[SM6.10] Add -fcgl variant to LinAlg Matrix codegen tests

hekota · hekota · commit 2ab314620db5 · 2026-03-13T12:52:12.000-07:00
diff --git a/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/copyconvertmatrix/nominal.hlsl b/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/copyconvertmatrix/nominal.hlsl
@@ -1,11 +1,17 @@
 // REQUIRES: dxil-1-10
 // RUN: %dxc -T cs_6_10 -E main %s | FileCheck %s
+// RUN: %dxc -T cs_6_10 -E main -fcgl %s | FileCheck %s --check-prefix=CHECK2
 
 [numthreads(1,1,1)]
 void main() {
   // CHECK-LABEL: define void @main()
 
-  // CHECK: call %dx.types.LinAlgMatrixC4M5N4U1S2 @dx.op.linAlgCopyConvertMatrix.mC4M5N4U1S2.mC2M5N4U1S2(i32 -2147483635, %dx.types.LinAlgMatrixC2M5N4U1S2 {{.*}}, i1 false)  ; LinAlgCopyConvertMatrix(srcMatrix,transpose)
+  // CHECK: call %dx.types.LinAlgMatrixC4M5N4U1S2 @dx.op.linAlgCopyConvertMatrix.mC4M5N4U1S2.mC2M5N4U1S2
+  // CHECK-SAME: (i32 -2147483635, %dx.types.LinAlgMatrixC2M5N4U1S2 {{.*}}, i1 false)  
+  // CHECK-SAME: ; LinAlgCopyConvertMatrix(srcMatrix,transpose)
+
+  // CHECK2: call void @"dx.hl.op..void (i32, %dx.types.LinAlgMatrixC4M5N4U1S2*, %dx.types.LinAlgMatrixC2M5N4U1S2, i1)"
+  // CHECK2-SAME: (i32 405, %dx.types.LinAlgMatrixC4M5N4U1S2* {{.*}}, %dx.types.LinAlgMatrixC2M5N4U1S2 {{.*}}, i1 false)
   __builtin_LinAlgMatrix [[__LinAlgMatrix_Attributes(2, 5, 4, 1, 2)]] mat1;
   __builtin_LinAlgMatrix [[__LinAlgMatrix_Attributes(4, 5, 4, 1, 2)]] mat2;
   __builtin_LinAlg_CopyConvertMatrix(mat2, mat1, false);
diff --git a/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/fillmatrix/nominal.hlsl b/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/fillmatrix/nominal.hlsl
@@ -1,14 +1,18 @@
 // REQUIRES: dxil-1-10
 // RUN: %dxc -T cs_6_10 -E main %s | FileCheck %s
+// RUN: %dxc -T cs_6_10 -E main -fcgl %s | FileCheck %s --check-prefix=CHECK2
 
 [numthreads(1,1,1)]
 void main() {
   // CHECK-LABEL: define void @main()
 
   // CHECK: %{{.*}} = call %dx.types.LinAlgMatrixC4M5N4U1S2 @dx.op.linAlgFillMatrix.mC4M5N4U1S2.i32(i32 -2147483636, i32 {{.*}})  ; LinAlgFillMatrix(value)
+  // CHECK2: call void @"dx.hl.op..void (i32, %dx.types.LinAlgMatrixC4M5N4U1S2*, i32)"(i32 406, %dx.types.LinAlgMatrixC4M5N4U1S2* {{.*}}, i32 5),
   __builtin_LinAlgMatrix [[__LinAlgMatrix_Attributes(4, 5, 4, 1, 2)]] mat1;
   __builtin_LinAlg_FillMatrix(mat1, 5);
+  
   // CHECK: %{{.*}} = call %dx.types.LinAlgMatrixC5M3N4U0S0 @dx.op.linAlgFillMatrix.mC5M3N4U0S0.f32(i32 -2147483636, float {{.*}})  ; LinAlgFillMatrix(value)
+  // CHECK2: call void @"dx.hl.op..void (i32, %dx.types.LinAlgMatrixC5M3N4U0S0*, float)"(i32 406, %dx.types.LinAlgMatrixC5M3N4U0S0* {{.*}}, float 0x40091EB860000000)
   __builtin_LinAlgMatrix [[__LinAlgMatrix_Attributes(5, 3, 4, 0, 0)]] mat2;
   __builtin_LinAlg_FillMatrix(mat2, 3.14);
 }
diff --git a/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/matrixaccumulate/nominal.hlsl b/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/matrixaccumulate/nominal.hlsl
@@ -1,5 +1,6 @@
 // REQUIRES: dxil-1-10
 // RUN: %dxc -T cs_6_10 -E main %s | FileCheck %s
+// RUN: %dxc -T cs_6_10 -E main -fcgl %s | FileCheck %s --check-prefix=CHECK2
 
 [numthreads(1,1,1)]
 void main() {
@@ -8,7 +9,12 @@ void main() {
   __builtin_LinAlgMatrix [[__LinAlgMatrix_Attributes(5, 3, 4, 0, 0)]] mat1;
   __builtin_LinAlgMatrix [[__LinAlgMatrix_Attributes(1, 1, 1, 0, 0)]] mat2;
   __builtin_LinAlgMatrix [[__LinAlgMatrix_Attributes(2, 2, 2, 2, 2)]] mat3;
+  
+  // CHECK: call %dx.types.LinAlgMatrixC2M2N2U2S2 @dx.op.linAlgMatrixAccumulate.mC2M2N2U2S2.mC1M1N1U0S0.mC5M3N4U0S
+  // CHECK-SAME: (i32 -2147483624, %dx.types.LinAlgMatrixC1M1N1U0S0 {{.*}}, %dx.types.LinAlgMatrixC5M3N4U0S0 {{.*}}) ; LinAlgMatrixAccumulate(matrixLHS,matrixRHS)
 
-  // CHECK: call %dx.types.LinAlgMatrixC2M2N2U2S2 @dx.op.linAlgMatrixAccumulate.mC2M2N2U2S2.mC1M1N1U0S0.mC5M3N4U0S0(i32 -2147483624, %dx.types.LinAlgMatrixC1M1N1U0S0 {{.*}}, %dx.types.LinAlgMatrixC5M3N4U0S0 {{.*}})  ; LinAlgMatrixAccumulate(matrixLHS,matrixRHS)
+  // CHECK2: call void @"dx.hl.op..void (i32, %dx.types.LinAlgMatrixC2M2N2U2S2*, %dx.types.LinAlgMatrixC1M1N1U0S0,
+  // CHECK2" %dx.types.LinAlgMatrixC5M3N4U0S0)"(i32 415, %dx.types.LinAlgMatrixC2M2N2U2S2* {{.*}},
+  // CHECK2" %dx.types.LinAlgMatrixC1M1N1U0S0 {{.*}}, %dx.types.LinAlgMatrixC5M3N4U0S0 {{.*}})
   __builtin_LinAlg_MatrixAccumulate(mat3, mat2, mat1);
 }
diff --git a/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/matrixaccumulatetodescriptor/nominal.hlsl b/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/matrixaccumulatetodescriptor/nominal.hlsl
@@ -1,13 +1,19 @@
 // REQUIRES: dxil-1-10
 // RUN: %dxc -T cs_6_10 -E main %s | FileCheck %s
+// RUN: %dxc -T cs_6_10 -E main -fcgl %s | FileCheck %s --check-prefix=CHECK2
 
 RWByteAddressBuffer outbuf;
 
 [numthreads(1,1,1)]
 void main() {
   // CHECK-LABEL: define void @main()
 
-  // CHECK: call void @dx.op.linAlgMatrixAccumulateToDescriptor.mC4M5N4U1S2(i32 -2147483621, %dx.types.LinAlgMatrixC4M5N4U1S2 {{.*}}, %dx.types.Handle %{{.*}}, i32 5, i32 5, i32 5)  ; LinAlgMatrixAccumulateToDescriptor(matrix,handle,offset,stride,layout)
+  // CHECK: call void @dx.op.linAlgMatrixAccumulateToDescriptor.mC4M5N4U1S2(i32 -2147483621,
+  // CHECK-SAME: %dx.types.LinAlgMatrixC4M5N4U1S2 {{.*}}, %dx.types.Handle %{{.*}}, i32 5, i32 5, i32 5)
+  // CHECK-SAME: ; LinAlgMatrixAccumulateToDescriptor(matrix,handle,offset,stride,layout)
+
+  // CHECK2: call void @"dx.hl.op..void (i32, %dx.types.LinAlgMatrixC4M5N4U1S2, %dx.types.Handle, i32, i32, i32)"
+  // CHECK2-SAME: (i32 419, %dx.types.LinAlgMatrixC4M5N4U1S2 %2, %dx.types.Handle {{.*}}, i32 5, i32 5, i32 5)
   __builtin_LinAlgMatrix [[__LinAlgMatrix_Attributes(4, 5, 4, 1, 2)]] mat;
   __builtin_LinAlg_MatrixAccumulateToDescriptor(mat, outbuf, 5, 5, 5);
 }
diff --git a/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/matrixgetcoordinate/nominal.hlsl b/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/matrixgetcoordinate/nominal.hlsl
@@ -1,11 +1,17 @@
 // REQUIRES: dxil-1-10
 // RUN: %dxc -T cs_6_10 -E main %s | FileCheck %s
+// RUN: %dxc -T cs_6_10 -E main -fcgl %s | FileCheck %s --check-prefix=CHECK2
 
 [numthreads(1,1,1)]
 void main() {
   // CHECK-LABEL: define void @main()
 
-  // CHECK: call <2 x i32> @dx.op.linAlgMatrixGetCoordinate.mC4M5N4U1S2(i32 -2147483631, %dx.types.LinAlgMatrixC4M5N4U1S2 {{.*}}, i32 1) ; LinAlgMatrixGetCoordinate(matrix,threadLocalIndex)
+  // CHECK: call <2 x i32> @dx.op.linAlgMatrixGetCoordinate.mC4M5N4U1S2(i32 -2147483631,
+  // CHECK-SAME: %dx.types.LinAlgMatrixC4M5N4U1S2 {{.*}}, i32 1) 
+  // CHECK-SAME: ; LinAlgMatrixGetCoordinate(matrix,threadLocalIndex)
+
+  // CHECK2: call <2 x i32> @"dx.hl.op..<2 x i32> (i32, %dx.types.LinAlgMatrixC4M5N4U1S2, i32)"
+  // CHECK2-SAME: (i32 407, %dx.types.LinAlgMatrixC4M5N4U1S2 {{.*}}, i32 1)
   __builtin_LinAlgMatrix [[__LinAlgMatrix_Attributes(4, 5, 4, 1, 2)]] mat;
   uint2 coord = __builtin_LinAlg_MatrixGetCoordinate(mat, 1);
 }
diff --git a/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/matrixgetelement/nominal.hlsl b/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/matrixgetelement/nominal.hlsl
@@ -1,16 +1,28 @@
 // REQUIRES: dxil-1-10
 // RUN: %dxc -T cs_6_10 -E main %s | FileCheck %s
+// RUN: %dxc -T cs_6_10 -E main -fcgl %s | FileCheck %s --check-prefix=CHECK2
 
 [numthreads(1,1,1)]
 void main() {
   // CHECK-LABEL: define void @main()
 
   __builtin_LinAlgMatrix [[__LinAlgMatrix_Attributes(4, 5, 4, 1, 2)]] mat;
 
-  // CHECK: call i32 @dx.op.linAlgMatrixGetElement.i32.mC4M5N4U1S2(i32 -2147483630, %dx.types.LinAlgMatrixC4M5N4U1S2 {{.*}}, i32 0)  ; LinAlgMatrixGetElement(matrix,threadLocalIndex)
+  // CHECK: call i32 @dx.op.linAlgMatrixGetElement.i32.mC4M5N4U1S2(i32 -2147483630,
+  // CHECK-SAME: %dx.types.LinAlgMatrixC4M5N4U1S2 {{.*}}, i32 0)  
+  // CHECK-SAME: ; LinAlgMatrixGetElement(matrix,threadLocalIndex)
+
+  // CHECK2: call void @"dx.hl.op..void (i32, i32*, %dx.types.LinAlgMatrixC4M5N4U1S2, i32)"
+  // CHECK2-SAME: (i32 408, i32* %elem1, %dx.types.LinAlgMatrixC4M5N4U1S2 %3, i32 0)
   uint elem1;
   __builtin_LinAlg_MatrixGetElement(elem1, mat, 0);
-  // CHECK: call float @dx.op.linAlgMatrixGetElement.f32.mC4M5N4U1S2(i32 -2147483630, %dx.types.LinAlgMatrixC4M5N4U1S2 {{.*}}, i32 1)  ; LinAlgMatrixGetElement(matrix,threadLocalIndex)
+
+  // CHECK: call float @dx.op.linAlgMatrixGetElement.f32.mC4M5N4U1S2(i32 -2147483630,
+  // CHECK-SAME: %dx.types.LinAlgMatrixC4M5N4U1S2 {{.*}}, i32 1)  
+  // CHECK-SAME: ; LinAlgMatrixGetElement(matrix,threadLocalIndex)
+
+  // CHECK2: call void @"dx.hl.op..void (i32, float*, %dx.types.LinAlgMatrixC4M5N4U1S2, i32)"
+  // CHECK2-SAME: (i32 408, float* %elem2, %dx.types.LinAlgMatrixC4M5N4U1S2 {{.*}}, i32 1)
   float elem2;
   __builtin_LinAlg_MatrixGetElement(elem2, mat, 1);
 }
diff --git a/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/matrixlength/nominal.hlsl b/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/matrixlength/nominal.hlsl
@@ -1,11 +1,16 @@
 // REQUIRES: dxil-1-10
 // RUN: %dxc -T cs_6_10 -E main %s | FileCheck %s
+// RUN: %dxc -T cs_6_10 -E main -fcgl %s | FileCheck %s --check-prefix=CHECK2
 
 [numthreads(1,1,1)]
 void main() {
   // CHECK-LABEL: define void @main()
 
-  // CHECK: call i32 @dx.op.linAlgMatrixLength.mC4M5N4U1S2(i32 -2147483632, %dx.types.LinAlgMatrixC4M5N4U1S2 {{.*}})  ; LinAlgMatrixLength(matrix)
+  // CHECK: call i32 @dx.op.linAlgMatrixLength.mC4M5N4U1S2(i32 -2147483632,
+  // CHECK-SAME: %dx.types.LinAlgMatrixC4M5N4U1S2 {{.*}})  ; LinAlgMatrixLength(matrix)
+
+  // CHECK2: call i32 @"dx.hl.op..i32 (i32, %dx.types.LinAlgMatrixC4M5N4U1S2)"
+  // CHECK2-SAME: (i32 409, %dx.types.LinAlgMatrixC4M5N4U1S2 {{.*}})
   __builtin_LinAlgMatrix [[__LinAlgMatrix_Attributes(4, 5, 4, 1, 2)]] mat;
   uint len = __builtin_LinAlg_MatrixLength(mat);
 }
diff --git a/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/matrixloadfromdescriptor/nominal.hlsl b/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/matrixloadfromdescriptor/nominal.hlsl
@@ -1,13 +1,19 @@
 // REQUIRES: dxil-1-10
 // RUN: %dxc -T cs_6_10 -E main %s | FileCheck %s
+// RUN: %dxc -T cs_6_10 -E main -fcgl %s | FileCheck %s --check-prefix=CHECK2
 
 ByteAddressBuffer inbuf;
 
 [numthreads(1,1,1)]
 void main() {
   // CHECK-LABEL: define void @main()
 
-  // CHECK: %{{.*}} = call %dx.types.LinAlgMatrixC1M1N1U0S0 @dx.op.linAlgMatrixLoadFromDescriptor.mC1M1N1U0S0(i32 -2147483634, %dx.types.Handle %{{.*}}, i32 0, i32 0, i32 0)  ; LinAlgMatrixLoadFromDescriptor(handle,offset,stride,layout)
+  // CHECK: %{{.*}} = call %dx.types.LinAlgMatrixC1M1N1U0S0 @dx.op.linAlgMatrixLoadFromDescriptor.mC1M1N1U0S0
+  // CHECK-SAME: (i32 -2147483634, %dx.types.Handle %{{.*}}, i32 0, i32 0, i32 0) 
+  // CHECK-SAME: ; LinAlgMatrixLoadFromDescriptor(handle,offset,stride,layout)
+
+  // CHECK2: call void @"dx.hl.op..void (i32, %dx.types.LinAlgMatrixC1M1N1U0S0*, %dx.types.Handle, i32, i32, i32)
+  // CHECK2-SAME: "(i32 410, %dx.types.LinAlgMatrixC1M1N1U0S0* %mat, %dx.types.Handle {{.*}}, i32 0, i32 0, i32 0)
   __builtin_LinAlgMatrix [[__LinAlgMatrix_Attributes(1, 1, 1, 0, 0)]] mat;
   __builtin_LinAlg_MatrixLoadFromDescriptor(mat, inbuf, 0, 0, 0);
 }
diff --git a/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/matrixmatrixmultiply/nominal.hlsl b/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/matrixmatrixmultiply/nominal.hlsl
@@ -1,11 +1,17 @@
 // REQUIRES: dxil-1-10
 // RUN: %dxc -T cs_6_10 -E main %s | FileCheck %s
+// RUN: %dxc -T cs_6_10 -E main -fcgl %s | FileCheck %s --check-prefix=CHECK2
 
 [numthreads(1,1,1)]
 void main() {
   // CHECK-LABEL: define void @main()
 
-  // CHECK: call %dx.types.LinAlgMatrixC4M5N4U1S2 @dx.op.linAlgMatrixMultiply.mC4M5N4U1S2.mC4M5N4U1S2.mC4M5N4U1S2(i32 -2147483625, %dx.types.LinAlgMatrixC4M5N4U1S2 {{.*}}, %dx.types.LinAlgMatrixC4M5N4U1S2 {{.*}})  ; LinAlgMatrixMultiply(matrixA,matrixB)
+  // CHECK: call %dx.types.LinAlgMatrixC4M5N4U1S2 @dx.op.linAlgMatrixMultiply.mC4M5N4U1S2.mC4M5N4U1S2.mC4M5N4U1S2(i32 -2147483625,
+  // CHECK-SAME: %dx.types.LinAlgMatrixC4M5N4U1S2 {{.*}}, %dx.types.LinAlgMatrixC4M5N4U1S2 {{.*}}) ; LinAlgMatrixMultiply(matrixA,matrixB)
+
+  // CHECK2: call void @"dx.hl.op..void (i32, %dx.types.LinAlgMatrixC4M5N4U1S2*, %dx.types.LinAlgMatrixC4M5N4U1S2,
+  // CHECK2-SAME: %dx.types.LinAlgMatrixC4M5N4U1S2)"(i32 416, %dx.types.LinAlgMatrixC4M5N4U1S2* %mat2,
+  // CHECK2-SAME: %dx.types.LinAlgMatrixC4M5N4U1S2 %{{[0-9]+}}, %dx.types.LinAlgMatrixC4M5N4U1S2 %{{[0-9]+}})
   __builtin_LinAlgMatrix [[__LinAlgMatrix_Attributes(4, 5, 4, 1, 2)]] mat1;
   __builtin_LinAlgMatrix [[__LinAlgMatrix_Attributes(4, 5, 4, 1, 2)]] mat2;
   __builtin_LinAlg_MatrixMatrixMultiply(mat2, mat1, mat1);
diff --git a/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/matrixmatrixmultiplyaccumulate/nominal.hlsl b/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/matrixmatrixmultiplyaccumulate/nominal.hlsl
@@ -1,12 +1,22 @@
 // REQUIRES: dxil-1-10
 // RUN: %dxc -T cs_6_10 -E main %s | FileCheck %s
+// RUN: %dxc -T cs_6_10 -E main -fcgl %s | FileCheck %s --check-prefix=CHECK2
 
 [numthreads(1,1,1)]
 void main() {
   // CHECK-LABEL: define void @main()
 
-  // CHECK: call %dx.types.LinAlgMatrixC4M5N4U1S2 @dx.op.linAlgMatrixMultiplyAccumulate.mC4M5N4U1S2.mC4M5N4U1S2.mC4M5N4U1S2.mC4M5N4U1S2(i32 -2147483637, %dx.types.LinAlgMatrixC4M5N4U1S2 {{.*}}, %dx.types.LinAlgMatrixC4M5N4U1S2 {{.*}}, %dx.types.LinAlgMatrixC4M5N4U1S2 {{.*}})  ; LinAlgMatrixMultiplyAccumulate(matrixA,matrixB,matrixC)
+  // CHECK: call %dx.types.LinAlgMatrixC4M5N3U1S2
+  // CHECK-SAME: @dx.op.linAlgMatrixMultiplyAccumulate.mC4M5N3U1S2.mC4M5N4U1S2.mC4M4N3U1S2.mC4M5N3U1S2
+  // CHECK-SAME: (i32 -2147483637, %dx.types.LinAlgMatrixC4M5N4U1S2 undef, %dx.types.LinAlgMatrixC4M4N3U1S2 undef, 
+  // CHECK-SAME: %dx.types.LinAlgMatrixC4M5N3U1S2 undef) ; LinAlgMatrixMultiplyAccumulate(matrixA,matrixB,matrixC)
+  
+  // CHECK2: call void @"dx.hl.op..void (i32, %dx.types.LinAlgMatrixC4M5N3U1S2*, %dx.types.LinAlgMatrixC4M5N4U1S2,
+  // CHECK2-SAME: %dx.types.LinAlgMatrixC4M4N3U1S2, %dx.types.LinAlgMatrixC4M5N3U1S2)"(i32 417,
+  // CHECK2-SAME: %dx.types.LinAlgMatrixC4M5N3U1S2* {{.*}}, %dx.types.LinAlgMatrixC4M5N4U1S2 %{{[0-9]+}},
+  // CHECK2-SAME: %dx.types.LinAlgMatrixC4M4N3U1S2 %{{[0-9]+}}, %dx.types.LinAlgMatrixC4M5N3U1S2 %{{[0-9]+}})
   __builtin_LinAlgMatrix [[__LinAlgMatrix_Attributes(4, 5, 4, 1, 2)]] mat1;
-  __builtin_LinAlgMatrix [[__LinAlgMatrix_Attributes(4, 5, 4, 1, 2)]] mat2;
-  __builtin_LinAlg_MatrixMatrixMultiplyAccumulate(mat2, mat1, mat1, mat1);
+  __builtin_LinAlgMatrix [[__LinAlgMatrix_Attributes(4, 4, 3, 1, 2)]] mat2;
+  __builtin_LinAlgMatrix [[__LinAlgMatrix_Attributes(4, 5, 3, 1, 2)]] mat3;
+  __builtin_LinAlg_MatrixMatrixMultiplyAccumulate(mat3, mat1, mat2, mat3);
 }
diff --git a/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/matrixouterproduct/nominal.hlsl b/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/matrixouterproduct/nominal.hlsl
@@ -1,13 +1,19 @@
 // REQUIRES: dxil-1-10
 // RUN: %dxc -T cs_6_10 -E main %s | FileCheck %s
+// RUN: %dxc -T cs_6_10 -E main -fcgl %s | FileCheck %s --check-prefix=CHECK2
 
 [numthreads(1,1,1)]
 void main() {
   // CHECK-LABEL: define void @main()
 
   float4 lhs = {1,2,3,4};
   float4 rhs = {4,3,2,1};
-  // CHECK: call %dx.types.LinAlgMatrixC2M2N2U2S2 @dx.op.linAlgMatrixOuterProduct.mC2M2N2U2S2.v4f32.v4f32(i32 -2147483619, <4 x float> {{.*}}, <4 x float> {{.*}})  ; LinAlgMatrixOuterProduct(vectorA,vectorB)
+
+  // CHECK: call %dx.types.LinAlgMatrixC2M2N2U2S2 @dx.op.linAlgMatrixOuterProduct.mC2M2N2U2S2.v4f32.v4f32
+  // CHECK-SAME: (i32 -2147483619, <4 x float> {{.*}}, <4 x float> {{.*}})  ; LinAlgMatrixOuterProduct(vectorA,vectorB)
+
+  // CHECK2: call void @"dx.hl.op..void (i32, %dx.types.LinAlgMatrixC2M2N2U2S2*, <4 x float>, <4 x float>)"
+  // CHECK2: (i32 421, %dx.types.LinAlgMatrixC2M2N2U2S2* {{.*}}, <4 x float> {{.*}}, <4 x float> {{.*}})
   __builtin_LinAlgMatrix [[__LinAlgMatrix_Attributes(2, 2, 2, 2, 2)]] mat;
   __builtin_LinAlg_MatrixOuterProduct(mat, lhs, rhs);
 }
diff --git a/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/matrixqueryaccumulatorlayout/nominal.hlsl b/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/matrixqueryaccumulatorlayout/nominal.hlsl
@@ -1,10 +1,13 @@
 // REQUIRES: dxil-1-10
 // RUN: %dxc -T cs_6_10 -E main %s | FileCheck %s
+// RUN: %dxc -T cs_6_10 -E main -fcgl %s | FileCheck %s --check-prefix=CHECK2
 
 [numthreads(1,1,1)]
 void main() {
   // CHECK-LABEL: define void @main()
 
   // CHECK: call i32 @dx.op.linAlgMatrixQueryAccumulatorLayout(i32 -2147483626)  ; LinAlgMatrixQueryAccumulatorLayout()
+
+  // CHECK2: call i32 @"dx.hl.op..i32 (i32)"(i32 418)
   uint layout = __builtin_LinAlg_MatrixQueryAccumulatorLayout();
 }
diff --git a/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/matrixsetelement/nominal.hlsl b/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/matrixsetelement/nominal.hlsl
@@ -1,5 +1,6 @@
 // REQUIRES: dxil-1-10
 // RUN: %dxc -T cs_6_10 -E main %s | FileCheck %s
+// RUN: %dxc -T cs_6_10 -E main -fcgl %s | FileCheck %s --check-prefix=CHECK2
 
 ByteAddressBuffer inbuf;
 RWByteAddressBuffer outbuf;
@@ -8,7 +9,13 @@ RWByteAddressBuffer outbuf;
 void main() {
   // CHECK-LABEL: define void @main()
 
-  // CHECK: call %dx.types.LinAlgMatrixC4M5N4U1S2 @dx.op.linAlgMatrixSetElement.mC4M5N4U1S2.mC4M5N4U1S2.i32(i32 -2147483629, %dx.types.LinAlgMatrixC4M5N4U1S2 {{.*}}, i32 1, i32 5)  ; LinAlgMatrixSetElement(matrix,threadLocalIndex,value)
+  // CHECK: call %dx.types.LinAlgMatrixC4M5N4U1S2 @dx.op.linAlgMatrixSetElement.mC4M5N4U1S2.mC4M5N4U1S2.i32
+  // CHECK-SAME: (i32 -2147483629, %dx.types.LinAlgMatrixC4M5N4U1S2 {{.*}}, i32 1, i32 5)
+  // CHECK-SAME: ; LinAlgMatrixSetElement(matrix,threadLocalIndex,value)
+
+  // CHECK2: call void @"dx.hl.op..void (i32, %dx.types.LinAlgMatrixC4M5N4U1S2*, %dx.types.LinAlgMatrixC4M5N4U1S2, i32, i32)
+  // CHECK2-SAME: "(i32 412, %dx.types.LinAlgMatrixC4M5N4U1S2* {{.*}}, %dx.types.LinAlgMatrixC4M5N4U1S2 {{.*}}, i32 1, i32 5)
+
   __builtin_LinAlgMatrix [[__LinAlgMatrix_Attributes(4, 5, 4, 1, 2)]] mat1;
   __builtin_LinAlgMatrix [[__LinAlgMatrix_Attributes(4, 5, 4, 1, 2)]] mat2;
   __builtin_LinAlg_MatrixSetElement(mat2, mat1, 1, 5);
diff --git a/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/matrixstoretodescriptor/nominal.hlsl b/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/matrixstoretodescriptor/nominal.hlsl
@@ -1,13 +1,19 @@
 // REQUIRES: dxil-1-10
 // RUN: %dxc -T cs_6_10 -E main %s | FileCheck %s
+// RUN: %dxc -T cs_6_10 -E main -fcgl %s | FileCheck %s --check-prefix=CHECK2
 
 RWByteAddressBuffer outbuf;
 
 [numthreads(1,1,1)]
 void main() {
   // CHECK-LABEL: define void @main()
 
-  // CHECK: call void @dx.op.linAlgMatrixStoreToDescriptor.mC4M5N4U1S2(i32 -2147483628, %dx.types.LinAlgMatrixC4M5N4U1S2 {{.*}}, %dx.types.Handle %{{.*}}, i32 1, i32 1, i32 0)  ; LinAlgMatrixStoreToDescriptor(matrix,handle,offset,stride,layout)
+  // CHECK: call void @dx.op.linAlgMatrixStoreToDescriptor.mC4M5N4U1S2(i32 -2147483628,
+  // CHECK-SAME: %dx.types.LinAlgMatrixC4M5N4U1S2 {{.*}}, %dx.types.Handle %{{.*}}, i32 1, i32 1, i32 0)
+  // CHECK-SAME:  ; LinAlgMatrixStoreToDescriptor(matrix,handle,offset,stride,layout)
+
+  // CHECK2: call void @"dx.hl.op..void (i32, %dx.types.LinAlgMatrixC4M5N4U1S2, %dx.types.Handle, i32, i32, i32)
+  // CHECK2-SAME: "(i32 413, %dx.types.LinAlgMatrixC4M5N4U1S2 {{.*}}, %dx.types.Handle {{.*}}, i32 1, i32 1, i32 0)
   __builtin_LinAlgMatrix [[__LinAlgMatrix_Attributes(4, 5, 4, 1, 2)]] mat1;
   __builtin_LinAlg_MatrixStoreToDescriptor(mat1, outbuf, 1, 1, 0);
 }
diff --git a/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/matrixvectormultiply/nominal.hlsl b/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/matrixvectormultiply/nominal.hlsl
@@ -1,5 +1,6 @@
 // REQUIRES: dxil-1-10
 // RUN: %dxc -T cs_6_10 -E main %s | FileCheck %s
+// RUN: %dxc -T cs_6_10 -E main -fcgl %s | FileCheck %s --check-prefix=CHECK2 
 
 [numthreads(1,1,1)]
 void main() {
@@ -9,6 +10,11 @@ void main() {
   float4 vec = {1,2,3,4};
   float4 result;
 
-  // CHECK: call <4 x float> @dx.op.linAlgMatVecMul.v4f32.mC4M5N4U1S2.v4f32(i32 -2147483623, %dx.types.LinAlgMatrixC4M5N4U1S2 {{.*}}, <4 x float> <float 1.000000e+00, float 2.000000e+00, float 3.000000e+00, float 4.000000e+00>, i32 1)  ; LinAlgMatVecMul(matrix,inputVector,interpretation)
+  // CHECK: call <4 x float> @dx.op.linAlgMatVecMul.v4f32.mC4M5N4U1S2.v4f32(i32 -2147483623,
+  // CHECK-SAME: %dx.types.LinAlgMatrixC4M5N4U1S2 {{.*}}, <4 x float> <float 1.000000e+00, float 2.000000e+00,
+  // CHECK-SAME: float 3.000000e+00, float 4.000000e+00>, i32 1)  ; LinAlgMatVecMul(matrix,inputVector,interpretation)
+
+  // CHECK2: call void @"dx.hl.op..void (i32, <4 x float>*, %dx.types.LinAlgMatrixC4M5N4U1S2, <4 x float>, i32)
+  // CHECK2-SAME: "(i32 422, <4 x float>* %result, %dx.types.LinAlgMatrixC4M5N4U1S2 %5, <4 x float> %3, i32 1)
   __builtin_LinAlg_MatrixVectorMultiply(result, mat, vec, 1);
 }
diff --git a/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/matrixvectormultiplyadd/nominal.hlsl b/tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/matrixvectormultiplyadd/nominal.hlsl
@@ -1,5 +1,6 @@
 // REQUIRES: dxil-1-10
 // RUN: %dxc -T cs_6_10 -E main %s | FileCheck %s
+// RUN: %dxc -T cs_6_10 -E main -fcgl %s | FileCheck %s --check-prefix=CHECK2
 
 [numthreads(1,1,1)]
 void main() {
@@ -8,6 +9,13 @@ void main() {
   float4 vec = {1,2,3,4};
   float4 result;
 
-  // CHECK: call <4 x float> @dx.op.linAlgMatVecMulAdd.v4f32.mC5M3N4U0S0.v4f32.v4f32(i32 -2147483622, %dx.types.LinAlgMatrixC5M3N4U0S0 {{.*}}, <4 x float> <float 1.000000e+00, float 2.000000e+00, float 3.000000e+00, float 4.000000e+00>, i32 1, <4 x float> {{.*}}, i32 0)  ; LinAlgMatVecMulAdd(matrix,inputVector,inputInterpretation,biasVector,biasInterpretation)
+  // CHECK: call <4 x float> @dx.op.linAlgMatVecMulAdd.v4f32.mC5M3N4U0S0.v4f32.v4f32(i32 -2147483622,
+  // CHECK-SAME: %dx.types.LinAlgMatrixC5M3N4U0S0 {{.*}}, <4 x float> <float 1.000000e+00, 
+  // CHECK-SAME: float 2.000000e+00, float 3.000000e+00, float 4.000000e+00>, i32 1, <4 x float> {{.*}}, i32 0)
+  // CHECK-SAME: ; LinAlgMatVecMulAdd(matrix,inputVector,inputInterpretation,biasVector,biasInterpretation)
+
+  // CHECK2:  call void @"dx.hl.op..void (i32, <4 x float>*, %dx.types.LinAlgMatrixC5M3N4U0S0, <4 x float>, i32,
+  // CHECK2-SAME: <4 x float>, i32)"(i32 423, <4 x float>* %result, %dx.types.LinAlgMatrixC5M3N4U0S0 %6,
+  // CHECK2-SAME: <4 x float> %4, i32 1, <4 x float> %3, i32 0)
   __builtin_LinAlg_MatrixVectorMultiplyAdd(result, mat, vec, 1, result, 0);
 }
