From asotona at openjdk.org  Wed Oct 29 11:45:51 2025
From: asotona at openjdk.org (Adam Sotona)
Date: Wed, 29 Oct 2025 11:45:51 GMT
Subject: [code-reflection] RFR: Fixed BytecodeGenerator INDY calls with invalid
 interface method type
Message-ID: <EgfeLfObILtFd8Gzk41ReEkj4fRTT7VUOPdS-0Bn7bU=.ef281405-f03d-42e6-bd20-5341bfa92924@github.com>

`BytecodeGenerator` provided INDY calls with invalid `interfaceMethodType` when it differs from the actual `dynamicMethodType`.

-------------

Commit messages:
 - Fixed BytecodeGenerator INDY

Changes: https://git.openjdk.org/babylon/pull/648/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=648&range=00
  Stats: 29 lines in 2 files changed: 17 ins; 0 del; 12 mod
  Patch: https://git.openjdk.org/babylon/pull/648.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/648/head:pull/648

PR: https://git.openjdk.org/babylon/pull/648

From psandoz at openjdk.org  Wed Oct 29 16:31:58 2025
From: psandoz at openjdk.org (Paul Sandoz)
Date: Wed, 29 Oct 2025 16:31:58 GMT
Subject: [code-reflection] RFR: Fixed BytecodeGenerator INDY calls with
 invalid interface method type
In-Reply-To: <EgfeLfObILtFd8Gzk41ReEkj4fRTT7VUOPdS-0Bn7bU=.ef281405-f03d-42e6-bd20-5341bfa92924@github.com>
References: <EgfeLfObILtFd8Gzk41ReEkj4fRTT7VUOPdS-0Bn7bU=.ef281405-f03d-42e6-bd20-5341bfa92924@github.com>
Message-ID: <g7vlfiptChmLxRpALpdllNbne_dzMDe9oc83wQcyAU0=.b288589c-367a-4b81-b8b5-da0b21ed66d5@github.com>

On Wed, 29 Oct 2025 11:40:01 GMT, Adam Sotona <asotona at openjdk.org> wrote:

> `BytecodeGenerator` provided INDY calls with invalid `interfaceMethodType` when it differs from the actual `dynamicMethodType`.

Marked as reviewed by psandoz (Lead).

-------------

PR Review: https://git.openjdk.org/babylon/pull/648#pullrequestreview-3394753473

From asotona at openjdk.org  Wed Oct 29 18:13:22 2025
From: asotona at openjdk.org (Adam Sotona)
Date: Wed, 29 Oct 2025 18:13:22 GMT
Subject: git: openjdk/babylon: code-reflection: Fixed BytecodeGenerator INDY
 calls with invalid interface method type
Message-ID: <b9356061-7056-4ef7-9fa3-7f46c4cb21c8@openjdk.org>

Changeset: 2dd513ad
Branch: code-reflection
Author:    Adam Sotona <asotona at openjdk.org>
Date:      2025-10-29 18:09:53 +0000
URL:       https://git.openjdk.org/babylon/commit/2dd513ad74b2c7a005a019f08739c9777ce719a8

Fixed BytecodeGenerator INDY calls with invalid interface method type

Reviewed-by: psandoz

! src/jdk.incubator.code/share/classes/jdk/incubator/code/bytecode/BytecodeGenerator.java
! test/jdk/java/lang/reflect/code/bytecode/TestBytecode.java


From asotona at openjdk.org  Wed Oct 29 18:13:21 2025
From: asotona at openjdk.org (Adam Sotona)
Date: Wed, 29 Oct 2025 18:13:21 GMT
Subject: [code-reflection] RFR: Fixed BytecodeGenerator INDY calls with
 invalid interface method type
In-Reply-To: <EgfeLfObILtFd8Gzk41ReEkj4fRTT7VUOPdS-0Bn7bU=.ef281405-f03d-42e6-bd20-5341bfa92924@github.com>
References: <EgfeLfObILtFd8Gzk41ReEkj4fRTT7VUOPdS-0Bn7bU=.ef281405-f03d-42e6-bd20-5341bfa92924@github.com>
Message-ID: <_DdmlOVPm4gUqqv9rjYqOql0NJqX0GzIZRjG0P3Zvy0=.df2ce599-71e3-4be4-9257-985be4cc77c0@github.com>

On Wed, 29 Oct 2025 11:40:01 GMT, Adam Sotona <asotona at openjdk.org> wrote:

> `BytecodeGenerator` provided INDY calls with invalid `interfaceMethodType` when it differs from the actual `dynamicMethodType`.

Thank you for the review.

-------------

PR Comment: https://git.openjdk.org/babylon/pull/648#issuecomment-3462939183

From asotona at openjdk.org  Wed Oct 29 18:13:22 2025
From: asotona at openjdk.org (Adam Sotona)
Date: Wed, 29 Oct 2025 18:13:22 GMT
Subject: [code-reflection] Integrated: Fixed BytecodeGenerator INDY calls with
 invalid interface method type
In-Reply-To: <EgfeLfObILtFd8Gzk41ReEkj4fRTT7VUOPdS-0Bn7bU=.ef281405-f03d-42e6-bd20-5341bfa92924@github.com>
References: <EgfeLfObILtFd8Gzk41ReEkj4fRTT7VUOPdS-0Bn7bU=.ef281405-f03d-42e6-bd20-5341bfa92924@github.com>
Message-ID: <CsJHUlMpiDP-iDSWp8WZSs81GZMzCO0dTcww0WtS1fg=.3a173c7c-04ce-4efd-bd20-fa7c9c1e64b4@github.com>

On Wed, 29 Oct 2025 11:40:01 GMT, Adam Sotona <asotona at openjdk.org> wrote:

> `BytecodeGenerator` provided INDY calls with invalid `interfaceMethodType` when it differs from the actual `dynamicMethodType`.

This pull request has now been integrated.

Changeset: 2dd513ad
Author:    Adam Sotona <asotona at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/2dd513ad74b2c7a005a019f08739c9777ce719a8
Stats:     29 lines in 2 files changed: 17 ins; 0 del; 12 mod

Fixed BytecodeGenerator INDY calls with invalid interface method type

Reviewed-by: psandoz

-------------

PR: https://git.openjdk.org/babylon/pull/648

From asotona at openjdk.org  Wed Oct 29 18:48:05 2025
From: asotona at openjdk.org (Adam Sotona)
Date: Wed, 29 Oct 2025 18:48:05 GMT
Subject: [code-reflection] RFR: Upgrade ONNX runtime to use API version
 1.23.1 [v3]
In-Reply-To: <zE7HVzNTY6U9gDdl4NmvUZWBks7W4VJMBcCqLYj4XfY=.ab2c502c-a8ad-4b92-a558-19d46eef62bb@github.com>
References: <ogjLjVBV2D868E0DIzs1f9fP4806ecVBz4nhwtDN5Jg=.86eabea2-9283-453f-956a-e01a3014df01@github.com>
 <zE7HVzNTY6U9gDdl4NmvUZWBks7W4VJMBcCqLYj4XfY=.ab2c502c-a8ad-4b92-a558-19d46eef62bb@github.com>
Message-ID: <MYcPCFflprY6f9F4OBDLj12qiiSZI7T5Y2xgm3-f3No=.3dfca7ce-f14b-497f-a115-796ff36167a4@github.com>

On Mon, 27 Oct 2025 21:59:52 GMT, Ana Maria Mihalceanu <duke at openjdk.org> wrote:

>> This PR contains the following:
>> 
>> - [x] Upgrade to version 1.23.1 ONNX runtime API.
>> - [x] Refactoring of the python script as `AttributeProto` is now in `onnx` (I used version `1.19.1` which is the latest)
>> - [x] Regeneration of tensors and tensor operations.
>> - [x] Regeneration of the runtime API based on `onnxruntime_c_api.h`.
>> - [x] Appends execution provider options to runtime, as shown in the [new API](https://onnxruntime.ai/docs/execution-providers/CoreML-ExecutionProvider.html). 
>> - [x] Allows users to append execution providers like `WebGpuExecutionProvider` and `CpuExecutionProvider` through `OnnxRuntime.appendExecutionProvider_V2` method.
>> 
>> CC @LizeRaes
>
> Ana Maria Mihalceanu has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Allow low-precision accumulation on a GPU.
>   
>   Signed-off-by: ammbra <mihalceanuana at gmail.com>

It looks and works great!
I found just some NITs.

cr-examples/onnx/src/main/java/oracle/code/onnx/OnnxProvider.java line 1:

> 1: package oracle.code.onnx;

Missing copyright header

cr-examples/onnx/src/test/java/oracle/code/onnx/fer/FERInference.java line 35:

> 33: import java.lang.foreign.Arena;
> 34: import java.net.URL;
> 35: import java.util.Map;

unused imports?

cr-examples/onnx/src/test/java/oracle/code/onnx/fer/FERInference.java line 39:

> 37: 
> 38: import static oracle.code.onnx.fer.FERCoreMLDemo.IMAGE_SIZE;
> 39: import static oracle.code.onnx.foreign.onnxruntime_c_api_h.*;

unused imports?

-------------

Marked as reviewed by asotona (Reviewer).

PR Review: https://git.openjdk.org/babylon/pull/645#pullrequestreview-3395251346
PR Review Comment: https://git.openjdk.org/babylon/pull/645#discussion_r2474595088
PR Review Comment: https://git.openjdk.org/babylon/pull/645#discussion_r2474677780
PR Review Comment: https://git.openjdk.org/babylon/pull/645#discussion_r2474681806

From duke at openjdk.org  Wed Oct 29 19:24:32 2025
From: duke at openjdk.org (Ana Maria Mihalceanu)
Date: Wed, 29 Oct 2025 19:24:32 GMT
Subject: [code-reflection] RFR: Upgrade ONNX runtime to use API version
 1.23.1 [v4]
In-Reply-To: <ogjLjVBV2D868E0DIzs1f9fP4806ecVBz4nhwtDN5Jg=.86eabea2-9283-453f-956a-e01a3014df01@github.com>
References: <ogjLjVBV2D868E0DIzs1f9fP4806ecVBz4nhwtDN5Jg=.86eabea2-9283-453f-956a-e01a3014df01@github.com>
Message-ID: <7oBiUeRNDc7teZmbg1ONM6l750ABYmGVDtdTiY9w-TY=.f37d6bf5-b2a9-44c6-bf36-e15ca593cd49@github.com>

> This PR contains the following:
> 
> - [x] Upgrade to version 1.23.1 ONNX runtime API.
> - [x] Refactoring of the python script as `AttributeProto` is now in `onnx` (I used version `1.19.1` which is the latest)
> - [x] Regeneration of tensors and tensor operations.
> - [x] Regeneration of the runtime API based on `onnxruntime_c_api.h`.
> - [x] Appends execution provider options to runtime, as shown in the [new API](https://onnxruntime.ai/docs/execution-providers/CoreML-ExecutionProvider.html). 
> - [x] Allows users to append execution providers like `WebGpuExecutionProvider` and `CpuExecutionProvider` through `OnnxRuntime.appendExecutionProvider_V2` method.
> 
> CC @LizeRaes

Ana Maria Mihalceanu has updated the pull request incrementally with one additional commit since the last revision:

  Remove unused imports.
  
  Signed-off-by: ammbra <mihalceanuana at gmail.com>

-------------

Changes:
  - all: https://git.openjdk.org/babylon/pull/645/files
  - new: https://git.openjdk.org/babylon/pull/645/files/1d67ef6a..354c1149

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=babylon&pr=645&range=03
 - incr: https://webrevs.openjdk.org/?repo=babylon&pr=645&range=02-03

  Stats: 2 lines in 1 file changed: 0 ins; 2 del; 0 mod
  Patch: https://git.openjdk.org/babylon/pull/645.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/645/head:pull/645

PR: https://git.openjdk.org/babylon/pull/645

From duke at openjdk.org  Wed Oct 29 19:24:37 2025
From: duke at openjdk.org (Ana Maria Mihalceanu)
Date: Wed, 29 Oct 2025 19:24:37 GMT
Subject: [code-reflection] RFR: Upgrade ONNX runtime to use API version
 1.23.1 [v3]
In-Reply-To: <MYcPCFflprY6f9F4OBDLj12qiiSZI7T5Y2xgm3-f3No=.3dfca7ce-f14b-497f-a115-796ff36167a4@github.com>
References: <ogjLjVBV2D868E0DIzs1f9fP4806ecVBz4nhwtDN5Jg=.86eabea2-9283-453f-956a-e01a3014df01@github.com>
 <zE7HVzNTY6U9gDdl4NmvUZWBks7W4VJMBcCqLYj4XfY=.ab2c502c-a8ad-4b92-a558-19d46eef62bb@github.com>
 <MYcPCFflprY6f9F4OBDLj12qiiSZI7T5Y2xgm3-f3No=.3dfca7ce-f14b-497f-a115-796ff36167a4@github.com>
Message-ID: <RQ9Ioo1F4hSU0KBjSyvXTXNsK7SqYC6zOtK3OO824CU=.c5ac6416-6a98-4aaa-b027-e3fcfe6dfcd7@github.com>

On Wed, 29 Oct 2025 18:37:50 GMT, Adam Sotona <asotona at openjdk.org> wrote:

>> Ana Maria Mihalceanu has updated the pull request incrementally with one additional commit since the last revision:
>> 
>>   Allow low-precision accumulation on a GPU.
>>   
>>   Signed-off-by: ammbra <mihalceanuana at gmail.com>
>
> cr-examples/onnx/src/test/java/oracle/code/onnx/fer/FERInference.java line 35:
> 
>> 33: import java.lang.foreign.Arena;
>> 34: import java.net.URL;
>> 35: import java.util.Map;
> 
> unused imports?

Yes, I removed those in my recent commit. Thank you for catching them.

> cr-examples/onnx/src/test/java/oracle/code/onnx/fer/FERInference.java line 39:
> 
>> 37: 
>> 38: import static oracle.code.onnx.fer.FERCoreMLDemo.IMAGE_SIZE;
>> 39: import static oracle.code.onnx.foreign.onnxruntime_c_api_h.*;
> 
> unused imports?

Yes, I forgot about these ? . Thank you for catching them.

-------------

PR Review Comment: https://git.openjdk.org/babylon/pull/645#discussion_r2475030349
PR Review Comment: https://git.openjdk.org/babylon/pull/645#discussion_r2475028608

From duke at openjdk.org  Wed Oct 29 19:34:31 2025
From: duke at openjdk.org (Ana Maria Mihalceanu)
Date: Wed, 29 Oct 2025 19:34:31 GMT
Subject: [code-reflection] RFR: Upgrade ONNX runtime to use API version
 1.23.1 [v5]
In-Reply-To: <ogjLjVBV2D868E0DIzs1f9fP4806ecVBz4nhwtDN5Jg=.86eabea2-9283-453f-956a-e01a3014df01@github.com>
References: <ogjLjVBV2D868E0DIzs1f9fP4806ecVBz4nhwtDN5Jg=.86eabea2-9283-453f-956a-e01a3014df01@github.com>
Message-ID: <u8q7-PLFM6ma12fp3brpv7Z540oVE4RcY9QtUh9lcO4=.3cd5673b-18c1-4278-a37d-0aafec34073d@github.com>

> This PR contains the following:
> 
> - [x] Upgrade to version 1.23.1 ONNX runtime API.
> - [x] Refactoring of the python script as `AttributeProto` is now in `onnx` (I used version `1.19.1` which is the latest)
> - [x] Regeneration of tensors and tensor operations.
> - [x] Regeneration of the runtime API based on `onnxruntime_c_api.h`.
> - [x] Appends execution provider options to runtime, as shown in the [new API](https://onnxruntime.ai/docs/execution-providers/CoreML-ExecutionProvider.html). 
> - [x] Allows users to append execution providers like `WebGpuExecutionProvider` and `CpuExecutionProvider` through `OnnxRuntime.appendExecutionProvider_V2` method.
> 
> CC @LizeRaes

Ana Maria Mihalceanu has updated the pull request incrementally with one additional commit since the last revision:

  Add copyright header.
  
  Signed-off-by: ammbra <mihalceanuana at gmail.com>

-------------

Changes:
  - all: https://git.openjdk.org/babylon/pull/645/files
  - new: https://git.openjdk.org/babylon/pull/645/files/354c1149..5f37308b

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=babylon&pr=645&range=04
 - incr: https://webrevs.openjdk.org/?repo=babylon&pr=645&range=03-04

  Stats: 25 lines in 1 file changed: 25 ins; 0 del; 0 mod
  Patch: https://git.openjdk.org/babylon/pull/645.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/645/head:pull/645

PR: https://git.openjdk.org/babylon/pull/645

From duke at openjdk.org  Wed Oct 29 19:34:34 2025
From: duke at openjdk.org (Ana Maria Mihalceanu)
Date: Wed, 29 Oct 2025 19:34:34 GMT
Subject: [code-reflection] RFR: Upgrade ONNX runtime to use API version
 1.23.1 [v3]
In-Reply-To: <MYcPCFflprY6f9F4OBDLj12qiiSZI7T5Y2xgm3-f3No=.3dfca7ce-f14b-497f-a115-796ff36167a4@github.com>
References: <ogjLjVBV2D868E0DIzs1f9fP4806ecVBz4nhwtDN5Jg=.86eabea2-9283-453f-956a-e01a3014df01@github.com>
 <zE7HVzNTY6U9gDdl4NmvUZWBks7W4VJMBcCqLYj4XfY=.ab2c502c-a8ad-4b92-a558-19d46eef62bb@github.com>
 <MYcPCFflprY6f9F4OBDLj12qiiSZI7T5Y2xgm3-f3No=.3dfca7ce-f14b-497f-a115-796ff36167a4@github.com>
Message-ID: <bn9nFKnJ3P8UZ-yZ-oWqII_KN56aGvGYoQLqm4TZRT4=.be2e99ca-ed17-4e52-a008-7141c2806f4c@github.com>

On Wed, 29 Oct 2025 18:27:16 GMT, Adam Sotona <asotona at openjdk.org> wrote:

>> Ana Maria Mihalceanu has updated the pull request incrementally with one additional commit since the last revision:
>> 
>>   Allow low-precision accumulation on a GPU.
>>   
>>   Signed-off-by: ammbra <mihalceanuana at gmail.com>
>
> cr-examples/onnx/src/main/java/oracle/code/onnx/OnnxProvider.java line 1:
> 
>> 1: package oracle.code.onnx;
> 
> Missing copyright header

Added in the recent commit ?

-------------

PR Review Comment: https://git.openjdk.org/babylon/pull/645#discussion_r2475055631

From duke at openjdk.org  Wed Oct 29 19:34:35 2025
From: duke at openjdk.org (Ana Maria Mihalceanu)
Date: Wed, 29 Oct 2025 19:34:35 GMT
Subject: [code-reflection] Integrated: Upgrade ONNX runtime to use API version
 1.23.1
In-Reply-To: <ogjLjVBV2D868E0DIzs1f9fP4806ecVBz4nhwtDN5Jg=.86eabea2-9283-453f-956a-e01a3014df01@github.com>
References: <ogjLjVBV2D868E0DIzs1f9fP4806ecVBz4nhwtDN5Jg=.86eabea2-9283-453f-956a-e01a3014df01@github.com>
Message-ID: <xZc2PBHwDIcyB98XvZSFysh4WLKm4JrIeBAneJgWqqk=.4b995576-62bd-4324-ae2c-6366b4be4032@github.com>

On Mon, 27 Oct 2025 20:37:13 GMT, Ana Maria Mihalceanu <duke at openjdk.org> wrote:

> This PR contains the following:
> 
> - [x] Upgrade to version 1.23.1 ONNX runtime API.
> - [x] Refactoring of the python script as `AttributeProto` is now in `onnx` (I used version `1.19.1` which is the latest)
> - [x] Regeneration of tensors and tensor operations.
> - [x] Regeneration of the runtime API based on `onnxruntime_c_api.h`.
> - [x] Appends execution provider options to runtime, as shown in the [new API](https://onnxruntime.ai/docs/execution-providers/CoreML-ExecutionProvider.html). 
> - [x] Allows users to append execution providers like `WebGpuExecutionProvider` and `CpuExecutionProvider` through `OnnxRuntime.appendExecutionProvider_V2` method.
> 
> CC @LizeRaes

This pull request has now been integrated.

Changeset: 59a20e81
Author:    ammbra <mihalceanuana at gmail.com>
Committer: Adam Sotona <asotona at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/59a20e814588b6efe49cb67284814c20b23f6f7c
Stats:     49874 lines in 41 files changed: 42971 ins; 3317 del; 3586 mod

Upgrade ONNX runtime to use API version 1.23.1

Reviewed-by: asotona

-------------

PR: https://git.openjdk.org/babylon/pull/645

From asotona at openjdk.org  Wed Oct 29 19:36:29 2025
From: asotona at openjdk.org (Adam Sotona)
Date: Wed, 29 Oct 2025 19:36:29 GMT
Subject: git: openjdk/babylon: code-reflection: Upgrade ONNX runtime to use API
 version 1.23.1
Message-ID: <f2f33b17-62c3-44a7-bc75-74c1e89d5823@openjdk.org>

Changeset: 59a20e81
Branch: code-reflection
Author:    ammbra <mihalceanuana at gmail.com>
Committer: Adam Sotona <asotona at openjdk.org>
Date:      2025-10-29 19:30:23 +0000
URL:       https://git.openjdk.org/babylon/commit/59a20e814588b6efe49cb67284814c20b23f6f7c

Upgrade ONNX runtime to use API version 1.23.1

Reviewed-by: asotona

! cr-examples/onnx/opgen/onnxOpSchemaToJSON.py
! cr-examples/onnx/opgen/setup.sh
! cr-examples/onnx/pom.xml
! cr-examples/onnx/src/main/java/oracle/code/onnx/OnnxOperators.java
+ cr-examples/onnx/src/main/java/oracle/code/onnx/OnnxProvider.java
! cr-examples/onnx/src/main/java/oracle/code/onnx/OnnxRuntime.java
+ cr-examples/onnx/src/main/java/oracle/code/onnx/foreign/EpSelectionDelegate.java
+ cr-examples/onnx/src/main/java/oracle/code/onnx/foreign/OrtAllocator.java
! cr-examples/onnx/src/main/java/oracle/code/onnx/foreign/OrtApi.java
! cr-examples/onnx/src/main/java/oracle/code/onnx/foreign/OrtApiBase.java
+ cr-examples/onnx/src/main/java/oracle/code/onnx/foreign/OrtCUDAProviderOptions.java
+ cr-examples/onnx/src/main/java/oracle/code/onnx/foreign/OrtCompileApi.java
+ cr-examples/onnx/src/main/java/oracle/code/onnx/foreign/OrtCustomCreateThreadFn.java
+ cr-examples/onnx/src/main/java/oracle/code/onnx/foreign/OrtCustomHandleType.java
+ cr-examples/onnx/src/main/java/oracle/code/onnx/foreign/OrtCustomJoinThreadFn.java
+ cr-examples/onnx/src/main/java/oracle/code/onnx/foreign/OrtCustomOp.java
+ cr-examples/onnx/src/main/java/oracle/code/onnx/foreign/OrtGetInitializerLocationFunc.java
+ cr-examples/onnx/src/main/java/oracle/code/onnx/foreign/OrtLoggingFunction.java
+ cr-examples/onnx/src/main/java/oracle/code/onnx/foreign/OrtMIGraphXProviderOptions.java
+ cr-examples/onnx/src/main/java/oracle/code/onnx/foreign/OrtModelEditorApi.java
+ cr-examples/onnx/src/main/java/oracle/code/onnx/foreign/OrtOpenVINOProviderOptions.java
+ cr-examples/onnx/src/main/java/oracle/code/onnx/foreign/OrtROCMProviderOptions.java
+ cr-examples/onnx/src/main/java/oracle/code/onnx/foreign/OrtTensorRTProviderOptions.java
+ cr-examples/onnx/src/main/java/oracle/code/onnx/foreign/OrtThreadWorkerFn.java
+ cr-examples/onnx/src/main/java/oracle/code/onnx/foreign/OrtWriteBufferFunc.java
+ cr-examples/onnx/src/main/java/oracle/code/onnx/foreign/RegisterCustomOpsFn.java
+ cr-examples/onnx/src/main/java/oracle/code/onnx/foreign/RunAsyncCallbackFn.java
! cr-examples/onnx/src/main/java/oracle/code/onnx/foreign/onnxruntime_c_api_h.java
! cr-examples/onnx/src/main/java/oracle/code/onnx/ir/OnnxOps.java
! cr-examples/onnx/src/main/java/oracle/code/onnx/ir/OnnxType.java
! cr-examples/onnx/src/main/java/oracle/code/onnx/proto/OnnxBuilder.java
! cr-examples/onnx/src/main/java/oracle/code/onnx/proto/OnnxConstants.java
! cr-examples/onnx/src/main/java/oracle/code/onnx/proto/OnnxModel.java
- cr-examples/onnx/src/main/java/oracle/code/onnx/provider/CoreMLProvider.java
- cr-examples/onnx/src/main/java/oracle/code/onnx/provider/OnnxProvider.java
! cr-examples/onnx/src/test/java/oracle/code/onnx/CNNTest.java
! cr-examples/onnx/src/test/java/oracle/code/onnx/RuntimeTest.java
! cr-examples/onnx/src/test/java/oracle/code/onnx/WalkTheMazeTest.java
! cr-examples/onnx/src/test/java/oracle/code/onnx/fer/FERCoreMLDemo.java
! cr-examples/onnx/src/test/java/oracle/code/onnx/fer/FERInference.java
! cr-examples/onnx/src/test/java/oracle/code/onnx/llm/LlamaModel.java


From gfrost at openjdk.org  Wed Oct 29 19:42:50 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Wed, 29 Oct 2025 19:42:50 GMT
Subject: [code-reflection] Integrated: Hat refactor view
Message-ID: <gHZqZRxQAQCPmlL9ziVhqWuu0B6qVdn1WokM6tmJWh0=.1b75d793-0125-4011-afc2-a13243ffa4e8@github.com>

Small changes to view example. 

I am playing with iface/record vector implementations 

A small regex change also in here

-------------

Commit messages:
 - regex changes
 - replace pools
 - merging types
 - merging types
 - merging types
 - merging types
 - merging types
 - Refactor index pools and delete dead code

Changes: https://git.openjdk.org/babylon/pull/649/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=649&range=00
  Stats: 1979 lines in 30 files changed: 852 ins; 918 del; 209 mod
  Patch: https://git.openjdk.org/babylon/pull/649.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/649/head:pull/649

PR: https://git.openjdk.org/babylon/pull/649

From gfrost at openjdk.org  Wed Oct 29 19:42:53 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Wed, 29 Oct 2025 19:42:53 GMT
Subject: [code-reflection] Integrated: Hat refactor view
In-Reply-To: <gHZqZRxQAQCPmlL9ziVhqWuu0B6qVdn1WokM6tmJWh0=.1b75d793-0125-4011-afc2-a13243ffa4e8@github.com>
References: <gHZqZRxQAQCPmlL9ziVhqWuu0B6qVdn1WokM6tmJWh0=.1b75d793-0125-4011-afc2-a13243ffa4e8@github.com>
Message-ID: <ng85z2VvdC325lmV7mXks61x6hcC8mTJqwuFyy4zzo0=.b58dc227-cc53-4e0a-bc2b-407676b6ad65@github.com>

On Wed, 29 Oct 2025 19:35:31 GMT, Gary Frost <gfrost at openjdk.org> wrote:

> Small changes to view example. 
> 
> I am playing with iface/record vector implementations 
> 
> A small regex change also in here

This pull request has now been integrated.

Changeset: d175d75a
Author:    Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/d175d75ae9287e2d858fc73a9dd352514ea37a3c
Stats:     1979 lines in 30 files changed: 852 ins; 918 del; 209 mod

Hat refactor view

-------------

PR: https://git.openjdk.org/babylon/pull/649

From gfrost at openjdk.org  Wed Oct 29 19:44:52 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Wed, 29 Oct 2025 19:44:52 GMT
Subject: git: openjdk/babylon: code-reflection: Hat refactor view
Message-ID: <16eecc6a-f214-4933-b827-ff4ba4f3c07b@openjdk.org>

Changeset: d175d75a
Branch: code-reflection
Author:    Gary Frost <gfrost at openjdk.org>
Date:      2025-10-29 19:39:43 +0000
URL:       https://git.openjdk.org/babylon/commit/d175d75ae9287e2d858fc73a9dd352514ea37a3c

Hat refactor view

! hat/core/src/main/java/hat/util/Regex.java
! hat/examples/view/src/main/java/view/EliteMeshReader.java
! hat/examples/view/src/main/java/view/Main.java
+ hat/examples/view/src/main/java/view/ModelHighWaterMark.java
= hat/examples/view/src/main/java/view/Physics.java
! hat/examples/view/src/main/java/view/Rasterizer.java
+ hat/examples/view/src/main/java/view/Renderer.java
! hat/examples/view/src/main/java/view/ViewFrame.java
! hat/examples/view/src/main/java/view/ZPos.java
! hat/examples/view/src/main/java/view/f32/F32Line3D.java
- hat/examples/view/src/main/java/view/f32/F32Mat4.java
+ hat/examples/view/src/main/java/view/f32/F32Matrix4x4.java
! hat/examples/view/src/main/java/view/f32/F32Mesh3D.java
+ hat/examples/view/src/main/java/view/f32/F32Triangle2D.java
! hat/examples/view/src/main/java/view/f32/F32Triangle3D.java
+ hat/examples/view/src/main/java/view/f32/F32Vec2.java
! hat/examples/view/src/main/java/view/f32/F32Vec3.java
+ hat/examples/view/src/main/java/view/f32/FloatPool.java
+ hat/examples/view/src/main/java/view/f32/IndexPool.java
+ hat/examples/view/src/main/java/view/f32/Pool.java
- hat/examples/view/src/main/java/view/f32/mat4.java
- hat/examples/view/src/main/java/view/f32/projectionMat4.java
- hat/examples/view/src/main/java/view/f32/rotationMat4.java
- hat/examples/view/src/main/java/view/f32/scaleMat4.java
- hat/examples/view/src/main/java/view/f32/translateMat4.java
- hat/examples/view/src/main/java/view/f32/tri.java
- hat/examples/view/src/main/java/view/f32/vec3.java
- hat/examples/view/src/main/java/view/i32/I32Line2D.java
- hat/examples/view/src/main/java/view/i32/I32Triangle2D.java
- hat/examples/view/src/main/java/view/i32/I32Vec2.java


From asotona at openjdk.org  Wed Oct 29 21:33:25 2025
From: asotona at openjdk.org (Adam Sotona)
Date: Wed, 29 Oct 2025 21:33:25 GMT
Subject: [code-reflection] RFR: OnnxLift prototype [v5]
In-Reply-To: <dFnT1a6oJdYYzER2-cOVWCIgNce9uh0mBloQ3T52EDc=.612309ef-e96a-4a48-9223-3cafbda9663d@github.com>
References: <dFnT1a6oJdYYzER2-cOVWCIgNce9uh0mBloQ3T52EDc=.612309ef-e96a-4a48-9223-3cafbda9663d@github.com>
Message-ID: <VSmKpPHa6iMbVRla-P4bqKbTtNss9lEPsHIFM5orQT8=.69acec2d-4680-4a28-9970-ce53d7c4806c@github.com>

> This is a prototype of OnnxLift tool.
> It lifts ONNX model binary to ONNX code reflection model, extracts weights, and generates Java model source.

Adam Sotona has updated the pull request with a new target base due to a merge or a rebase. The incremental webrev excludes the unrelated changes brought in by the merge/rebase. The pull request contains eight additional commits since the last revision:

 - fixed imports
   
 - rollback of Tensor factory as Constant attribute
   
 - fixed imports
   
 - Merge branch 'code-reflection' into onnx-lift-prototype
 - added class name cmd option
   
 - refactoring split to JavaTemplate
   
 - missing annotation
   
 - OnnxLift prototype

-------------

Changes:
  - all: https://git.openjdk.org/babylon/pull/638/files
  - new: https://git.openjdk.org/babylon/pull/638/files/bf54e820..943b85af

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=babylon&pr=638&range=04
 - incr: https://webrevs.openjdk.org/?repo=babylon&pr=638&range=03-04

  Stats: 52309 lines in 88 files changed: 44061 ins; 4353 del; 3895 mod
  Patch: https://git.openjdk.org/babylon/pull/638.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/638/head:pull/638

PR: https://git.openjdk.org/babylon/pull/638

From asotona at openjdk.org  Wed Oct 29 21:36:17 2025
From: asotona at openjdk.org (Adam Sotona)
Date: Wed, 29 Oct 2025 21:36:17 GMT
Subject: git: openjdk/babylon: code-reflection: OnnxLift prototype
Message-ID: <652ca2d4-dde6-44b5-a38e-cbeda103e78b@openjdk.org>

Changeset: df377fb2
Branch: code-reflection
Author:    Adam Sotona <asotona at openjdk.org>
Date:      2025-10-29 21:35:10 +0000
URL:       https://git.openjdk.org/babylon/commit/df377fb293c362210b1e7a55f02b064992e0c727

OnnxLift prototype

! cr-examples/onnx/README.md
! cr-examples/onnx/pom.xml
+ cr-examples/onnx/src/main/java/oracle/code/onnx/lift/JavaTemplate.java
+ cr-examples/onnx/src/main/java/oracle/code/onnx/lift/OnnxLift.java


From asotona at openjdk.org  Wed Oct 29 21:38:35 2025
From: asotona at openjdk.org (Adam Sotona)
Date: Wed, 29 Oct 2025 21:38:35 GMT
Subject: [code-reflection] Integrated: OnnxLift prototype
In-Reply-To: <dFnT1a6oJdYYzER2-cOVWCIgNce9uh0mBloQ3T52EDc=.612309ef-e96a-4a48-9223-3cafbda9663d@github.com>
References: <dFnT1a6oJdYYzER2-cOVWCIgNce9uh0mBloQ3T52EDc=.612309ef-e96a-4a48-9223-3cafbda9663d@github.com>
Message-ID: <KxvbJqxhVT1daJCi2I5ErzSakv-dbtNCML-YB5GOJ2g=.b5b9762c-999b-4893-937d-29e0b9af03de@github.com>

On Fri, 24 Oct 2025 09:19:15 GMT, Adam Sotona <asotona at openjdk.org> wrote:

> This is a prototype of OnnxLift tool.
> It lifts ONNX model binary to ONNX code reflection model, extracts weights, and generates Java model source.

This pull request has now been integrated.

Changeset: df377fb2
Author:    Adam Sotona <asotona at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/df377fb293c362210b1e7a55f02b064992e0c727
Stats:     784 lines in 4 files changed: 782 ins; 1 del; 1 mod

OnnxLift prototype

-------------

PR: https://git.openjdk.org/babylon/pull/638

From mabbay at openjdk.org  Thu Oct 30 10:16:57 2025
From: mabbay at openjdk.org (Mourad Abbay)
Date: Thu, 30 Oct 2025 10:16:57 GMT
Subject: [code-reflection] RFR: Replace methods "create" in Op subclasses
 by constructors [v3]
In-Reply-To: <U_X6K2k0xo7jYVxHu70A-pqiabYJUMs7LXzH-5TxDKk=.d0b686fd-d3c8-4b9b-93cf-1f5938b3bada@github.com>
References: <U_X6K2k0xo7jYVxHu70A-pqiabYJUMs7LXzH-5TxDKk=.d0b686fd-d3c8-4b9b-93cf-1f5938b3bada@github.com>
Message-ID: <P-nXRS7-6k5chFp0yqtVWNf80VFgyqXaZsxQQsuzirI=.e41adee4-3d3b-4c6e-b843-d233cd5005d3@github.com>

> By using flexible constructor bodies feature, we get rid of methods "create" in Op subclasses and we replace it with constructors. We also remove unused constructors and methods in Op subclasses.

Mourad Abbay has updated the pull request incrementally with one additional commit since the last revision:

  Apply suggestion

-------------

Changes:
  - all: https://git.openjdk.org/babylon/pull/584/files
  - new: https://git.openjdk.org/babylon/pull/584/files/372cc17b..49d4367b

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=babylon&pr=584&range=02
 - incr: https://webrevs.openjdk.org/?repo=babylon&pr=584&range=01-02

  Stats: 14 lines in 1 file changed: 7 ins; 7 del; 0 mod
  Patch: https://git.openjdk.org/babylon/pull/584.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/584/head:pull/584

PR: https://git.openjdk.org/babylon/pull/584

From mabbay at openjdk.org  Thu Oct 30 10:16:58 2025
From: mabbay at openjdk.org (Mourad Abbay)
Date: Thu, 30 Oct 2025 10:16:58 GMT
Subject: [code-reflection] Integrated: Replace methods "create" in Op
 subclasses by constructors
In-Reply-To: <U_X6K2k0xo7jYVxHu70A-pqiabYJUMs7LXzH-5TxDKk=.d0b686fd-d3c8-4b9b-93cf-1f5938b3bada@github.com>
References: <U_X6K2k0xo7jYVxHu70A-pqiabYJUMs7LXzH-5TxDKk=.d0b686fd-d3c8-4b9b-93cf-1f5938b3bada@github.com>
Message-ID: <SM6qiZH-m6SzVsT3nrHLsOQPGxRZimkP2Muc4M24tXI=.b44c1c5f-0a47-46cf-b086-1828ae142530@github.com>

On Fri, 26 Sep 2025 18:46:41 GMT, Mourad Abbay <mabbay at openjdk.org> wrote:

> By using flexible constructor bodies feature, we get rid of methods "create" in Op subclasses and we replace it with constructors. We also remove unused constructors and methods in Op subclasses.

This pull request has now been integrated.

Changeset: 837bb95f
Author:    Mourad Abbay <mabbay at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/837bb95fb29152fcd827de623399136c5974d1e2
Stats:     84 lines in 2 files changed: 10 ins; 19 del; 55 mod

Replace methods "create" in Op subclasses by constructors

Reviewed-by: psandoz

-------------

PR: https://git.openjdk.org/babylon/pull/584

From mabbay at openjdk.org  Thu Oct 30 10:21:13 2025
From: mabbay at openjdk.org (Mourad Abbay)
Date: Thu, 30 Oct 2025 10:21:13 GMT
Subject: git: openjdk/babylon: code-reflection: Replace methods "create" in Op
 subclasses by constructors
Message-ID: <2041c095-a524-443f-8ee0-2b2d0be6cb0f@openjdk.org>

Changeset: 837bb95f
Branch: code-reflection
Author:    Mourad Abbay <mabbay at openjdk.org>
Date:      2025-10-30 10:13:52 +0000
URL:       https://git.openjdk.org/babylon/commit/837bb95fb29152fcd827de623399136c5974d1e2

Replace methods "create" in Op subclasses by constructors

Reviewed-by: psandoz

! src/jdk.incubator.code/share/classes/jdk/incubator/code/dialect/core/CoreOp.java
! src/jdk.incubator.code/share/classes/jdk/incubator/code/dialect/java/JavaOp.java


From gfrost at openjdk.org  Thu Oct 30 13:48:10 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Thu, 30 Oct 2025 13:48:10 GMT
Subject: [code-reflection] Integrated: Hat view more tinkering
Message-ID: <4AWK_ypeQnQ2ktpxn7pdWZMGWJni3IE2XuywmIZSHWQ=.bf47230b-4c57-4240-9331-f9058cb0632e@github.com>

Ignore this ;) I am migrating view example across to use vector types compatible with HAT.  

Work in progress.. more to come,

-------------

Commit messages:
 - slowly migrating to removing pools
 - safety checkin #4 simplify
 - safety checkin #3 - use old for now
 - safety checkin #2
 - safety checkin

Changes: https://git.openjdk.org/babylon/pull/650/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=650&range=00
  Stats: 1913 lines in 10 files changed: 1673 ins; 165 del; 75 mod
  Patch: https://git.openjdk.org/babylon/pull/650.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/650/head:pull/650

PR: https://git.openjdk.org/babylon/pull/650

From gfrost at openjdk.org  Thu Oct 30 13:48:11 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Thu, 30 Oct 2025 13:48:11 GMT
Subject: [code-reflection] Integrated: Hat view more tinkering
In-Reply-To: <4AWK_ypeQnQ2ktpxn7pdWZMGWJni3IE2XuywmIZSHWQ=.bf47230b-4c57-4240-9331-f9058cb0632e@github.com>
References: <4AWK_ypeQnQ2ktpxn7pdWZMGWJni3IE2XuywmIZSHWQ=.bf47230b-4c57-4240-9331-f9058cb0632e@github.com>
Message-ID: <S3fg_DWkH_0hEi6gkJEo3tjZASSmClGrBXPRfJSHxvI=.cbbffc59-991e-4b95-8d69-c7a665612f49@github.com>

On Thu, 30 Oct 2025 13:39:02 GMT, Gary Frost <gfrost at openjdk.org> wrote:

> Ignore this ;) I am migrating view example across to use vector types compatible with HAT.  
> 
> Work in progress.. more to come,

This pull request has now been integrated.

Changeset: ceb9524c
Author:    Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/ceb9524c6bcce6166b736043397599ebd3b99024
Stats:     1913 lines in 10 files changed: 1673 ins; 165 del; 75 mod

Hat view more tinkering

-------------

PR: https://git.openjdk.org/babylon/pull/650

From gfrost at openjdk.org  Thu Oct 30 13:50:53 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Thu, 30 Oct 2025 13:50:53 GMT
Subject: git: openjdk/babylon: code-reflection: Hat view more tinkering
Message-ID: <6703de2a-2ab4-434d-a35f-ca7f283d3262@openjdk.org>

Changeset: ceb9524c
Branch: code-reflection
Author:    Gary Frost <gfrost at openjdk.org>
Date:      2025-10-30 13:43:39 +0000
URL:       https://git.openjdk.org/babylon/commit/ceb9524c6bcce6166b736043397599ebd3b99024

Hat view more tinkering

! hat/examples/experiments/src/main/java/experiments/PrefixSum.java
! hat/examples/view/src/main/java/view/EliteMeshReader.java
+ hat/examples/view/src/main/java/view/F32.java
! hat/examples/view/src/main/java/view/Main.java
- hat/examples/view/src/main/java/view/Physics.java
! hat/examples/view/src/main/java/view/Rasterizer.java
! hat/examples/view/src/main/java/view/Renderer.java
! hat/examples/view/src/main/java/view/ViewFrame.java
+ hat/examples/view/src/main/java/view/ViewFrameNew.java
+ hat/examples/view/src/main/java/view/ViewFrameOld.java


From gfrost at openjdk.org  Thu Oct 30 15:06:13 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Thu, 30 Oct 2025 15:06:13 GMT
Subject: [code-reflection] Integrated: Moved job src into hat project
Message-ID: <m8P8fj_scnBAWstvsTCbCAGCX6qNLq9C62d0agAbK_k=.a2032668-5e1f-45f9-9fd7-0422c3fda6e4@github.com>

We have been using job (java opinionated builder) for some time.  The code has been sitting on my local disk. 

To create job.jar (should only be needed if we change the builder source )


 bash hat/bootstrap.bash 


After that `java @hat/xxxx` will use the modified job.jar.  

Please check job.jar back in if you make changes that we need.

-------------

Commit messages:
 - job src into hat project

Changes: https://git.openjdk.org/babylon/pull/651/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=651&range=00
  Stats: 3740 lines in 31 files changed: 3723 ins; 16 del; 1 mod
  Patch: https://git.openjdk.org/babylon/pull/651.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/651/head:pull/651

PR: https://git.openjdk.org/babylon/pull/651

From gfrost at openjdk.org  Thu Oct 30 15:06:13 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Thu, 30 Oct 2025 15:06:13 GMT
Subject: [code-reflection] Integrated: Moved job src into hat project
In-Reply-To: <m8P8fj_scnBAWstvsTCbCAGCX6qNLq9C62d0agAbK_k=.a2032668-5e1f-45f9-9fd7-0422c3fda6e4@github.com>
References: <m8P8fj_scnBAWstvsTCbCAGCX6qNLq9C62d0agAbK_k=.a2032668-5e1f-45f9-9fd7-0422c3fda6e4@github.com>
Message-ID: <PzeXFiK5UY6N6zNvsl7tZY7qyD_u6yzEAqP-XHLEkxE=.1417fa40-e1a7-4852-aa6e-21be339bb836@github.com>

On Thu, 30 Oct 2025 14:56:55 GMT, Gary Frost <gfrost at openjdk.org> wrote:

> We have been using job (java opinionated builder) for some time.  The code has been sitting on my local disk. 
> 
> To create job.jar (should only be needed if we change the builder source )
> 
> 
>  bash hat/bootstrap.bash 
> 
> 
> After that `java @hat/xxxx` will use the modified job.jar.  
> 
> Please check job.jar back in if you make changes that we need.

This pull request has now been integrated.

Changeset: b6b09e85
Author:    Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/b6b09e85cdbe57d2817f6f7e5f4b2ecab2850a5a
Stats:     3740 lines in 31 files changed: 3723 ins; 16 del; 1 mod

Moved job src into hat project

-------------

PR: https://git.openjdk.org/babylon/pull/651

From gfrost at openjdk.org  Thu Oct 30 15:09:25 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Thu, 30 Oct 2025 15:09:25 GMT
Subject: git: openjdk/babylon: code-reflection: Moved job src into hat project
Message-ID: <b277100b-53d4-4f2d-9f74-16b523de91d6@openjdk.org>

Changeset: b6b09e85
Branch: code-reflection
Author:    Gary Frost <gfrost at openjdk.org>
Date:      2025-10-30 15:02:38 +0000
URL:       https://git.openjdk.org/babylon/commit/b6b09e85cdbe57d2817f6f7e5f4b2ecab2850a5a

Moved job src into hat project

! hat/examples/pom.xml
! hat/examples/view/src/main/java/view/F32.java
! hat/hat.java
+ hat/hat/bootstrap.bash
! hat/hat/job.jar
+ hat/hat/job/.gitignore
+ hat/hat/job/src/main/java/job/CMake.java
+ hat/hat/job/src/main/java/job/CMakeInfo.java
+ hat/hat/job/src/main/java/job/Cuda.java
+ hat/hat/job/src/main/java/job/Dag.java
+ hat/hat/job/src/main/java/job/Dependency.java
+ hat/hat/job/src/main/java/job/DependencyImpl.java
+ hat/hat/job/src/main/java/job/ForkExec.java
+ hat/hat/job/src/main/java/job/Intellij.java
+ hat/hat/job/src/main/java/job/JExtract.java
+ hat/hat/job/src/main/java/job/JExtractOptProvider.java
+ hat/hat/job/src/main/java/job/Jar.java
+ hat/hat/job/src/main/java/job/Linux.java
+ hat/hat/job/src/main/java/job/Mac.java
+ hat/hat/job/src/main/java/job/Maven.java
+ hat/hat/job/src/main/java/job/OpenCL.java
+ hat/hat/job/src/main/java/job/OpenGL.java
+ hat/hat/job/src/main/java/job/Opt.java
+ hat/hat/job/src/main/java/job/Project.java
+ hat/hat/job/src/main/java/job/Reporter.java
+ hat/hat/job/src/main/java/job/Util.java
+ hat/hat/job/src/main/java/job/Windows.java
+ hat/hat/job/src/main/java/job/XMLNode.java
! hat/intellij/.idea/modules.xml
- hat/intellij/example_arrayview.iml
+ hat/intellij/job.iml


From psandoz at openjdk.org  Thu Oct 30 17:39:04 2025
From: psandoz at openjdk.org (Paul Sandoz)
Date: Thu, 30 Oct 2025 17:39:04 GMT
Subject: [code-reflection] RFR: Merge master
Message-ID: <TBT1Y6bk-ILnapRnSxO-c1uyMpnd2QudQTBvcDMsSDg=.86182088-a811-4822-a931-0dbb8c2e6b94@github.com>

Merge master into code-reflection.

Building Babylon depends on JDK 25 and jtreg >= 8.1

-------------

Commit messages:
 - Merge master.
 - 8370851: Mark hotspot and jdk tests incompatible with test thread factory
 - 8370908: Remove test javax/management/remote/mandatory/connection/DeadLockTest.java from ProblemList-Virtual
 - 8369609: calls from Continuations to invalidate_jvmti_stack must be more accurate
 - 8370730: Test serviceability/attach/EarlyDynamicLoad/EarlyDynamicLoad.java needs to be resilient about warnings
 - 8351159: Remaining cleanups in cpu/x86 after 32-bit x86 removal
 - 8369435: C2: transform (LShiftX (SubX con0 a), con1) into (SubX con0<<con1 (LShiftX a con1))
 - 8370804: G1: Make G1HeapRegionAttr::remset_is_tracked() conform to coding style
 - 8370261: Test runtime/NMT/NMTPrintMallocSiteOfCorruptedMemory.java timed out
 - 8368739: [AIX] java/net/httpclient/http3/H3SimpleGet.java#useNioSelector and #with-continuations fail
 - ... and 1884 more: https://git.openjdk.org/babylon/compare/b6b09e85...d2fb4e29

The webrevs contain the adjustments done while merging with regards to each parent branch:
 - code-reflection: https://webrevs.openjdk.org/?repo=babylon&pr=652&range=00.0
 - master: https://webrevs.openjdk.org/?repo=babylon&pr=652&range=00.1

Changes: https://git.openjdk.org/babylon/pull/652/files
  Stats: 436122 lines in 6606 files changed: 291354 ins; 100262 del; 44506 mod
  Patch: https://git.openjdk.org/babylon/pull/652.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/652/head:pull/652

PR: https://git.openjdk.org/babylon/pull/652

From psandoz at openjdk.org  Thu Oct 30 18:33:12 2025
From: psandoz at openjdk.org (Paul Sandoz)
Date: Thu, 30 Oct 2025 18:33:12 GMT
Subject: [code-reflection] RFR: Merge master [v2]
In-Reply-To: <TBT1Y6bk-ILnapRnSxO-c1uyMpnd2QudQTBvcDMsSDg=.86182088-a811-4822-a931-0dbb8c2e6b94@github.com>
References: <TBT1Y6bk-ILnapRnSxO-c1uyMpnd2QudQTBvcDMsSDg=.86182088-a811-4822-a931-0dbb8c2e6b94@github.com>
Message-ID: <oU0YIy2k0mvlHIfeRlYHvqYPi7YQfQ2ewN3yUBOmcMg=.efa5ce39-6f4f-4c7a-8a01-a1ddee31cddd@github.com>

> Merge master into code-reflection.
> 
> Building Babylon depends on JDK 25 and jtreg >= 8.1

Paul Sandoz has updated the pull request with a new target base due to a merge or a rebase. The pull request now contains 614 commits:

 - Merge master.
 - Moved job src into hat project
 - Hat view more tinkering
 - Replace methods "create" in Op subclasses by constructors
   
   Reviewed-by: psandoz
 - OnnxLift prototype
 - Hat refactor view
 - Upgrade ONNX runtime to use API version 1.23.1
   
   Reviewed-by: asotona
 - Fixed BytecodeGenerator INDY calls with invalid interface method type
   
   Reviewed-by: psandoz
 - Coverted enum base view parser to iface based
 - Hat util regex
 - ... and 604 more: https://git.openjdk.org/babylon/compare/ed36b9bb...d2fb4e29

-------------

Changes: https://git.openjdk.org/babylon/pull/652/files
  Webrev: Webrev is not available because diff is too large
  Stats: 419432 lines in 1106 files changed: 419251 ins; 73 del; 108 mod
  Patch: https://git.openjdk.org/babylon/pull/652.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/652/head:pull/652

PR: https://git.openjdk.org/babylon/pull/652

From psandoz at openjdk.org  Thu Oct 30 18:33:16 2025
From: psandoz at openjdk.org (Paul Sandoz)
Date: Thu, 30 Oct 2025 18:33:16 GMT
Subject: [code-reflection] Integrated: Merge master
In-Reply-To: <TBT1Y6bk-ILnapRnSxO-c1uyMpnd2QudQTBvcDMsSDg=.86182088-a811-4822-a931-0dbb8c2e6b94@github.com>
References: <TBT1Y6bk-ILnapRnSxO-c1uyMpnd2QudQTBvcDMsSDg=.86182088-a811-4822-a931-0dbb8c2e6b94@github.com>
Message-ID: <U6pH1JNChXWIys5328-ym6kSCs65RdSv5ZTRV9eMIUM=.e7af8e66-c331-452f-b06e-806c9a959aac@github.com>

On Thu, 30 Oct 2025 17:31:15 GMT, Paul Sandoz <psandoz at openjdk.org> wrote:

> Merge master into code-reflection.
> 
> Building Babylon depends on JDK 25 and jtreg >= 8.1

This pull request has now been integrated.

Changeset: c87e5746
Author:    Paul Sandoz <psandoz at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/c87e5746aaa35efb02ab5d2eb5026a3f20c00878
Stats:     436122 lines in 6606 files changed: 291354 ins; 100262 del; 44506 mod

Merge

-------------

PR: https://git.openjdk.org/babylon/pull/652

From jfumero at openjdk.org  Fri Oct 31 07:29:47 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Fri, 31 Oct 2025 07:29:47 GMT
Subject: git: openjdk/babylon: created branch hat/fp16/extension based on the
 branch code-reflection containing 1 unique commit
Message-ID: <f84717c8-64c5-40f2-8031-5610dbd718b1@openjdk.org>

The following commits are unique to the hat/fp16/extension branch:
========================================================
23da5523: [hat][wip] fp16 extensions

From jfumero at openjdk.org  Fri Oct 31 07:50:07 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Fri, 31 Oct 2025 07:50:07 GMT
Subject: [code-reflection] RFR: Proposal for supporting Float4.of for CPU
 host code and GPU device code [v7]
In-Reply-To: <paNHesfhOTHU4U8SWowRhNWUroeVKz9GA4SJVTaimnI=.23d8e3d6-26a5-4ff4-901e-200306e5ad7e@github.com>
References: <paNHesfhOTHU4U8SWowRhNWUroeVKz9GA4SJVTaimnI=.23d8e3d6-26a5-4ff4-901e-200306e5ad7e@github.com>
Message-ID: <Qx81EYULTfGevygdM-1ZgxCgXutz-hpOC2kqdPSyFB8=.f411086b-4ad0-46b8-8a1c-c9dc4baf1281@github.com>

> This PR adds support for creating float4 objects within the device code:
> 
> 
> Float4 f = Float4.of(x, y, z ,w)
> 
> 
> In addition, it provides a CPU implementation for this type a new method to operate in `lanewise`. This operation is CPU only at the moment:
> 
> 
> Float4 vA ... 
> Float4 vB ... 
> Float4 vC = vA.linewise(vB, Float::sum);

Juan Fumero has updated the pull request with a new target base due to a merge or a rebase. The incremental webrev excludes the unrelated changes brought in by the merge/rebase. The pull request contains 15 additional commits since the last revision:

 - Merge branch 'code-reflection' into hat/float4/of
 - [hat] config bits message fixed
 - [hat] Log Config bits for OpenCL moved to OpenCL backend instantiation
 - [hat] Dialect for handling vector types simplified
 - [hat] minor refactor ocl-codegen
 - [hat] codegen - cleanup old OpenCL typecasts
 - [hat] Move vectorOf codegen method to common C99 codegen
 - [hat] minor change
 - [hat] makeFloat4 from immutable supported on GPU
 - [hat][wip] float4 implementation with mutable/immutable variants
 - ... and 5 more: https://git.openjdk.org/babylon/compare/8d67a217...aec780a9

-------------

Changes:
  - all: https://git.openjdk.org/babylon/pull/642/files
  - new: https://git.openjdk.org/babylon/pull/642/files/251f1a52..aec780a9

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=babylon&pr=642&range=06
 - incr: https://webrevs.openjdk.org/?repo=babylon&pr=642&range=05-06

  Stats: 494433 lines in 6718 files changed: 341377 ins; 104665 del; 48391 mod
  Patch: https://git.openjdk.org/babylon/pull/642.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/642/head:pull/642

PR: https://git.openjdk.org/babylon/pull/642

From jfumero at openjdk.org  Fri Oct 31 08:27:19 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Fri, 31 Oct 2025 08:27:19 GMT
Subject: [code-reflection] RFR: Proposal for supporting Float4.of for CPU
 host code and GPU device code [v8]
In-Reply-To: <paNHesfhOTHU4U8SWowRhNWUroeVKz9GA4SJVTaimnI=.23d8e3d6-26a5-4ff4-901e-200306e5ad7e@github.com>
References: <paNHesfhOTHU4U8SWowRhNWUroeVKz9GA4SJVTaimnI=.23d8e3d6-26a5-4ff4-901e-200306e5ad7e@github.com>
Message-ID: <Px5I1GfX0rXJPLCdh8n5fRFiw2Lpl2QyB3-c2fgNu88=.482cf922-54b7-40e2-ae05-94fcd9f294c9@github.com>

> This PR adds support for creating float4 objects within the device code:
> 
> 
> Float4 f = Float4.of(x, y, z ,w)
> 
> 
> In addition, it provides a CPU implementation for this type a new method to operate in `lanewise`. This operation is CPU only at the moment:
> 
> 
> Float4 vA ... 
> Float4 vB ... 
> Float4 vC = vA.linewise(vB, Float::sum);

Juan Fumero has updated the pull request incrementally with one additional commit since the last revision:

  [hat][cuda] buildType method for vload/vstore restored

-------------

Changes:
  - all: https://git.openjdk.org/babylon/pull/642/files
  - new: https://git.openjdk.org/babylon/pull/642/files/aec780a9..dfb349ce

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=babylon&pr=642&range=07
 - incr: https://webrevs.openjdk.org/?repo=babylon&pr=642&range=06-07

  Stats: 14 lines in 3 files changed: 14 ins; 0 del; 0 mod
  Patch: https://git.openjdk.org/babylon/pull/642.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/642/head:pull/642

PR: https://git.openjdk.org/babylon/pull/642

From jfumero at openjdk.org  Fri Oct 31 08:30:39 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Fri, 31 Oct 2025 08:30:39 GMT
Subject: [code-reflection] RFR: Proposal for supporting Float4.of for CPU
 host code and GPU device code [v8]
In-Reply-To: <Px5I1GfX0rXJPLCdh8n5fRFiw2Lpl2QyB3-c2fgNu88=.482cf922-54b7-40e2-ae05-94fcd9f294c9@github.com>
References: <paNHesfhOTHU4U8SWowRhNWUroeVKz9GA4SJVTaimnI=.23d8e3d6-26a5-4ff4-901e-200306e5ad7e@github.com>
 <Px5I1GfX0rXJPLCdh8n5fRFiw2Lpl2QyB3-c2fgNu88=.482cf922-54b7-40e2-ae05-94fcd9f294c9@github.com>
Message-ID: <EQalyLlXw_2GAc5qdgKbi3m0f26Zx3SgPYNzi2rOdCk=.5f206248-8e98-4441-a03a-c14ff18ddcba@github.com>

On Fri, 31 Oct 2025 08:27:19 GMT, Juan Fumero <jfumero at openjdk.org> wrote:

>> This PR adds support for creating float4 objects within the device code:
>> 
>> 
>> Float4 f = Float4.of(x, y, z ,w)
>> 
>> 
>> In addition, it provides a CPU implementation for this type a new method to operate in `lanewise`. This operation is CPU only at the moment:
>> 
>> 
>> Float4 vA ... 
>> Float4 vB ... 
>> Float4 vC = vA.linewise(vB, Float::sum);
>
> Juan Fumero has updated the pull request incrementally with one additional commit since the last revision:
> 
>   [hat][cuda] buildType method for vload/vstore restored

PR ready and passing for both OpenCL and CUDA.

-------------

PR Comment: https://git.openjdk.org/babylon/pull/642#issuecomment-3471830825

From jfumero at openjdk.org  Fri Oct 31 08:30:41 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Fri, 31 Oct 2025 08:30:41 GMT
Subject: [code-reflection] Integrated: Proposal for supporting Float4.of for
 CPU host code and GPU device code
In-Reply-To: <paNHesfhOTHU4U8SWowRhNWUroeVKz9GA4SJVTaimnI=.23d8e3d6-26a5-4ff4-901e-200306e5ad7e@github.com>
References: <paNHesfhOTHU4U8SWowRhNWUroeVKz9GA4SJVTaimnI=.23d8e3d6-26a5-4ff4-901e-200306e5ad7e@github.com>
Message-ID: <PVnNgFre7ZkqB20UukjGsQ-v6V8IMDgSSK_r2K7o42M=.4e98ed10-f397-4ee4-899a-58db03739cc9@github.com>

On Fri, 24 Oct 2025 15:39:47 GMT, Juan Fumero <jfumero at openjdk.org> wrote:

> This PR adds support for creating float4 objects within the device code:
> 
> 
> Float4 f = Float4.of(x, y, z ,w)
> 
> 
> In addition, it provides a CPU implementation for this type a new method to operate in `lanewise`. This operation is CPU only at the moment:
> 
> 
> Float4 vA ... 
> Float4 vB ... 
> Float4 vC = vA.linewise(vB, Float::sum);

This pull request has now been integrated.

Changeset: 420c933a
Author:    Juan Fumero <jfumero at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/420c933a9e718fa193db4dc76c203fa0e4b2212e
Stats:     811 lines in 29 files changed: 593 ins; 140 del; 78 mod

Proposal for supporting Float4.of for CPU host code and GPU device code

-------------

PR: https://git.openjdk.org/babylon/pull/642

From jfumero at openjdk.org  Fri Oct 31 08:32:48 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Fri, 31 Oct 2025 08:32:48 GMT
Subject: git: openjdk/babylon: code-reflection: Proposal for supporting
 Float4.of for CPU host code and GPU device code
Message-ID: <d4d20ae5-663b-4d3e-a234-9fc08ac4258e@openjdk.org>

Changeset: 420c933a
Branch: code-reflection
Author:    Juan Fumero <jfumero at openjdk.org>
Date:      2025-10-31 08:28:19 +0000
URL:       https://git.openjdk.org/babylon/commit/420c933a9e718fa193db4dc76c203fa0e4b2212e

Proposal for supporting Float4.of for CPU host code and GPU device code

! hat/backends/ffi/cuda/src/main/java/hat/backend/ffi/CudaHATKernelBuilder.java
! hat/backends/ffi/opencl/src/main/java/hat/backend/ffi/OpenCLHATKernelBuilder.java
! hat/backends/ffi/opencl/src/main/native/cpp/opencl_backend.cpp
! hat/backends/jextracted/opencl/src/main/java/hat/backend/jextracted/OpenCLHatKernelBuilder.java
! hat/core/src/main/java/hat/buffer/Buffer.java
! hat/core/src/main/java/hat/buffer/F32ArrayPadded.java
! hat/core/src/main/java/hat/buffer/Float4.java
! hat/core/src/main/java/hat/codebuilders/BabylonOpBuilder.java
! hat/core/src/main/java/hat/codebuilders/C99HATKernelBuilder.java
! hat/core/src/main/java/hat/codebuilders/CodeBuilder.java
! hat/core/src/main/java/hat/dialect/HATF16BinaryOp.java
! hat/core/src/main/java/hat/dialect/HATVectorBinaryOp.java
! hat/core/src/main/java/hat/dialect/HATVectorLoadOp.java
+ hat/core/src/main/java/hat/dialect/HATVectorMakeOfOp.java
+ hat/core/src/main/java/hat/dialect/HATVectorOfOp.java
+ hat/core/src/main/java/hat/dialect/HATVectorOp.java
! hat/core/src/main/java/hat/dialect/HATVectorSelectLoadOp.java
! hat/core/src/main/java/hat/dialect/HATVectorSelectStoreOp.java
! hat/core/src/main/java/hat/dialect/HATVectorStoreView.java
! hat/core/src/main/java/hat/dialect/HATVectorVarLoadOp.java
! hat/core/src/main/java/hat/dialect/HATVectorVarOp.java
- hat/core/src/main/java/hat/dialect/HATVectorViewOp.java
! hat/core/src/main/java/hat/phases/HATDialectifyTier.java
! hat/core/src/main/java/hat/phases/HATDialectifyVectorOpPhase.java
! hat/core/src/main/java/hat/phases/HATDialectifyVectorSelectPhase.java
! hat/core/src/main/java/hat/phases/HATDialectifyVectorStorePhase.java
! hat/tests/src/main/java/hat/test/TestVectorTypes.java
! hat/tests/src/main/java/hat/test/engine/HatAsserts.java
! hat/tools/src/main/java/hat/tools/text/JavaHATCodeBuilder.java


From jfumero at openjdk.org  Fri Oct 31 09:28:16 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Fri, 31 Oct 2025 09:28:16 GMT
Subject: [code-reflection] RFR: Second Proposal to generalize handling of
 Vector Types based on 644
Message-ID: <eYskseFAJp2ELA7iAne_ZLJM4rXjSOKacyLqFI_UVr8=.e7c2ebd8-7352-4e0d-ae53-bcd5368f62a5@github.com>

This PR adds another proposal for generalising HAT Vector Types. 

It is based on PR #644 as an alternative approach. So we can evaluate each proposal and device which one to move forward. 

This approach replaces the `@HATVectorType` annotation with a set of interfaces to identify base type and width of the vector type being represented. 

For instance:


public interface _V {
    PrimitiveType type();
    int width();
}

public interface _V4 extends _V {

    @CodeReflection
    @Override
    default int width() {
        return 4;
    }
}


public interface Float4 extends _V4 {

    float x();
    float y();
    float z();
    float w();

    @CodeReflection
    @Override
    default PrimitiveType type() {
        return JavaType.FLOAT;
    }

   ...

}


Note that both methods, width and type, are annotated with `@CodeReflection`. During the "dialectify" process, this approach builds the code models for both methods and inspects the meta-data values. 

Some restrictions:

- `Float4` is not an `iFace` interface due to the small granularity. 
- Usually, these types are inspected from the kernel code, so we don't have instance objects.
- Returning `PrimitiveType` from the `type` method allows us to be restricted in the amount of base types to support. 


How to test:

For OpenCL:


HAT=SHOW_CODE,SHOW_COMPILATION_PHASES java @hat/otest ffi-opencl hat.test.TestVectorTypes


For CUDA:



HAT=SHOW_CODE,SHOW_COMPILATION_PHASES java @hat/otest ffi-cuda hat.test.TestVectorTypes


This PR has been tested with both OpenCL and CUDA backends.

-------------

Commit messages:
 - Merge branch 'code-reflection' into hat/vectors/generalv2
 - Merge remote-tracking branch 'refs/remotes/personal/hat/vectors/generalv2' into hat/vectors/generalv2
 - Merge remote-tracking branch 'refs/remotes/personal/hat/vectors/generalv2' into hat/vectors/generalv2
 - [hat] CUDA backend fixed
 - [hat] RW vector float4 sample added
 - [hat] missing header license added
 - [hat] Generalise vectorisation type system and dialectify for HAT v2
 - [hat] refactor for vector phases and scaffolding for vectors and interfaces
 - minor change
 - [hat] Proposal to generalise dialect for vector types
 - ... and 14 more: https://git.openjdk.org/babylon/compare/420c933a...6166cd63

Changes: https://git.openjdk.org/babylon/pull/647/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=647&range=00
  Stats: 724 lines in 25 files changed: 468 ins; 147 del; 109 mod
  Patch: https://git.openjdk.org/babylon/pull/647.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/647/head:pull/647

PR: https://git.openjdk.org/babylon/pull/647

From jfumero at openjdk.org  Fri Oct 31 09:38:41 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Fri, 31 Oct 2025 09:38:41 GMT
Subject: [code-reflection] RFR: Second Proposal to generalize handling of
 Vector Types based on 644 [v2]
In-Reply-To: <eYskseFAJp2ELA7iAne_ZLJM4rXjSOKacyLqFI_UVr8=.e7c2ebd8-7352-4e0d-ae53-bcd5368f62a5@github.com>
References: <eYskseFAJp2ELA7iAne_ZLJM4rXjSOKacyLqFI_UVr8=.e7c2ebd8-7352-4e0d-ae53-bcd5368f62a5@github.com>
Message-ID: <kwk7nt3Ns13YwdzJmzBXlVNYF0_wfhhYjHxpENaGytU=.2b2efb25-4e06-4f19-95d8-82d29a364fc7@github.com>

> This PR adds another proposal for generalising HAT Vector Types. 
> 
> It is based on PR #644 as an alternative approach. So we can evaluate each proposal and device which one to move forward. 
> 
> This approach replaces the `@HATVectorType` annotation with a set of interfaces to identify base type and width of the vector type being represented. 
> 
> For instance:
> 
> 
> public interface _V {
>     PrimitiveType type();
>     int width();
> }
> 
> public interface _V4 extends _V {
> 
>     @CodeReflection
>     @Override
>     default int width() {
>         return 4;
>     }
> }
> 
> 
> public interface Float4 extends _V4 {
> 
>     float x();
>     float y();
>     float z();
>     float w();
> 
>     @CodeReflection
>     @Override
>     default PrimitiveType type() {
>         return JavaType.FLOAT;
>     }
> 
>    ...
> 
> }
> 
> 
> Note that both methods, width and type, are annotated with `@CodeReflection`. During the "dialectify" process, this approach builds the code models for both methods and inspects the meta-data values. 
> 
> Some restrictions:
> 
> - `Float4` is not an `iFace` interface due to the small granularity. 
> - Usually, these types are inspected from the kernel code, so we don't have instance objects.
> - Returning `PrimitiveType` from the `type` method allows us to be restricted in the amount of base types to support. 
> 
> 
> How to test:
> 
> For OpenCL:
> 
> 
> HAT=SHOW_CODE,SHOW_COMPILATION_PHASES java @hat/otest ffi-opencl hat.test.TestVectorTypes
> 
> 
> For CUDA:
> 
> 
> 
> HAT=SHOW_CODE,SHOW_COMPILATION_PHASES java @hat/otest ffi-cuda hat.test.TestVectorTypes
> 
> 
> This PR has been tested with both OpenCL and CUDA backends.

Juan Fumero has updated the pull request incrementally with one additional commit since the last revision:

  [hat] Cleanup phase utils

-------------

Changes:
  - all: https://git.openjdk.org/babylon/pull/647/files
  - new: https://git.openjdk.org/babylon/pull/647/files/6166cd63..0471a646

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=babylon&pr=647&range=01
 - incr: https://webrevs.openjdk.org/?repo=babylon&pr=647&range=00-01

  Stats: 393 lines in 5 files changed: 172 ins; 217 del; 4 mod
  Patch: https://git.openjdk.org/babylon/pull/647.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/647/head:pull/647

PR: https://git.openjdk.org/babylon/pull/647

From jfumero at openjdk.org  Fri Oct 31 09:43:41 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Fri, 31 Oct 2025 09:43:41 GMT
Subject: [code-reflection] RFR: Second Proposal to generalize handling of
 Vector Types based on 644 [v2]
In-Reply-To: <kwk7nt3Ns13YwdzJmzBXlVNYF0_wfhhYjHxpENaGytU=.2b2efb25-4e06-4f19-95d8-82d29a364fc7@github.com>
References: <eYskseFAJp2ELA7iAne_ZLJM4rXjSOKacyLqFI_UVr8=.e7c2ebd8-7352-4e0d-ae53-bcd5368f62a5@github.com>
 <kwk7nt3Ns13YwdzJmzBXlVNYF0_wfhhYjHxpENaGytU=.2b2efb25-4e06-4f19-95d8-82d29a364fc7@github.com>
Message-ID: <ohOJ70AqPdCGVrlV_ISSZ6DOxEREide3CY6rOTFg8W0=.4a8aec43-62f6-432f-ab45-79a895186017@github.com>

On Fri, 31 Oct 2025 09:38:41 GMT, Juan Fumero <jfumero at openjdk.org> wrote:

>> This PR adds another proposal for generalising HAT Vector Types. 
>> 
>> It is based on PR #644 as an alternative approach. So we can evaluate each proposal and device which one to move forward. 
>> 
>> This approach replaces the `@HATVectorType` annotation with a set of interfaces to identify base type and width of the vector type being represented. 
>> 
>> For instance:
>> 
>> 
>> public interface _V {
>>     PrimitiveType type();
>>     int width();
>> }
>> 
>> public interface _V4 extends _V {
>> 
>>     @CodeReflection
>>     @Override
>>     default int width() {
>>         return 4;
>>     }
>> }
>> 
>> 
>> public interface Float4 extends _V4 {
>> 
>>     float x();
>>     float y();
>>     float z();
>>     float w();
>> 
>>     @CodeReflection
>>     @Override
>>     default PrimitiveType type() {
>>         return JavaType.FLOAT;
>>     }
>> 
>>    ...
>> 
>> }
>> 
>> 
>> Note that both methods, width and type, are annotated with `@CodeReflection`. During the "dialectify" process, this approach builds the code models for both methods and inspects the meta-data values. 
>> 
>> Some restrictions:
>> 
>> - `Float4` is not an `iFace` interface due to the small granularity. 
>> - Usually, these types are inspected from the kernel code, so we don't have instance objects.
>> - Returning `PrimitiveType` from the `type` method allows us to be restricted in the amount of base types to support. 
>> 
>> 
>> How to test:
>> 
>> For OpenCL:
>> 
>> 
>> HAT=SHOW_CODE,SHOW_COMPILATION_PHASES java @hat/otest ffi-opencl hat.test.TestVectorTypes
>> 
>> 
>> For CUDA:
>> 
>> 
>> 
>> HAT=SHOW_CODE,SHOW_COMPILATION_PHASES java @hat/otest ffi-cuda hat.test.TestVectorTypes
>> 
>> 
>> This PR has been tested with both OpenCL and CUDA backends.
>
> Juan Fumero has updated the pull request incrementally with one additional commit since the last revision:
> 
>   [hat] Cleanup phase utils

This version works with the latest version of code-reflection and the OpenCL/CUDA backends.

-------------

PR Comment: https://git.openjdk.org/babylon/pull/647#issuecomment-3472153689

From jfumero at openjdk.org  Fri Oct 31 09:43:43 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Fri, 31 Oct 2025 09:43:43 GMT
Subject: [code-reflection] Integrated: Second Proposal to generalize handling
 of Vector Types based on 644
In-Reply-To: <eYskseFAJp2ELA7iAne_ZLJM4rXjSOKacyLqFI_UVr8=.e7c2ebd8-7352-4e0d-ae53-bcd5368f62a5@github.com>
References: <eYskseFAJp2ELA7iAne_ZLJM4rXjSOKacyLqFI_UVr8=.e7c2ebd8-7352-4e0d-ae53-bcd5368f62a5@github.com>
Message-ID: <6m3L58G7sRyukDw7ixgeNyEWScfgt2YJeVU7ypAAX3w=.50304430-c7c5-4496-af52-4cf3a7a68a1f@github.com>

On Tue, 28 Oct 2025 14:24:05 GMT, Juan Fumero <jfumero at openjdk.org> wrote:

> This PR adds another proposal for generalising HAT Vector Types. 
> 
> It is based on PR #644 as an alternative approach. So we can evaluate each proposal and device which one to move forward. 
> 
> This approach replaces the `@HATVectorType` annotation with a set of interfaces to identify base type and width of the vector type being represented. 
> 
> For instance:
> 
> 
> public interface _V {
>     PrimitiveType type();
>     int width();
> }
> 
> public interface _V4 extends _V {
> 
>     @CodeReflection
>     @Override
>     default int width() {
>         return 4;
>     }
> }
> 
> 
> public interface Float4 extends _V4 {
> 
>     float x();
>     float y();
>     float z();
>     float w();
> 
>     @CodeReflection
>     @Override
>     default PrimitiveType type() {
>         return JavaType.FLOAT;
>     }
> 
>    ...
> 
> }
> 
> 
> Note that both methods, width and type, are annotated with `@CodeReflection`. During the "dialectify" process, this approach builds the code models for both methods and inspects the meta-data values. 
> 
> Some restrictions:
> 
> - `Float4` is not an `iFace` interface due to the small granularity. 
> - Usually, these types are inspected from the kernel code, so we don't have instance objects.
> - Returning `PrimitiveType` from the `type` method allows us to be restricted in the amount of base types to support. 
> 
> 
> How to test:
> 
> For OpenCL:
> 
> 
> HAT=SHOW_CODE,SHOW_COMPILATION_PHASES java @hat/otest ffi-opencl hat.test.TestVectorTypes
> 
> 
> For CUDA:
> 
> 
> 
> HAT=SHOW_CODE,SHOW_COMPILATION_PHASES java @hat/otest ffi-cuda hat.test.TestVectorTypes
> 
> 
> This PR has been tested with both OpenCL and CUDA backends.

This pull request has now been integrated.

Changeset: cb06b855
Author:    Juan Fumero <jfumero at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/cb06b8551bcf9ca02c02023161c78a142ed70dc9
Stats:     679 lines in 24 files changed: 423 ins; 147 del; 109 mod

Second Proposal to generalize handling of Vector Types based on 644

-------------

PR: https://git.openjdk.org/babylon/pull/647

From jfumero at openjdk.org  Fri Oct 31 09:46:14 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Fri, 31 Oct 2025 09:46:14 GMT
Subject: git: openjdk/babylon: code-reflection: Second Proposal to generalize
 handling of Vector Types based on 644
Message-ID: <119d7bb0-7e57-4546-b1b1-b0d405d991e1@openjdk.org>

Changeset: cb06b855
Branch: code-reflection
Author:    Juan Fumero <jfumero at openjdk.org>
Date:      2025-10-31 09:41:20 +0000
URL:       https://git.openjdk.org/babylon/commit/cb06b8551bcf9ca02c02023161c78a142ed70dc9

Second Proposal to generalize handling of Vector Types based on 644

! hat/core/src/main/java/hat/buffer/Float4.java
= hat/core/src/main/java/hat/buffer/HATVector.java
+ hat/core/src/main/java/hat/dialect/HATPhaseUtils.java
! hat/core/src/main/java/hat/dialect/HATVectorAddOp.java
! hat/core/src/main/java/hat/dialect/HATVectorBinaryOp.java
! hat/core/src/main/java/hat/dialect/HATVectorDivOp.java
! hat/core/src/main/java/hat/dialect/HATVectorLoadOp.java
! hat/core/src/main/java/hat/dialect/HATVectorMakeOfOp.java
! hat/core/src/main/java/hat/dialect/HATVectorMulOp.java
! hat/core/src/main/java/hat/dialect/HATVectorOfOp.java
! hat/core/src/main/java/hat/dialect/HATVectorOp.java
! hat/core/src/main/java/hat/dialect/HATVectorSelectLoadOp.java
! hat/core/src/main/java/hat/dialect/HATVectorSelectStoreOp.java
! hat/core/src/main/java/hat/dialect/HATVectorStoreView.java
! hat/core/src/main/java/hat/dialect/HATVectorSubOp.java
! hat/core/src/main/java/hat/dialect/HATVectorVarLoadOp.java
! hat/core/src/main/java/hat/dialect/HATVectorVarOp.java
! hat/core/src/main/java/hat/phases/HATDialect.java
! hat/core/src/main/java/hat/phases/HATDialectifyVectorOpPhase.java
! hat/core/src/main/java/hat/phases/HATDialectifyVectorSelectPhase.java
! hat/core/src/main/java/hat/phases/HATDialectifyVectorStorePhase.java
+ hat/core/src/main/java/hat/types/_V.java
+ hat/core/src/main/java/hat/types/_V4.java
! hat/tests/src/main/java/hat/test/TestVectorTypes.java


From gfrost at openjdk.org  Fri Oct 31 10:28:32 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Fri, 31 Oct 2025 10:28:32 GMT
Subject: git: openjdk/babylon: code-reflection: Added TLDR  build+run
 instructions to README and simplified hat.java
Message-ID: <f74f3889-7980-4b6f-8020-4367d202d898@openjdk.org>

Changeset: 3db08a3e
Branch: code-reflection
Author:    Gary Frost <gfrost at openjdk.org>
Date:      2025-10-31 10:26:03 +0000
URL:       https://git.openjdk.org/babylon/commit/3db08a3e9cfe71dba1409580c4c04c8744f98af8

Added TLDR  build+run instructions to README and simplified hat.java

! hat/README.md
! hat/hat.java
! hat/hat/job.jar
! hat/hat/job/src/main/java/job/Project.java


From gfrost at openjdk.org  Fri Oct 31 10:29:28 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Fri, 31 Oct 2025 10:29:28 GMT
Subject: [code-reflection] Integrated: Added TLDR  build+run instructions to
 README and simplified hat.java
Message-ID: <vyWXiLPucyuiWDQDArlITcGtzq0WS6tmiSQWJAMnfQY=.2bcfaeac-5b21-4cf5-974d-ffe6937d1ae3@github.com>

Added a TLDR stanza to the README.md 

Also simplified the `hat.java` bld/run script.

-------------

Commit messages:
 - Added TLDR  build+run instructions to README and simplified hat.java script

Changes: https://git.openjdk.org/babylon/pull/653/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=653&range=00
  Stats: 118 lines in 4 files changed: 49 ins; 3 del; 66 mod
  Patch: https://git.openjdk.org/babylon/pull/653.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/653/head:pull/653

PR: https://git.openjdk.org/babylon/pull/653

From gfrost at openjdk.org  Fri Oct 31 10:29:29 2025
From: gfrost at openjdk.org (Gary Frost)
Date: Fri, 31 Oct 2025 10:29:29 GMT
Subject: [code-reflection] Integrated: Added TLDR  build+run instructions to
 README and simplified hat.java
In-Reply-To: <vyWXiLPucyuiWDQDArlITcGtzq0WS6tmiSQWJAMnfQY=.2bcfaeac-5b21-4cf5-974d-ffe6937d1ae3@github.com>
References: <vyWXiLPucyuiWDQDArlITcGtzq0WS6tmiSQWJAMnfQY=.2bcfaeac-5b21-4cf5-974d-ffe6937d1ae3@github.com>
Message-ID: <PN_GesunUMbVI7xXnVX29dzZuWxXnVHxmEIFhuIPypA=.8ec85e68-1f5a-452a-b49d-804740ec929e@github.com>

On Fri, 31 Oct 2025 10:22:23 GMT, Gary Frost <gfrost at openjdk.org> wrote:

> Added a TLDR stanza to the README.md 
> 
> Also simplified the `hat.java` bld/run script.

This pull request has now been integrated.

Changeset: 3db08a3e
Author:    Gary Frost <gfrost at openjdk.org>
URL:       https://git.openjdk.org/babylon/commit/3db08a3e9cfe71dba1409580c4c04c8744f98af8
Stats:     118 lines in 4 files changed: 49 ins; 3 del; 66 mod

Added TLDR  build+run instructions to README and simplified hat.java

-------------

PR: https://git.openjdk.org/babylon/pull/653

From duke at openjdk.org  Fri Oct 31 22:04:49 2025
From: duke at openjdk.org (Ruby Chen)
Date: Fri, 31 Oct 2025 22:04:49 GMT
Subject: [code-reflection] RFR: Float4 arrayView support
Message-ID: <973wTu2S2Qa4ZxbuDfy4LmdBrG74Zr5GIrLTldYRma8=.64c976d7-a59e-4c6f-a163-711f7bf022e4@github.com>

Preliminary `Float4` support for arrayViews. Currently only supported for `F32ArrayPadded` buffers, which can be accessed like a `Float4[]` as shown below:

Float4[] arr = buffer.float4ArrayView(); // for a F32ArrayPadded buffer

At the moment, to use an element in the array, the `Float4` must be loaded into a separate variable first, i.e.

Float4 a = arr[index * 4];

A value can be stored into the array through the following syntax:

arr[index * 4] = ...

-------------

Commit messages:
 - Cleaning up vector arrayView
 - Merge branch 'code-reflection' into array-view
 - Float4 arrayView support

Changes: https://git.openjdk.org/babylon/pull/646/files
  Webrev: https://webrevs.openjdk.org/?repo=babylon&pr=646&range=00
  Stats: 1644 lines in 13 files changed: 1222 ins; 399 del; 23 mod
  Patch: https://git.openjdk.org/babylon/pull/646.diff
  Fetch: git fetch https://git.openjdk.org/babylon.git pull/646/head:pull/646

PR: https://git.openjdk.org/babylon/pull/646

From jfumero at openjdk.org  Fri Oct 31 22:04:50 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Fri, 31 Oct 2025 22:04:50 GMT
Subject: [code-reflection] RFR: Float4 arrayView support
In-Reply-To: <973wTu2S2Qa4ZxbuDfy4LmdBrG74Zr5GIrLTldYRma8=.64c976d7-a59e-4c6f-a163-711f7bf022e4@github.com>
References: <973wTu2S2Qa4ZxbuDfy4LmdBrG74Zr5GIrLTldYRma8=.64c976d7-a59e-4c6f-a163-711f7bf022e4@github.com>
Message-ID: <NoJ9hGQAUOWTtn5sMsHIuwZvQrqcGhiOJZFI60XiBVM=.916407f2-1c3e-4db2-86dd-87f039b8f66e@github.com>

On Mon, 27 Oct 2025 23:56:01 GMT, Ruby Chen <duke at openjdk.org> wrote:

> Preliminary `Float4` support for arrayViews. Currently only supported for `F32ArrayPadded` buffers, which can be accessed like a `Float4[]` as shown below:
> 
> Float4[] arr = buffer.float4ArrayView(); // for a F32ArrayPadded buffer
> 
> At the moment, to use an element in the array, the `Float4` must be loaded into a separate variable first, i.e.
> 
> Float4 a = arr[index * 4];
> 
> A value can be stored into the array through the following syntax:
> 
> arr[index * 4] = ...

This is great work. I have a few comments/questions. 

All test passing also for my setup (OpenCL backend).

hat/backends/ffi/opencl/src/main/java/hat/backend/ffi/OpenCLHATKernelBuilder.java line 89:

> 87:                 .oparen();
> 88:         // if the value to be stored is an operation, recurse on the operation
> 89:         if (hatVectorStoreView.operands().get(1) instanceof Op.Result r && r.op() instanceof HATVectorBinaryOp) {

Shouldn't this be automatically visited when `HATVectorBinaryOp` is found? Do you have the use case in which this is needed?  The reason I am asking is that this also works with the prev. proposal for single Float4 loads./stores . So I wonder which pattern triggers the new check.

hat/core/src/main/java/hat/phases/HATDialectifyArrayViewPhase.java line 99:

> 97:             switch (op) {
> 98:                 case JavaOp.InvokeOp iop -> {
> 99:                     if (isVectorOperation(iop)) {

This HAT Phase is invoked after all vector operations have been inserted in the new code-model. So I wonder if this is actually triggered. My take is that, at this point in the compilation pipeline, all VectoLoads, Stores and VSelect are already in place. Is this correct?

hat/core/src/main/java/hat/phases/HATDialectifyArrayViewPhase.java line 165:

> 163:                 }
> 164:                 // handles only 1D and 2D arrays
> 165:                 case JavaOp.ArrayAccessOp.ArrayLoadOp alop -> {

My take is that this phase is hitting this `case` for all the actual array view. If the prev. block still insert the dialect ops for vectors, we could substitute the prev.. pipeline of hat phases with this one.

hat/core/src/main/java/hat/phases/HATDialectifyArrayViewPhase.java line 249:

> 247:                                     4,
> 248:                                     HATVectorViewOp.VectorType.FLOAT4,
> 249:                                     false, //TODO: fix

This `flag` was to indentify if the store is set to a private or shared array. In this case, we don't have a pointer. 

Using this search will probably work:


    private boolean findIsSharedOrPrivateSpace(CoreOp.VarAccessOp.VarLoadOp varLoadOp) {
        return findIsSharedOrPrivateSpace(varLoadOp.operands().get(0));
    }

    private boolean findIsSharedOrPrivateSpace(Value v) {
        if (v instanceof Op.Result r && r.op() instanceof CoreOp.VarAccessOp.VarLoadOp varLoadOp) {
            return findIsSharedOrPrivateSpace(varLoadOp);
        } else  if (v instanceof CoreOp.Result r && (r.op() instanceof HATLocalVarOp || r.op() instanceof HATPrivateVarOp)) {
            return true;
        }else{
            return false;
        }
    }


I think in the future we can add a new Op for expressing global array accesses too (from parameters), so we can simplify how we handle different levels of memory within the code gen.

-------------

PR Review: https://git.openjdk.org/babylon/pull/646#pullrequestreview-3386912390
PR Review Comment: https://git.openjdk.org/babylon/pull/646#discussion_r2468120990
PR Review Comment: https://git.openjdk.org/babylon/pull/646#discussion_r2468126202
PR Review Comment: https://git.openjdk.org/babylon/pull/646#discussion_r2468131031
PR Review Comment: https://git.openjdk.org/babylon/pull/646#discussion_r2468143307

From jfumero at openjdk.org  Fri Oct 31 22:04:50 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Fri, 31 Oct 2025 22:04:50 GMT
Subject: [code-reflection] RFR: Float4 arrayView support
In-Reply-To: <NoJ9hGQAUOWTtn5sMsHIuwZvQrqcGhiOJZFI60XiBVM=.916407f2-1c3e-4db2-86dd-87f039b8f66e@github.com>
References: <973wTu2S2Qa4ZxbuDfy4LmdBrG74Zr5GIrLTldYRma8=.64c976d7-a59e-4c6f-a163-711f7bf022e4@github.com>
 <NoJ9hGQAUOWTtn5sMsHIuwZvQrqcGhiOJZFI60XiBVM=.916407f2-1c3e-4db2-86dd-87f039b8f66e@github.com>
Message-ID: <3yJGeHdP2iNHaFCBvwLX8TeQfCBNUX52jr-B1gUAW7U=.5a705205-3afd-40ba-86d0-757533c6dd65@github.com>

On Tue, 28 Oct 2025 06:50:35 GMT, Juan Fumero <jfumero at openjdk.org> wrote:

>> Preliminary `Float4` support for arrayViews. Currently only supported for `F32ArrayPadded` buffers, which can be accessed like a `Float4[]` as shown below:
>> 
>> Float4[] arr = buffer.float4ArrayView(); // for a F32ArrayPadded buffer
>> 
>> At the moment, to use an element in the array, the `Float4` must be loaded into a separate variable first, i.e.
>> 
>> Float4 a = arr[index * 4];
>> 
>> A value can be stored into the array through the following syntax:
>> 
>> arr[index * 4] = ...
>
> hat/backends/ffi/opencl/src/main/java/hat/backend/ffi/OpenCLHATKernelBuilder.java line 89:
> 
>> 87:                 .oparen();
>> 88:         // if the value to be stored is an operation, recurse on the operation
>> 89:         if (hatVectorStoreView.operands().get(1) instanceof Op.Result r && r.op() instanceof HATVectorBinaryOp) {
> 
> Shouldn't this be automatically visited when `HATVectorBinaryOp` is found? Do you have the use case in which this is needed?  The reason I am asking is that this also works with the prev. proposal for single Float4 loads./stores . So I wonder which pattern triggers the new check.

We will need to propagate this change also for the CUDA codegen.

-------------

PR Review Comment: https://git.openjdk.org/babylon/pull/646#discussion_r2468132770

From duke at openjdk.org  Fri Oct 31 22:04:51 2025
From: duke at openjdk.org (Ruby Chen)
Date: Fri, 31 Oct 2025 22:04:51 GMT
Subject: [code-reflection] RFR: Float4 arrayView support
In-Reply-To: <3yJGeHdP2iNHaFCBvwLX8TeQfCBNUX52jr-B1gUAW7U=.5a705205-3afd-40ba-86d0-757533c6dd65@github.com>
References: <973wTu2S2Qa4ZxbuDfy4LmdBrG74Zr5GIrLTldYRma8=.64c976d7-a59e-4c6f-a163-711f7bf022e4@github.com>
 <NoJ9hGQAUOWTtn5sMsHIuwZvQrqcGhiOJZFI60XiBVM=.916407f2-1c3e-4db2-86dd-87f039b8f66e@github.com>
 <3yJGeHdP2iNHaFCBvwLX8TeQfCBNUX52jr-B1gUAW7U=.5a705205-3afd-40ba-86d0-757533c6dd65@github.com>
Message-ID: <_1u5TCXLpIlETxYZXIJsMCbZWjmXqMSTsW6L2aX2Kds=.4b7facb0-fe56-4c6d-9d7f-edb88a00b319@github.com>

On Tue, 28 Oct 2025 06:56:53 GMT, Juan Fumero <jfumero at openjdk.org> wrote:

>> hat/backends/ffi/opencl/src/main/java/hat/backend/ffi/OpenCLHATKernelBuilder.java line 89:
>> 
>>> 87:                 .oparen();
>>> 88:         // if the value to be stored is an operation, recurse on the operation
>>> 89:         if (hatVectorStoreView.operands().get(1) instanceof Op.Result r && r.op() instanceof HATVectorBinaryOp) {
>> 
>> Shouldn't this be automatically visited when `HATVectorBinaryOp` is found? Do you have the use case in which this is needed?  The reason I am asking is that this also works with the prev. proposal for single Float4 loads./stores . So I wonder which pattern triggers the new check.
>
> We will need to propagate this change also for the CUDA codegen.

Good point; currently, a line like `vC[index * 4] = Float4.add(vA[index * 4], vB[index * 4])` won't be handled properly (the add operation won't be recursed over, and the resulting code will look like `vstore4(vA, 0, &c->array[index*4])` instead of storing the add operation's result). 

(Edit: I originally thought that if we extracted the array accesses `vA[index * 4]` and `vB[index * 4]` into Float4 variables first before operating on them, the original code would work, but it doesn't seem to be the case)

There might also be something I'm missing that causes a case like `vC[index * 4] = Float4.add(vA[index * 4], vB[index * 4])` to fail with the original code; I'll give it another pass.

-------------

PR Review Comment: https://git.openjdk.org/babylon/pull/646#discussion_r2470395187

From jfumero at openjdk.org  Fri Oct 31 22:04:52 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Fri, 31 Oct 2025 22:04:52 GMT
Subject: [code-reflection] RFR: Float4 arrayView support
In-Reply-To: <_1u5TCXLpIlETxYZXIJsMCbZWjmXqMSTsW6L2aX2Kds=.4b7facb0-fe56-4c6d-9d7f-edb88a00b319@github.com>
References: <973wTu2S2Qa4ZxbuDfy4LmdBrG74Zr5GIrLTldYRma8=.64c976d7-a59e-4c6f-a163-711f7bf022e4@github.com>
 <NoJ9hGQAUOWTtn5sMsHIuwZvQrqcGhiOJZFI60XiBVM=.916407f2-1c3e-4db2-86dd-87f039b8f66e@github.com>
 <3yJGeHdP2iNHaFCBvwLX8TeQfCBNUX52jr-B1gUAW7U=.5a705205-3afd-40ba-86d0-757533c6dd65@github.com>
 <_1u5TCXLpIlETxYZXIJsMCbZWjmXqMSTsW6L2aX2Kds=.4b7facb0-fe56-4c6d-9d7f-edb88a00b319@github.com>
Message-ID: <43Eu8jrIPnMBuk4UM-_IiUqu9OmwlI2uCW6o-fZL4U4=.b81319e7-4f83-45a7-ba86-41892179b47c@github.com>

On Tue, 28 Oct 2025 17:16:23 GMT, Ruby Chen <duke at openjdk.org> wrote:

>> We will need to propagate this change also for the CUDA codegen.
>
> Good point; currently, a line like `vC[index * 4] = Float4.add(vA[index * 4], vB[index * 4])` won't be handled properly (the add operation won't be recursed over, and the resulting code will look like `vstore4(vA, 0, &c->array[index*4])` instead of storing the add operation's result). 
> 
> (Edit: I originally thought that if we extracted the array accesses `vA[index * 4]` and `vB[index * 4]` into Float4 variables first before operating on them, the original code would work, but it doesn't seem to be the case)
> 
> There might also be something I'm missing that causes a case like `vC[index * 4] = Float4.add(vA[index * 4], vB[index * 4])` to fail with the original code; I'll give it another pass.

Ok, got it. Thanks for the clarification. So it looks to me that possibly there are some missing `HATVectorVarLoadOps` in this case. 

For example:

`Float4.add(vA[index * 4], vB[index * 4])`, would have two `HATVectorVarLoadOp` loading a `float4`. But I get now, with the array views, we will see actually two plain arrays being loaded. So if that's the case, it makes sense to invoke `recurse`. 

I am not suggesting to use `HATVectorVarLoadOp`, but it is good we have the cases identified.

-------------

PR Review Comment: https://git.openjdk.org/babylon/pull/646#discussion_r2472182044

From duke at openjdk.org  Fri Oct 31 22:04:54 2025
From: duke at openjdk.org (Ruby Chen)
Date: Fri, 31 Oct 2025 22:04:54 GMT
Subject: [code-reflection] RFR: Float4 arrayView support
In-Reply-To: <NoJ9hGQAUOWTtn5sMsHIuwZvQrqcGhiOJZFI60XiBVM=.916407f2-1c3e-4db2-86dd-87f039b8f66e@github.com>
References: <973wTu2S2Qa4ZxbuDfy4LmdBrG74Zr5GIrLTldYRma8=.64c976d7-a59e-4c6f-a163-711f7bf022e4@github.com>
 <NoJ9hGQAUOWTtn5sMsHIuwZvQrqcGhiOJZFI60XiBVM=.916407f2-1c3e-4db2-86dd-87f039b8f66e@github.com>
Message-ID: <MoHHoJulRYMkwzoJcSj0fpYwqZycm38affK9k-kjDCQ=.f040ff84-434e-494b-ae48-ba3d67bc969f@github.com>

On Tue, 28 Oct 2025 06:53:27 GMT, Juan Fumero <jfumero at openjdk.org> wrote:

>> Preliminary `Float4` support for arrayViews. Currently only supported for `F32ArrayPadded` buffers, which can be accessed like a `Float4[]` as shown below:
>> 
>> Float4[] arr = buffer.float4ArrayView(); // for a F32ArrayPadded buffer
>> 
>> At the moment, to use an element in the array, the `Float4` must be loaded into a separate variable first, i.e.
>> 
>> Float4 a = arr[index * 4];
>> 
>> A value can be stored into the array through the following syntax:
>> 
>> arr[index * 4] = ...
>
> hat/core/src/main/java/hat/phases/HATDialectifyArrayViewPhase.java line 99:
> 
>> 97:             switch (op) {
>> 98:                 case JavaOp.InvokeOp iop -> {
>> 99:                     if (isVectorOperation(iop)) {
> 
> This HAT Phase is invoked after all vector operations have been inserted in the new code-model. So I wonder if this is actually triggered. My take is that, at this point in the compilation pipeline, all VectoLoads, Stores and VSelect are already in place. Is this correct?

That's correct; this is leftover from when I was testing earlier haha. Good catch!

> hat/core/src/main/java/hat/phases/HATDialectifyArrayViewPhase.java line 165:
> 
>> 163:                 }
>> 164:                 // handles only 1D and 2D arrays
>> 165:                 case JavaOp.ArrayAccessOp.ArrayLoadOp alop -> {
> 
> My take is that this phase is hitting this `case` for all the actual array view. If the prev. block still insert the dialect ops for vectors, we could substitute the prev.. pipeline of hat phases with this one.

Oh interesting, are you proposing combining some of the hat phases into one larger overall phase using this `case` logic?

-------------

PR Review Comment: https://git.openjdk.org/babylon/pull/646#discussion_r2470578540
PR Review Comment: https://git.openjdk.org/babylon/pull/646#discussion_r2470585428

From jfumero at openjdk.org  Fri Oct 31 22:04:56 2025
From: jfumero at openjdk.org (Juan Fumero)
Date: Fri, 31 Oct 2025 22:04:56 GMT
Subject: [code-reflection] RFR: Float4 arrayView support
In-Reply-To: <MoHHoJulRYMkwzoJcSj0fpYwqZycm38affK9k-kjDCQ=.f040ff84-434e-494b-ae48-ba3d67bc969f@github.com>
References: <973wTu2S2Qa4ZxbuDfy4LmdBrG74Zr5GIrLTldYRma8=.64c976d7-a59e-4c6f-a163-711f7bf022e4@github.com>
 <NoJ9hGQAUOWTtn5sMsHIuwZvQrqcGhiOJZFI60XiBVM=.916407f2-1c3e-4db2-86dd-87f039b8f66e@github.com>
 <MoHHoJulRYMkwzoJcSj0fpYwqZycm38affK9k-kjDCQ=.f040ff84-434e-494b-ae48-ba3d67bc969f@github.com>
Message-ID: <ELMKm3iuissPBhB9DQ9jbLW0gPEJa_qtezaV3hqhzwo=.61185520-3d6c-45ef-bde6-2b2472434c13@github.com>

On Tue, 28 Oct 2025 18:13:01 GMT, Ruby Chen <duke at openjdk.org> wrote:

>> hat/core/src/main/java/hat/phases/HATDialectifyArrayViewPhase.java line 165:
>> 
>>> 163:                 }
>>> 164:                 // handles only 1D and 2D arrays
>>> 165:                 case JavaOp.ArrayAccessOp.ArrayLoadOp alop -> {
>> 
>> My take is that this phase is hitting this `case` for all the actual array view. If the prev. block still insert the dialect ops for vectors, we could substitute the prev.. pipeline of hat phases with this one.
>
> Oh interesting, are you proposing combining some of the hat phases into one larger overall phase using this `case` logic?

If the whole logic works, then yes, we can condense it in one phase. I personally prefer different compiler phases, so if something goes wrong, or we find a new use-case that does not match the patterns for one of the phases, we can easily detect phases involved. But not strong opinion.

-------------

PR Review Comment: https://git.openjdk.org/babylon/pull/646#discussion_r2472259888