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

Collection of changes to fix clang build. (#1200)

Browse Source 
* Remove unused variables

* Qualify calls to make_fragment_? from templated base class.

Fixes clang build error.

* Add missing `#include <cstdio>`

* Various changes to fix clang compile errors.

* More changes to fix clang build.

Remaining issues:

- `params` initializer of `CollectiveEpilogue`.
- `ops` initializer of `Sm90VisitorImplBase`.
- `__usAtomicCAS` needs to be added to clang upstream.

* Fix remaining clang build issues.

* Qualify `cute::rank()` calls.

* Qualify some more calls that are otherwise ambiguous between `cute` and `std` namespace.

* Double-escape special registers in inline asm.

* small change

---------

Co-authored-by: Haicheng Wu <haichengw@nvidia.com>
This commit is contained in:
Christian Sigg
2023-12-08 20:42:12 +01:00
committed by GitHub
parent f4a0216601
commit e1483d5fa0
46 changed files with 308 additions and 273 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
tools/util/include/cutlass/util/reference/device/gett.hpp
View File

@ -126,7 +126,7 @@ gett(
cudaStream_t stream = 0) {
using namespace cute;
static_assert(rank(ProblemShapeMNKL{}) == 4);
static_assert(cute::rank(ProblemShapeMNKL{}) == 4);
auto M = get<0>(problem_shape_mnkl);
auto N = get<1>(problem_shape_mnkl);
auto K = get<2>(problem_shape_mnkl);

8
tools/util/include/cutlass/util/reference/host/gett.hpp
View File

@ -431,11 +431,11 @@ void Gemm3x(
{
using namespace cute;
static_assert(rank(typename MainloopParams::LayoutA{}) == rank(typename MainloopParams::LayoutB{}));
static_assert(rank(typename EpilogueParams::LayoutC{}) == rank(typename EpilogueParams::LayoutD{}));
static_assert(rank(typename MainloopParams::LayoutA{}) == rank(typename EpilogueParams::LayoutC{}));
static_assert(cute::rank(typename MainloopParams::LayoutA{}) == cute::rank(typename MainloopParams::LayoutB{}));
static_assert(cute::rank(typename EpilogueParams::LayoutC{}) == cute::rank(typename EpilogueParams::LayoutD{}));
static_assert(cute::rank(typename MainloopParams::LayoutA{}) == cute::rank(typename EpilogueParams::LayoutC{}));
if constexpr (rank(typename MainloopParams::LayoutA{}) == 2) {
if constexpr (cute::rank(typename MainloopParams::LayoutA{}) == 2) {
Layout layout_A = make_layout_rank3(mainloop_params.A);
Layout layout_B = make_layout_rank3(mainloop_params.B);
Layout layout_C = make_layout_rank3(epilogue_params.C);