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

releaase 2.11 (#703)

Browse Source
This commit is contained in:
Aditya Atluri
2022-11-19 06:02:15 -08:00
committed by GitHub
parent 3c90f6aea6
commit c975e2ccbb
329 changed files with 47332 additions and 10607 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
include/cutlass/gemm/kernel/grouped_problem_visitor.h
View File

@@ -126,11 +126,7 @@ struct BaseGroupedProblemVisitor {
/// Get the grid shape
CUTLASS_HOST_DEVICE
static cutlass::gemm::GemmCoord grid_shape(const cutlass::gemm::GemmCoord& problem) {
return cutlass::gemm::GemmCoord(
((problem.m() - 1 + ThreadblockShape::kM) / ThreadblockShape::kM),
((problem.n() - 1 + ThreadblockShape::kN) / ThreadblockShape::kN),
1);
return ProblemSizeHelper::grid_shape(problem);
}
/// Gets the global tile index
@@ -346,7 +342,7 @@ struct GroupedProblemVisitor<ProblemSizeHelper,
PrefetchTileCount,
ThreadCount> : public BaseGroupedProblemVisitor<ProblemSizeHelper, ThreadblockShape> {
static_assert(PrefetchTileCount > 0,
"GroupedProblemVisitor with GroupScheduleMode `kHost` currently requires prefetching to shared memory");
"GroupedProblemVisitor with GroupScheduleMode `kHostPrecompute` currently requires prefetching to shared memory");
using Base = BaseGroupedProblemVisitor<ProblemSizeHelper, ThreadblockShape>;
using Params = typename Base::Params;