leavelet/cutlass
1
0
Fork 0
Code Issues Pull Requests Actions 27 Packages Projects Releases Wiki Activity

CUTLASS 3.5.0 (#1411)

Browse Source
This commit is contained in:
Vijay Thakkar
2024-03-19 17:51:04 -04:00
committed by GitHub
parent ffa34e7075
commit 629f4653c3
468 changed files with 48729 additions and 7252 deletions
Download Patch File Download Diff File Expand all files Collapse all files
test/unit/epilogue/threadblock/epilogue_with_reduction_testbed.h
+1 -19
View File
@@ -133,29 +133,11 @@ __global__ void epilogue_with_reduction_threadblock(
// For debugging, enable this block of code to fill each accumulator element with its
// source thread ID.
CUTLASS_PRAGMA_UNROLL
for (int i = 0; i < accumulators.size(); ++i) {
for (size_t i = 0; i < accumulators.size(); ++i) {
typename Epilogue::WarpMmaOperator::ElementC x(threadIdx.x);
//typename Epilogue::WarpMmaOperator::ElementC x(i);
accumulators[i] = x;
}
/*
#pragma unroll 1
for (int tid = 0; tid < 32; ++tid) {
if (tid == thread_idx) {
printf("\nT%d: ", thread_idx);
CUTLASS_PRAGMA_UNROLL
for (int i = 0; i < accumulators.size(); ++i) {
printf("%d ", int(accumulators[i]));
}
}
}
if (thread_idx == 0) {
printf("\n\n");
}
*/
__syncthreads();
#endif
test/unit/epilogue/threadblock/testbed.h
+1 -19
View File
@@ -122,29 +122,11 @@ __global__ void epilogue_threadblock(
// For debugging, enable this block of code to fill each accumulator element with its
// source thread ID.
CUTLASS_PRAGMA_UNROLL
for (int i = 0; i < accumulators.size(); ++i) {
for (size_t i = 0; i < accumulators.size(); ++i) {
typename Epilogue::WarpMmaOperator::ElementC x(threadIdx.x);
//typename Epilogue::WarpMmaOperator::ElementC x(i);
accumulators[i] = x;
}
/*
#pragma unroll 1
for (int tid = 0; tid < 32; ++tid) {
if (tid == thread_idx) {
printf("\nT%d: ", thread_idx);
CUTLASS_PRAGMA_UNROLL
for (int i = 0; i < accumulators.size(); ++i) {
printf("%d ", int(accumulators[i]));
}
}
}
if (thread_idx == 0) {
printf("\n\n");
}
*/
__syncthreads();
#endif