RFR: Save/load nmethod without going through CodeBuffer [v6]
Ashutosh Mehra
asmehra at openjdk.org
Mon Mar 31 08:40:50 UTC 2025
On Fri, 28 Mar 2025 11:51:12 GMT, Ashutosh Mehra <asmehra at openjdk.org> wrote:
>> This is the prototype for storing and loading nmethods without going through the CodeBuffer.
>> The new implementation is protected by the flag -XX:+UseNewCode2.
>>
>> Some numbers using this implementation:
>> spring-boot-getting-started [0] shows startup improvement of ~ 7.5%% and quarkus-getting-started [1] shows improvement of around ~3.5%.
>>
>> Numbers for Springboot
>>
>> Old build = /home/asmehra/data/ashu-mehra/leyden/build/nmethod-single-copy-load-release/images/jdk with options -XX:+UnlockDiagnosticVMOptions -XX:-UseNewCode2
>> New build = /home/asmehra/data/ashu-mehra/leyden/build/nmethod-single-copy-load-release/images/jdk with options -XX:+UnlockDiagnosticVMOptions -XX:+UseNewCode2
>> Run,Old CDS + AOT,New CDS + AOT
>> 1,544,512
>> 2,554,525
>> 3,550,508
>> 4,571,515
>> 5,550,506
>> 6,552,515
>> 7,568,510
>> 8,554,517
>> 9,551,506
>> 10,556,508
>> Geomean,554.94,512.17
>> Stdev,7.90,5.65
>>
>> Numbers for Quarkus:
>>
>> Old build = /home/asmehra/data/ashu-mehra/leyden/build/nmethod-single-copy-load-release/images/jdk with options -XX:+UnlockDiagnosticVMOptions -XX:-UseNewCode2
>> New build = /home/asmehra/data/ashu-mehra/leyden/build/nmethod-single-copy-load-release/images/jdk with options -XX:+UnlockDiagnosticVMOptions -XX:+UseNewCode2
>> Run,Old CDS + AOT,New CDS + AOT
>> 1,359,346
>> 2,360,353
>> 3,373,357
>> 4,376,356
>> 5,366,353
>> 6,360,356
>> 7,361,361
>> 8,347,349
>> 9,355,336
>> 10,374,342
>> Geomean,363.00,350.82
>> Stdev,8.70,7.27
>>
>>
>> -Xlog:init logs the load time from AOT code cache at JVM exit.
>> For spring-boot-getting-started without `UseNewCode2`:
>>
>> [3.459s][info][init] SC Load Time: 0.202 s
>> [3.459s][info][init] nmethod register: 0.135 s
>> [3.459s][info][init] find cached code: 0.007 s
>>
>>
>> For spring-boot-getting-started with `UseNewCode2`:
>>
>> [3.192s][info][init]
>> [3.192s][info][init] SC Load Time: 0.138 s
>> [3.192s][info][init] nmethod register: 0.111 s
>> [3.192s][info][init] find cached code: 0.006 s
>>
>>
>> For quarkus-getting-started without `UseNewCode2`
>>
>> [0.392s][info][init] SC Load Time: 0.060 s
>> [0.392s][info][init] nmethod register: 0.039 s
>> [0.392s][info][init] find cached code: 0.002 s
>>
>>
>> For quarkus-getting-started with `UseNewCode2`
>>
>> [0.386s][info][init] SC Load Time: 0.033 s
>> [0.386s][info][init] nmethod register: 0.027 s
>> [0.386s][info][init] find cached code: ...
>
> Ashutosh Mehra 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 17 additional commits since the last revision:
>
> - Merge branch 'premain' into nmethod-single-copy-load
> - Rename write_nmethod_extra_relocations to write_nmethod_loadtime_relocations
>
> Signed-off-by: Ashutosh Mehra <asmehra at redhat.com>
> - Flush code block when loading the nmethod from archive
>
> Signed-off-by: Ashutosh Mehra <asmehra at redhat.com>
> - Some more cleanup
>
> Signed-off-by: Ashutosh Mehra <asmehra at redhat.com>
> - Set compile id
>
> Signed-off-by: Ashutosh Mehra <asmehra at redhat.com>
> - Minor cleanup
>
> Signed-off-by: Ashutosh Mehra <asmehra at redhat.com>
> - Handle mutable nmethod data properly when storing/loading from AOT code
> cache
>
> Signed-off-by: Ashutosh Mehra <asmehra at redhat.com>
> - Fix win compile failures
>
> Signed-off-by: Ashutosh Mehra <asmehra at redhat.com>
> - Remove trailing whitespaces
>
> Signed-off-by: Ashutosh Mehra <asmehra at redhat.com>
> - Include os.hpp to fix compile failures
>
> Signed-off-by: Ashutosh Mehra <asmehra at redhat.com>
> - ... and 7 more: https://git.openjdk.org/leyden/compare/d5429885...45b086ad
I got some numbers for aarch64 as well and they are in similar range as for x86-64
On a 128 cpu server:
Old build = /tmp/ashu/leyden/build/premain-release/images/jdk with options
New build = /tmp/ashu/leyden/build/nmethod-single-copy-load-release/images/jdk with options
Run,Old CDS + AOT,New CDS + AOT
1,463,444 2,457,443 3,465,453
4,452,447
5,460,446
6,458,443
7,465,445
8,464,454
9,461,449
10,455,449
Geomean,459.98,447.28
Stdev,4.22,3.72
On same 128 cpu server but bound to 0-7 cpus:
Old build = /tmp/ashu/leyden/build/premain-release/images/jdk with options
New build = /tmp/ashu/leyden/build/nmethod-single-copy-load-release/images/jdk with options
Run,Old CDS + AOT,New CDS + AOT
1,441,424
2,449,420
3,435,427
4,429,427
5,429,430
6,437,424
7,441,425
8,447,426
9,429,432
10,438,421
Geomean,437.45,425.59
Stdev,6.86,3.50
Load time as reported by `Xlog:init` for premain:
[0.680s][info][init] SC Load Time: 0.210 s
[0.680s][info][init] nmethod register: 0.148 s
[0.680s][info][init] find cached code: 0.004 s
Load time as reported by `Xlog:init` for this PR:
[0.675s][info][init] SC Load Time: 0.143 s
[0.675s][info][init] nmethod register: 0.128 s
[0.675s][info][init] find cached code: 0.004 s
-------------
PR Comment: https://git.openjdk.org/leyden/pull/27#issuecomment-2765510815
More information about the leyden-dev
mailing list