[llvm] AMDGPU/GlobalISel: RegBankLegalize rules for mfma_scale (PR #203279)

[via llvm-commits]

llvmorg-github-actions[bot] wrote:


<!--LLVM PR SUMMARY COMMENT-->

@llvm/pr-subscribers-backend-amdgpu

Author: Petar Avramovic (petar-avramovic)

<details>
<summary>Changes</summary>



---
Full diff: https://github.com/llvm/llvm-project/pull/203279.diff


3 Files Affected:

- (modified) llvm/lib/Target/AMDGPU/AMDGPURegBankLegalizeRules.cpp (+7) 
- (modified) llvm/test/CodeGen/AMDGPU/llvm.amdgcn.mfma.scale.f32.16x16x128.f8f6f4.ll (+1-1) 
- (modified) llvm/test/CodeGen/AMDGPU/llvm.amdgcn.mfma.scale.f32.32x32x64.f8f6f4.ll (+1-1) 


``````````diff
diff --git a/llvm/lib/Target/AMDGPU/AMDGPURegBankLegalizeRules.cpp b/llvm/lib/Target/AMDGPU/AMDGPURegBankLegalizeRules.cpp
index 838951bcd8d26..7e9724670f1d3 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPURegBankLegalizeRules.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPURegBankLegalizeRules.cpp
@@ -2532,6 +2532,13 @@ RegBankLegalizeRules::RegBankLegalizeRules(const GCNSubtarget &_ST,
             {{VgprOrAgprAnyTy},
              {IntrId, VgprAnyTy, VgprAnyTy, VgprOrAgprAnyTy, VgprAnyTy}}});
 
+  addRulesForIOpcs({amdgcn_mfma_scale_f32_32x32x64_f8f6f4,
+                    amdgcn_mfma_scale_f32_16x16x128_f8f6f4})
+      .Any({{DivAnyTy},
+            {{VgprOrAgprAnyTy},
+             {IntrId, VgprAnyTy, VgprAnyTy, VgprAnyTy, Imm, Imm, Imm, Vgpr32,
+              Imm, Vgpr32}}});
+
   // WMMA/SWMMAC intrinsics: all register operands map to VGPR.
   addRulesForIOpcs(
       {// WMMA GFX11+
diff --git a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.mfma.scale.f32.16x16x128.f8f6f4.ll b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.mfma.scale.f32.16x16x128.f8f6f4.ll
index 20f6f5021d5ef..66469851b9e21 100644
--- a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.mfma.scale.f32.16x16x128.f8f6f4.ll
+++ b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.mfma.scale.f32.16x16x128.f8f6f4.ll
@@ -1,6 +1,6 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
 ; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx950 -global-isel=0 < %s | FileCheck -check-prefixes=GCN,SDAG %s
-; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx950 -global-isel=1 < %s | FileCheck -check-prefixes=GCN,GISEL %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx950 -global-isel=1 -new-reg-bank-select < %s | FileCheck -check-prefixes=GCN,GISEL %s
 
 ; 0 = fp8
 ; 1 = bf8
diff --git a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.mfma.scale.f32.32x32x64.f8f6f4.ll b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.mfma.scale.f32.32x32x64.f8f6f4.ll
index f9332d752482d..41d5adefd1fab 100644
--- a/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.mfma.scale.f32.32x32x64.f8f6f4.ll
+++ b/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.mfma.scale.f32.32x32x64.f8f6f4.ll
@@ -1,6 +1,6 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
 ; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx950 -global-isel=0 < %s | FileCheck -check-prefixes=GCN,SDAG %s
-; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx950 -global-isel=1 < %s | FileCheck -check-prefixes=GCN,GISEL %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx950 -global-isel=1 -new-reg-bank-select < %s | FileCheck -check-prefixes=GCN,GISEL %s
 
 ; 0 = fp8
 ; 1 = bf8

``````````

</details>


https://github.com/llvm/llvm-project/pull/203279