[llvm] [AMDGPU] Form V_MAD_U64_U32 from mul24 (PR #72393)
Jay Foad via llvm-commits
llvm-commits at lists.llvm.org
Fri Dec 8 06:43:23 PST 2023
================
@@ -678,9 +678,26 @@ multiclass IMAD32_Pats <VOP3_Pseudo inst> {
>;
}
+// Handle cases where amdgpu-codegenprepare-mul24 made a mul24 instead of a normal mul.
+// We need to separate this because otherwise OtherPredicates would be overriden.
+multiclass IMAD32_Mul24_Pats <VOP3_Pseudo inst> {
+ def : GCNPat <
+ (i64 (add (i64 (AMDGPUmul_u24 i32:$src0, i32:$src1)), i64:$src2)),
+ (inst $src0, $src1, $src2, 0 /* clamp */)
+ >;
----------------
jayfoad wrote:
Agreed, if the multiply is followed by a full 64 bit add then we should always use mad_u64 regardless of speed model.
https://github.com/llvm/llvm-project/pull/72393
More information about the llvm-commits
mailing list