On Wed, 29 Jul 2026 21:29:39 GMT, Vladimir Ivanov <[email protected]> wrote:
>> Jatin Bhateja has updated the pull request incrementally with one additional >> commit since the last revision: >> >> Review comments resolution > > Thanks for the detailed analysis. I agree with the results. > >> Would it be fine to bring Float16 FMA to parity with single/double precision >> in this PR, and experiment with generalized encoding selection for all three >> (FmaVF/VD/VHF) in a follow-up? > > Sure. Please, file an RFE. > > I submitted the PR for testing. Hi @iwanowww , Your comments have been addressed. > src/hotspot/cpu/x86/assembler_x86.hpp line 2563: > >> 2561: void vdivsh(XMMRegister dst, XMMRegister nds, XMMRegister src); >> 2562: void vsqrtsh(XMMRegister dst, XMMRegister src); >> 2563: void vfmadd132sh(XMMRegister dst, XMMRegister src1, XMMRegister >> src2); > > Do we still need 132 instruction variants? They are effectively unused after > the change. Addressed ------------- PR Comment: https://git.openjdk.org/jdk/pull/31724#issuecomment-5162491625 PR Review Comment: https://git.openjdk.org/jdk/pull/31724#discussion_r3701432199
