Skip to content

[Legalizer] Cache extracted element when lowering G_SHUFFLE_VECTOR. #163893

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Open
wants to merge 1 commit into
base: main
Choose a base branch
from
Open

[Legalizer] Cache extracted element when lowering G_SHUFFLE_VECTOR. #163893

wants to merge 1 commit into from
+57 −54

Conversation

yunqingy180
Copy link

Cache extracted elements in lowerShuffleVector(). For example, when lowering

%0:_(<2 x s32>) = G_BUILD_VECTOR %0, %1
%2:_(<4 x s32>) = G_SHUFFLE_VECTOR %1, shufflemask(0, 0, 0, 0)

Only generate one G_EXTRACT_VECTOR_ELT from %0 and reuse it for all four result elements.

@github-actions GitHub Actions
Copy link

Thank you for submitting a Pull Request (PR) to the LLVM Project!

This PR will be automatically labeled and the relevant teams will be notified.

If you wish to, you can add reviewers by using the "Reviewers" section on this page.

If this is not working for you, it is probably because you do not have write permissions for the repository. In which case you can instead tag reviewers by name in a comment by using @ followed by their GitHub username.

If you have received no comments on your PR for a week, you can request a review by "ping"ing the PR by adding a comment “Ping”. The common courtesy "ping" rate is once a week. Please remember that you are asking for valuable time from other developers.

If you have further questions, they may be answered by the LLVM GitHub User Guide.

You can also ask questions in a comment on this PR, on the LLVM Discord or on the forums.

@llvmbot
Copy link
Member

llvmbot commented Oct 16, 2025 •
edited
Loading

@llvm/pr-subscribers-llvm-globalisel

@llvm/pr-subscribers-backend-amdgpu

Author: Yunqing Yu (yunqingy180)

Changes

Cache extracted elements in lowerShuffleVector(). For example, when lowering

%0:_(&lt;2 x s32&gt;) = G_BUILD_VECTOR %0, %1
%2:_(&lt;4 x s32&gt;) = G_SHUFFLE_VECTOR %1, shufflemask(0, 0, 0, 0)

Only generate one G_EXTRACT_VECTOR_ELT from %0 and reuse it for all four result elements.

Full diff: https://github.com/llvm/llvm-project/pull/163893.diff

3 Files Affected:

  • (modified) llvm/lib/CodeGen/GlobalISel/LegalizerHelper.cpp (+9-3)
  • (modified) llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-shuffle-vector.mir (+28-3)
  • (modified) llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-shuffle-vector.s16.mir (+20-48)
diff --git a/llvm/lib/CodeGen/GlobalISel/LegalizerHelper.cpp b/llvm/lib/CodeGen/GlobalISel/LegalizerHelper.cpp
index 38ec83f81f471..16dea7e54b4b9 100644
--- a/llvm/lib/CodeGen/GlobalISel/LegalizerHelper.cpp
+++ b/llvm/lib/CodeGen/GlobalISel/LegalizerHelper.cpp
@@ -9008,6 +9008,8 @@ LegalizerHelper::lowerShuffleVector(MachineInstr &MI) {
   SmallVector<Register, 32> BuildVec;
   LLT EltTy = DstTy.getScalarType();
 
+  DenseMap<unsigned, Register> CachedExtract;
+
   for (int Idx : Mask) {
     if (Idx < 0) {
       if (!Undef.isValid())
@@ -9022,9 +9024,13 @@ LegalizerHelper::lowerShuffleVector(MachineInstr &MI) {
       int NumElts = Src0Ty.getNumElements();
       Register SrcVec = Idx < NumElts ? Src0Reg : Src1Reg;
       int ExtractIdx = Idx < NumElts ? Idx : Idx - NumElts;
-      auto IdxK = MIRBuilder.buildConstant(IdxTy, ExtractIdx);
-      auto Extract = MIRBuilder.buildExtractVectorElement(EltTy, SrcVec, IdxK);
-      BuildVec.push_back(Extract.getReg(0));
+      auto [It, Inserted] = CachedExtract.try_emplace(Idx);
+      if (Inserted) {
+        auto IdxK = MIRBuilder.buildConstant(IdxTy, ExtractIdx);
+        It->second =
+            MIRBuilder.buildExtractVectorElement(EltTy, SrcVec, IdxK).getReg(0);
+      }
+      BuildVec.push_back(It->second);
     }
   }
 
diff --git a/llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-shuffle-vector.mir b/llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-shuffle-vector.mir
index c8bd8ab33f18c..8e8e78e484b85 100644
--- a/llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-shuffle-vector.mir
+++ b/llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-shuffle-vector.mir
@@ -136,9 +136,7 @@ body: |
     ; CHECK-NEXT: [[COPY1:%[0-9]+]]:_(<2 x s32>) = COPY $vgpr2_vgpr3
     ; CHECK-NEXT: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](<2 x s32>)
     ; CHECK-NEXT: [[COPY2:%[0-9]+]]:_(s32) = COPY [[UV]](s32)
-    ; CHECK-NEXT: [[UV2:%[0-9]+]]:_(s32), [[UV3:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY]](<2 x s32>)
-    ; CHECK-NEXT: [[COPY3:%[0-9]+]]:_(s32) = COPY [[UV2]](s32)
-    ; CHECK-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY2]](s32), [[COPY3]](s32)
+    ; CHECK-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s32>) = G_BUILD_VECTOR [[COPY2]](s32), [[COPY2]](s32)
     ; CHECK-NEXT: $vgpr0_vgpr1 = COPY [[BUILD_VECTOR]](<2 x s32>)
     %0:_(<2 x s32>) = COPY $vgpr0_vgpr1
     %1:_(<2 x s32>) = COPY $vgpr2_vgpr3
@@ -374,3 +372,30 @@ body: |
     $vgpr0_vgpr1 = COPY %4
 
 ...
+
+---
+name: shufflevector_cache_element
+tracksRegLiveness: true
+
+body: |
+  bb.0:
+    liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
+
+    ; CHECK-LABEL: name: shufflevector_cache_element
+    ; CHECK: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
+    ; CHECK-NEXT: {{  $}}
+    ; CHECK-NEXT: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
+    ; CHECK-NEXT: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
+    ; CHECK-NEXT: [[COPY2:%[0-9]+]]:_(s32) = COPY [[COPY]](s32)
+    ; CHECK-NEXT: [[COPY3:%[0-9]+]]:_(s32) = COPY [[COPY1]](s32)
+    ; CHECK-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<8 x s32>) = G_BUILD_VECTOR [[COPY2]](s32), [[COPY2]](s32), [[COPY2]](s32), [[COPY2]](s32), [[COPY3]](s32), [[COPY3]](s32), [[COPY3]](s32), [[COPY3]](s32)
+    ; CHECK-NEXT: [[EVEC:%[0-9]+]]:_(s32) = G_EXTRACT_VECTOR_ELT [[BUILD_VECTOR]](<8 x s32>), [[COPY1]](s32)
+    ; CHECK-NEXT: $vgpr0 = COPY [[EVEC]](s32)
+    %0:_(s32) = COPY $vgpr0
+    %1:_(s32) = COPY $vgpr1
+    %2:_(<2 x s32>) = G_BUILD_VECTOR %0, %1
+    %3:_(<8 x s32>) = G_SHUFFLE_VECTOR %2, %2, shufflemask(0, 0, 0, 0, 3, 3, 3, 3)
+    %4:_(s32) = G_EXTRACT_VECTOR_ELT %3(<8 x s32>), %1(s32)
+    $vgpr0 = COPY %4
+
+...
diff --git a/llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-shuffle-vector.s16.mir b/llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-shuffle-vector.s16.mir
index 10f425c96e6f4..dff8111de72d9 100644
--- a/llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-shuffle-vector.s16.mir
+++ b/llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-shuffle-vector.s16.mir
@@ -237,16 +237,13 @@ body: |
     ; GFX8-NEXT: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
     ; GFX8-NEXT: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>)
     ; GFX8-NEXT: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
-    ; GFX8-NEXT: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>)
-    ; GFX8-NEXT: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
     ; GFX8-NEXT: [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 65535
     ; GFX8-NEXT: [[AND:%[0-9]+]]:_(s32) = G_AND [[LSHR]], [[C1]]
-    ; GFX8-NEXT: [[AND1:%[0-9]+]]:_(s32) = G_AND [[LSHR1]], [[C1]]
     ; GFX8-NEXT: [[C2:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
-    ; GFX8-NEXT: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[AND1]], [[C2]](s32)
+    ; GFX8-NEXT: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[AND]], [[C2]](s32)
     ; GFX8-NEXT: [[OR:%[0-9]+]]:_(s32) = G_OR [[AND]], [[SHL]]
-    ; GFX8-NEXT: [[BITCAST2:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32)
-    ; GFX8-NEXT: $vgpr0 = COPY [[BITCAST2]](<2 x s16>)
+    ; GFX8-NEXT: [[BITCAST1:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32)
+    ; GFX8-NEXT: $vgpr0 = COPY [[BITCAST1]](<2 x s16>)
     ;
     ; GFX9-LABEL: name: shufflevector_v2s16_v2s16_0_0
     ; GFX9: liveins: $vgpr0, $vgpr1
@@ -257,10 +254,7 @@ body: |
     ; GFX9-NEXT: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>)
     ; GFX9-NEXT: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
     ; GFX9-NEXT: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32)
-    ; GFX9-NEXT: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>)
-    ; GFX9-NEXT: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
-    ; GFX9-NEXT: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32)
-    ; GFX9-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s16>) = G_BUILD_VECTOR [[TRUNC]](s16), [[TRUNC1]](s16)
+    ; GFX9-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s16>) = G_BUILD_VECTOR [[TRUNC]](s16), [[TRUNC]](s16)
     ; GFX9-NEXT: $vgpr0 = COPY [[BUILD_VECTOR]](<2 x s16>)
     %0:_(<2 x s16>) = COPY $vgpr0
     %1:_(<2 x s16>) = COPY $vgpr1
@@ -285,12 +279,10 @@ body: |
     ; GFX8-NEXT: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>)
     ; GFX8-NEXT: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
     ; GFX8-NEXT: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
-    ; GFX8-NEXT: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>)
-    ; GFX8-NEXT: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
-    ; GFX8-NEXT: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[LSHR1]], [[C]](s32)
+    ; GFX8-NEXT: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[LSHR]], [[C]](s32)
     ; GFX8-NEXT: [[OR:%[0-9]+]]:_(s32) = G_OR [[LSHR]], [[SHL]]
-    ; GFX8-NEXT: [[BITCAST2:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32)
-    ; GFX8-NEXT: $vgpr0 = COPY [[BITCAST2]](<2 x s16>)
+    ; GFX8-NEXT: [[BITCAST1:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32)
+    ; GFX8-NEXT: $vgpr0 = COPY [[BITCAST1]](<2 x s16>)
     ;
     ; GFX9-LABEL: name: shufflevector_v2s16_v2s16_1_1
     ; GFX9: liveins: $vgpr0, $vgpr1
@@ -301,10 +293,7 @@ body: |
     ; GFX9-NEXT: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
     ; GFX9-NEXT: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
     ; GFX9-NEXT: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32)
-    ; GFX9-NEXT: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>)
-    ; GFX9-NEXT: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
-    ; GFX9-NEXT: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32)
-    ; GFX9-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s16>) = G_BUILD_VECTOR [[TRUNC]](s16), [[TRUNC1]](s16)
+    ; GFX9-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s16>) = G_BUILD_VECTOR [[TRUNC]](s16), [[TRUNC]](s16)
     ; GFX9-NEXT: $vgpr0 = COPY [[BUILD_VECTOR]](<2 x s16>)
     %0:_(<2 x s16>) = COPY $vgpr0
     %1:_(<2 x s16>) = COPY $vgpr1
@@ -329,16 +318,13 @@ body: |
     ; GFX8-NEXT: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
     ; GFX8-NEXT: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[COPY1]](<2 x s16>)
     ; GFX8-NEXT: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
-    ; GFX8-NEXT: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[COPY1]](<2 x s16>)
-    ; GFX8-NEXT: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
     ; GFX8-NEXT: [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 65535
     ; GFX8-NEXT: [[AND:%[0-9]+]]:_(s32) = G_AND [[LSHR]], [[C1]]
-    ; GFX8-NEXT: [[AND1:%[0-9]+]]:_(s32) = G_AND [[LSHR1]], [[C1]]
     ; GFX8-NEXT: [[C2:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
-    ; GFX8-NEXT: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[AND1]], [[C2]](s32)
+    ; GFX8-NEXT: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[AND]], [[C2]](s32)
     ; GFX8-NEXT: [[OR:%[0-9]+]]:_(s32) = G_OR [[AND]], [[SHL]]
-    ; GFX8-NEXT: [[BITCAST2:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32)
-    ; GFX8-NEXT: $vgpr0 = COPY [[BITCAST2]](<2 x s16>)
+    ; GFX8-NEXT: [[BITCAST1:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32)
+    ; GFX8-NEXT: $vgpr0 = COPY [[BITCAST1]](<2 x s16>)
     ;
     ; GFX9-LABEL: name: shufflevector_v2s16_v2s16_2_2
     ; GFX9: liveins: $vgpr0, $vgpr1
@@ -349,10 +335,7 @@ body: |
     ; GFX9-NEXT: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[COPY1]](<2 x s16>)
     ; GFX9-NEXT: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
     ; GFX9-NEXT: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32)
-    ; GFX9-NEXT: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[COPY1]](<2 x s16>)
-    ; GFX9-NEXT: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
-    ; GFX9-NEXT: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32)
-    ; GFX9-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s16>) = G_BUILD_VECTOR [[TRUNC]](s16), [[TRUNC1]](s16)
+    ; GFX9-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s16>) = G_BUILD_VECTOR [[TRUNC]](s16), [[TRUNC]](s16)
     ; GFX9-NEXT: $vgpr0 = COPY [[BUILD_VECTOR]](<2 x s16>)
     %0:_(<2 x s16>) = COPY $vgpr0
     %1:_(<2 x s16>) = COPY $vgpr1
@@ -641,12 +624,10 @@ body: |
     ; GFX8-NEXT: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[COPY1]](<2 x s16>)
     ; GFX8-NEXT: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
     ; GFX8-NEXT: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
-    ; GFX8-NEXT: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[COPY1]](<2 x s16>)
-    ; GFX8-NEXT: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
-    ; GFX8-NEXT: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[LSHR1]], [[C]](s32)
+    ; GFX8-NEXT: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[LSHR]], [[C]](s32)
     ; GFX8-NEXT: [[OR:%[0-9]+]]:_(s32) = G_OR [[LSHR]], [[SHL]]
-    ; GFX8-NEXT: [[BITCAST2:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32)
-    ; GFX8-NEXT: $vgpr0 = COPY [[BITCAST2]](<2 x s16>)
+    ; GFX8-NEXT: [[BITCAST1:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32)
+    ; GFX8-NEXT: $vgpr0 = COPY [[BITCAST1]](<2 x s16>)
     ;
     ; GFX9-LABEL: name: shufflevector_v2s16_v2s16_3_3
     ; GFX9: liveins: $vgpr0, $vgpr1
@@ -657,10 +638,7 @@ body: |
     ; GFX9-NEXT: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
     ; GFX9-NEXT: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
     ; GFX9-NEXT: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32)
-    ; GFX9-NEXT: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[COPY1]](<2 x s16>)
-    ; GFX9-NEXT: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
-    ; GFX9-NEXT: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32)
-    ; GFX9-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s16>) = G_BUILD_VECTOR [[TRUNC]](s16), [[TRUNC1]](s16)
+    ; GFX9-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s16>) = G_BUILD_VECTOR [[TRUNC]](s16), [[TRUNC]](s16)
     ; GFX9-NEXT: $vgpr0 = COPY [[BUILD_VECTOR]](<2 x s16>)
     %0:_(<2 x s16>) = COPY $vgpr0
     %1:_(<2 x s16>) = COPY $vgpr1
@@ -733,16 +711,13 @@ body: |
     ; GFX8-NEXT: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 0
     ; GFX8-NEXT: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>)
     ; GFX8-NEXT: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
-    ; GFX8-NEXT: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>)
-    ; GFX8-NEXT: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
     ; GFX8-NEXT: [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 65535
     ; GFX8-NEXT: [[AND:%[0-9]+]]:_(s32) = G_AND [[LSHR]], [[C1]]
-    ; GFX8-NEXT: [[AND1:%[0-9]+]]:_(s32) = G_AND [[LSHR1]], [[C1]]
     ; GFX8-NEXT: [[C2:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
-    ; GFX8-NEXT: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[AND1]], [[C2]](s32)
+    ; GFX8-NEXT: [[SHL:%[0-9]+]]:_(s32) = G_SHL [[AND]], [[C2]](s32)
     ; GFX8-NEXT: [[OR:%[0-9]+]]:_(s32) = G_OR [[AND]], [[SHL]]
-    ; GFX8-NEXT: [[BITCAST2:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32)
-    ; GFX8-NEXT: $vgpr0 = COPY [[BITCAST2]](<2 x s16>)
+    ; GFX8-NEXT: [[BITCAST1:%[0-9]+]]:_(<2 x s16>) = G_BITCAST [[OR]](s32)
+    ; GFX8-NEXT: $vgpr0 = COPY [[BITCAST1]](<2 x s16>)
     ;
     ; GFX9-LABEL: name: shufflevector_v2s16_v2s16_2_0
     ; GFX9: liveins: $vgpr0, $vgpr1
@@ -753,10 +728,7 @@ body: |
     ; GFX9-NEXT: [[BITCAST:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>)
     ; GFX9-NEXT: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST]], [[C]](s32)
     ; GFX9-NEXT: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32)
-    ; GFX9-NEXT: [[BITCAST1:%[0-9]+]]:_(s32) = G_BITCAST [[COPY]](<2 x s16>)
-    ; GFX9-NEXT: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[BITCAST1]], [[C]](s32)
-    ; GFX9-NEXT: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR1]](s32)
-    ; GFX9-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s16>) = G_BUILD_VECTOR [[TRUNC]](s16), [[TRUNC1]](s16)
+    ; GFX9-NEXT: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s16>) = G_BUILD_VECTOR [[TRUNC]](s16), [[TRUNC]](s16)
     ; GFX9-NEXT: $vgpr0 = COPY [[BUILD_VECTOR]](<2 x s16>)
     %0:_(<2 x s16>) = COPY $vgpr0
     %1:_(<2 x s16>) = COPY $vgpr1

@yunqingy180
Copy link
Author

Hi @topperc could you please take a look at this Legalizer change? Thanks a lot!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

backend:AMDGPU llvm:globalisel

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

2 participants

@yunqingy180 @llvmbot