[all-commits] [llvm/llvm-project] 60941f: [NVPTX] Lower v2f16 and v2bf16 stores as 32-bit sc...

[Artem Belevich via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 60941f1d28ce17958d1e931711d93b8f629d4ad3
      https://github.com/llvm/llvm-project/commit/60941f1d28ce17958d1e931711d93b8f629d4ad3
  Author: Artem Belevich <tra at google.com>
  Date:   2023-06-23 (Fri, 23 Jun 2023)

  Changed paths:
    M llvm/lib/Target/NVPTX/NVPTXISelLowering.cpp
    M llvm/test/CodeGen/NVPTX/f16x2-instructions.ll

  Log Message:
  -----------
  [NVPTX] Lower v2f16 and v2bf16 stores as 32-bit scalars.

This avoids unnecessary vector splitting that was needed for vectorized store
instruction.

Differential Revision: https://reviews.llvm.org/D152593