[llvm] [AMDGPU] - Add s_quadmask intrinsics (PR #70804)

Jessica Del via llvm-commits llvm-commits at lists.llvm.org
Tue Oct 31 07:41:24 PDT 2023


================
@@ -326,8 +326,10 @@ def S_XNOR_SAVEEXEC_B64 : SOP1_64 <"s_xnor_saveexec_b64">;
 
 } // End hasSideEffects = 1, Uses = [EXEC], Defs = [EXEC, SCC]
 
-def S_QUADMASK_B32 : SOP1_32 <"s_quadmask_b32">;
-def S_QUADMASK_B64 : SOP1_64 <"s_quadmask_b64">;
+def S_QUADMASK_B32 : SOP1_32 <"s_quadmask_b32",
+[(set i32:$sdst, (int_amdgcn_s_quadmask i32:$src0))]>;
----------------
OutOfCache wrote:

I indented it two spaces, as I have seen in other instances of these patterns. Is it better to put the `>;` on the next line? I saw it a couple times, but not every time.

https://github.com/llvm/llvm-project/pull/70804


More information about the llvm-commits mailing list