[llvm] [LoopIdiom] Select llvm.experimental.memset.pattern intrinsic rather than memset_pattern16 libcall (PR #126736)

Alex Bradbury via llvm-commits llvm-commits at lists.llvm.org
Wed Jul 9 05:41:47 PDT 2025


================
@@ -150,6 +150,11 @@ static cl::opt<bool> UseLIRCodeSizeHeurs(
              "with -Os/-Oz"),
     cl::init(true), cl::Hidden);
 
+static cl::opt<bool> ForceMemsetPatternIntrinsic(
----------------
asb wrote:

I think force enabling of a TLI function is slightly orthogonal. This patch aims to introduce memset.pattern in just those cases where memset_pattern16 would have been available, even though we _do_ have codegen support for targets without memset_pattern16. The rationale is that using memset.pattern on all targets in this way would be a big codegen change. So the flag means it's easy to test the more eager selection of the intrinsic.

https://github.com/llvm/llvm-project/pull/126736


More information about the llvm-commits mailing list