JDK-8324724 : Add Stub routines for FP16 conversions on aarch64
  • Type: Enhancement
  • Component: hotspot
  • Sub-Component: compiler
  • Affected Version: 23
  • Priority: P4
  • Status: Resolved
  • Resolution: Fixed
  • OS: generic
  • CPU: aarch64
  • Submitted: 2024-01-25
  • Updated: 2024-02-12
  • Resolved: 2024-02-05
The Version table provides details related to the release that this issue/RFE will be addressed.

Unresolved : Release in which this issue/RFE will be addressed.
Resolved: Release in which this issue/RFE has been resolved.
Fixed : Release in which this issue/RFE has been fixed. The release containing this fix may be available for download as an Early Access Release or a General Availability Release.

To download the current JDK release, click here.
JDK 23
23 b09Fixed
Related Reports
Relates :  
JDK-8302976 - C2 Intrinsification of Float.floatToFloat16 and Float.float16ToFloat Yields Different Result than the Interpreter
Relates :  
JDK-8325264 - two compiler/intrinsics/float16 tests fail after JDK-8324724
Description
This commit - https://github.com/openjdk/jdk/commit/8cfd74f76afc9e5d50c52104fef9974784718dd4 adds stub routines on x86 for FP16 conversions for float/float16 constants to get an accurate compile time value of the node. This task adds similar stub routines for aarch64.
Comments
Changeset: 51853f74 Author: Bhavana Kilambi <bkilambi@openjdk.org> Committer: Nick Gasson <ngasson@openjdk.org> Date: 2024-02-05 18:29:32 +0000 URL: https://git.openjdk.org/jdk/commit/51853f7488afa69c0d14b0e96f1da84822cd83f1
05-02-2024
A pull request was submitted for review. URL: https://git.openjdk.org/jdk/pull/17706 Date: 2024-02-05 09:12:23 +0000
05-02-2024
Thanks [~bkilambi] for the update. I'm therefore turning this into a performance RFE.
26-01-2024
Yes, I remember aarch64 also being affected by inconsistency in results between interpreter, c1 and c2 before it was fixed in https://bugs.openjdk.org/browse/JDK-8302976. However, this is just a performance enhancement.
26-01-2024
[~bkilambi] does that mean that AArch64 is affected by JDK-8302976 as well? I.e., incorrect/different results or is adding the stubs just a performance enhancement?
26-01-2024