[llvm] [AMDGPU] Allow dpp in v_pk_fmac_f16 for GFX9 and GFX10 (PR #144782)

Stanislav Mekhanoshin via llvm-commits llvm-commits at lists.llvm.org
Thu Jun 19 11:46:52 PDT 2025


================
@@ -2172,6 +2172,7 @@ defm V_LDEXP_F16       : VOP2_Real_gfx10<0x03b>;
 let IsSingle = 1 in {
   defm V_PK_FMAC_F16     : VOP2_Real_e32_gfx10<0x03c>;
 }
+defm V_PK_FMAC_F16     :  VOP2_Real_dpp_gfx10<0x03c>, VOP2_Real_dpp8_gfx10<0x03c>;
----------------
rampitec wrote:

Actually gfx9 does support SDWA for this instruction. Add few tests as your change has enabled it?

For gfx10 I think you need to add a test to the gfx10_unsupported_sdwa.s, because it is unsupported there.

https://github.com/llvm/llvm-project/pull/144782


More information about the llvm-commits mailing list