[llvm] [AMDGPU] Introduce iglp_opt(2): Generalized exp/mfma interleaving for select kernels (PR #81342)
Jeffrey Byrnes via llvm-commits
llvm-commits at lists.llvm.org
Fri Feb 23 16:56:00 PST 2024
================
@@ -1726,6 +1749,8 @@ bool MFMAExpInterleaveOpt::applyIGLPStrategy(
// In loop MFMAs
auto MFMAInLoop = MFMAPipeCount - (MFMAEnablement * 2);
auto MFMALoopCount = MFMAInLoop / MFMARatio;
+ auto VALUOps =
----------------
jrbyrnes wrote:
Added a couple checks to be safe, but we should be good
https://github.com/llvm/llvm-project/pull/81342
More information about the llvm-commits
mailing list