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

CUTLASS 3.2.1 (#1113)

Browse Source 
* Updates for 3.2.1 release.

* Minor fix in gemm op profiler for raster order.

* Add scheduler mapping for raster order in the kernels.
This commit is contained in:
ANIKET SHIVAM
2023-09-26 14:24:26 -07:00
committed by GitHub
parent e0aaa3c3b3
commit 90d3b0fb18
428 changed files with 22253 additions and 21762 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
include/cute/algorithm/gemm.hpp
View File

@@ -218,7 +218,6 @@ gemm(MMA_Atom<MMA> const& mma,
CUTE_STATIC_ASSERT_V(size<0>(A) == size<0>(C)); // AM == CM
CUTE_STATIC_ASSERT_V(size<0>(B) == size<1>(C)); // BN == CN
CUTE_STATIC_ASSERT_V(size<0>(C) == size<0>(D) && size<1>(C) == size<1>(D));
gemm(mma,
D, // (M,N)
make_tensor(A.data(), append<2>(A.layout())), // (M,1)
@@ -253,7 +252,7 @@ gemm(MMA_Atom<MMA> const& mma,
CUTE_STATIC_ASSERT_V(size<1>(typename MMA_Atom<MMA>::LayoutC_TV{}) == Int<1>{});
CUTE_STATIC_ASSERT_V(size<1>(typename MMA_Atom<MMA>::LayoutA_TV{}) == Int<1>{});
CUTE_STATIC_ASSERT_V(size<1>(typename MMA_Atom<MMA>::LayoutB_TV{}) == Int<1>{});
gemm(mma,
make_tensor(D.data(), prepend<3>(D.layout())), // (1,M,N)
make_tensor(A.data(), prepend<3>(A.layout())), // (1,M,K)
@@ -282,7 +281,6 @@ gemm(MMA_Atom<MMA> const& mma,
CUTE_STATIC_ASSERT_V(size<1>(A) == size<1>(C)); // AM == CM
CUTE_STATIC_ASSERT_V(size<1>(B) == size<2>(C)); // BN == CN
CUTE_STATIC_ASSERT_V(size<0>(C) == size<0>(D) && size<1>(C) == size<1>(D) && size<2>(C) == size<2>(D));
auto M = size<1>(A);
auto N = size<1>(B);
// REGISTER .reuse OPTIMIZATIONS
@@ -409,7 +407,6 @@ gemm(MMA_Atom<MMA> const& mma,
CUTE_STATIC_ASSERT_V(size<1>(B) == size<2>(C)); // BN == CN
CUTE_STATIC_ASSERT_V(size<2>(A) == size<2>(B)); // AK == BK
CUTE_STATIC_ASSERT_V(size<0>(C) == size<0>(D) && size<1>(C) == size<1>(D) && size<2>(C) == size<2>(D));
auto K = size<2>(A);
CUTE_UNROLL
@@ -454,7 +451,6 @@ gemm(MMA_Atom<MMA> const& mma,
CUTE_STATIC_ASSERT_V(size<1>(typename MMA_Atom<MMA>::LayoutC_TV{}) == Int<1>{});
CUTE_STATIC_ASSERT_V(size<1>(typename MMA_Atom<MMA>::LayoutA_TV{}) == Int<1>{});
CUTE_STATIC_ASSERT_V(size<1>(typename MMA_Atom<MMA>::LayoutB_TV{}) == Int<1>{});
gemm(mma,
make_tensor(D.data(), prepend<3>(D.layout())), // (1,M,N)
make_tensor(A.data(), prepend<3>(A.layout())), // (1,M,K)