Add new coopmat2 functions: Reduce and Transpose (#7027)

This commit adds three new functions for CoopMat as described in the proposal document,

Cooperative matrix 2 proposal spec#12
The new functions are:

CoopMat<T,S,M,N,R>::Transpose
CoopMat<T,S,M,N,R>::ReduceRow
CoopMat<T,S,M,N,R>::ReduceColumn
CoopMat<T,S,M,N,R>::ReduceRowAndColumn
CoopMat<T,S,M,N,R>::Reduce2x2

1 files changed, 9 insertions, 4 deletions

diff --git a/tests/cooperative-matrix/load-store-rwstructuredbuffer.slang b/tests/cooperative-matrix/load-store-rwstructuredbuffer.slang
index e161fb7b2..6a94fd30e 100644
--- a/tests/cooperative-matrix/load-store-rwstructuredbuffer.slang
+++ b/tests/cooperative-matrix/load-store-rwstructuredbuffer.slang
@@ -1,4 +1,5 @@
 //TEST(compute):COMPARE_COMPUTE(filecheck-buffer=CHECK):-vk -output-using-type -emit-spirv-directly -skip-spirv-validation
+//TEST(compute):COMPARE_COMPUTE(filecheck-buffer=CHECK):-vk -output-using-type -emit-spirv-directly -skip-spirv-validation -Xslang -DRWSB
 
 // CHECK: type: int32_t
 // CHECK-NEXT: 1
@@ -11,17 +12,21 @@
 // CHECK-NEXT: 8
 
 //TEST_INPUT:ubuffer(data=[1 2 3 4 5 6 7 8], stride=4, count=256),name=buf
+#if defined(RWSB)
 RWStructuredBuffer<int32_t> inputBuffer;
+#else
+StructuredBuffer<int32_t> inputBuffer;
+#endif
 
 //TEST_INPUT:ubuffer(stride=4, count=256):out,name=outputBuffer
 RWStructuredBuffer<int32_t> outputBuffer;
 
+using namespace linalg;
+
 [numthreads(32, 1, 1)]
 void computeMain()
 {
     let stride = 16;
-    let matrixLayout = CoopMatMatrixLayout::RowMajor;
-
-    let mat = coopMatLoad<int32_t, CoopMatScope::Subgroup, 16, 16, CoopMatMatrixUse::MatrixAccumulator>(inputBuffer, 0, stride, matrixLayout);
-    mat.store(outputBuffer, 0, stride, matrixLayout);
+    let mat = coopMatLoad<int32_t, MemoryScope.Subgroup, 16, 16, CoopMatMatrixUse.MatrixAccumulator, CoopMatMatrixLayout.RowMajor>(inputBuffer, 0, stride);
+    mat.Store<CoopMatMatrixLayout.RowMajor>(outputBuffer, 0, stride);
 }