[lworld+fp16] Integrated: 8341414: Add support for FP16 conversion routines

[Bhavana Kilambi]

On Wed, 23 Oct 2024 14:07:52 GMT, Bhavana Kilambi <bkilambi at openjdk.org> wrote:

> This patch adds intrinsic support for FP16 conversion routines to int/long/double and also the aarch64 backend support. This patch implements both scalar and vector versions for these conversions.
> 
> Performance numbers on aarch64 machine with SVE support :
> 
> 
> Benchmark                         (vectorDim)   Gain
> Float16OpsBenchmark.fp16ToDouble  1024          18.23
> Float16OpsBenchmark.fp16ToInt     1024          1.93
> Float16OpsBenchmark.fp16ToLong    1024          3.95
> 
> 
> The Gain column is the ratio between thrpt of this patch and the thrpt with the intrinsics disabled (which generates FP32 arithmetic).

This pull request has now been integrated.

Changeset: 6bcf899e
Author:    Bhavana Kilambi <bkilambi at openjdk.org>
URL:       https://git.openjdk.org/valhalla/commit/6bcf899e70f2615216166da009f28c2a3d5778f8
Stats:     1580 lines in 31 files changed: 771 ins; 74 del; 735 mod

8341414: Add support for FP16 conversion routines

Reviewed-by: jbhateja

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

PR: https://git.openjdk.org/valhalla/pull/1283