On Mon, 28 Apr 2025 19:41:28 GMT, Vladimir Ivanov <vliva...@openjdk.org> wrote:
> Both SVML and SLEEF libraries reuse 128-bit versions for vector of 2 floats. > Usually, only `Float64Vector` shape satisfies that condiditon, but > with`-XX:MaxVectorSize=8` `FloatMaxVector` becomes 2 element vector as well. > > Adjust the relevant logic to match vector of 2 floats shape structurally. > > Testing: failing regression test, hs-tier1 - hs-tier4 This pull request has now been integrated. Changeset: 6850757f Author: Vladimir Ivanov <vliva...@openjdk.org> URL: https://git.openjdk.org/jdk/commit/6850757fe6c6c7868fab2314babb7c30bce2662e Stats: 4 lines in 1 file changed: 2 ins; 0 del; 2 mod 8355689: Wrong native entry name for FloatMaxVector vector math stubs with -XX:MaxVectorSize=8 Reviewed-by: psandoz, haosun, kvn ------------- PR: https://git.openjdk.org/jdk/pull/24936
