[llvm] [AMDGPU] Extend llvm.amdgcn.update.dpp intrinsic to support f64 (PR #91190)

Jay Foad via llvm-commits llvm-commits at lists.llvm.org
Mon May 6 09:33:06 PDT 2024


================
@@ -3338,14 +3338,17 @@ def : GCNPat <
                         (as_i1timm $bound_ctrl))
 >;
 
-def : GCNPat <
-  (i64 (int_amdgcn_update_dpp i64:$old, i64:$src, timm:$dpp_ctrl, timm:$row_mask,
+class UpdateDPP64Pat<ValueType vt> : GCNPat <
----------------
jayfoad wrote:

Nit: instead of creating a class you could do this with `foreach vt = [i64, f64] in ...`

https://github.com/llvm/llvm-project/pull/91190


More information about the llvm-commits mailing list