[all-commits] [llvm/llvm-project] c29d82: [AMDGPU] Deallocate VGPRs before exiting in dynami...

[Diana Picus via All-commits]

  Branch: refs/heads/users/rovka/dvgpr-3
  Home:   https://github.com/llvm/llvm-project
  Commit: c29d8202c06488a9466aea49dda4cf2b4663236e
      https://github.com/llvm/llvm-project/commit/c29d8202c06488a9466aea49dda4cf2b4663236e
  Author: Diana Picus <Diana-Magda.Picus at amd.com>
  Date:   2025-03-06 (Thu, 06 Mar 2025)

  Changed paths:
    M llvm/lib/Target/AMDGPU/SIInsertWaitcnts.cpp
    A llvm/test/CodeGen/AMDGPU/release-vgprs-gfx12.mir

  Log Message:
  -----------
  [AMDGPU] Deallocate VGPRs before exiting in dynamic VGPR mode

In dynamic VGPR mode, Waves must deallocate all VGPRs before exiting. If
the shader program does not do this, hardware inserts `S_ALLOC_VGPR 0`
before S_ENDPGM, but this may incur some performance cost. Therefore
it's better if the compiler proactively generates that instruction.

This patch extends `si-insert-waitcnts` to deallocate the VGPRs via
a `S_ALLOC_VGPR 0` before any `S_ENDPGM` when in dynamic VGPR mode.



To unsubscribe from these emails, change your notification settings at https://github.com/llvm/llvm-project/settings/notifications