leavelet/cutlass
1
0
Fork 0
Code Issues Pull Requests Actions 25 Packages Projects Releases Wiki Activity

v3.9 (#2185)

Browse Source 
* v3.8 update x

* fix blackwell gg

* doc change

* doc change

* doc change

---------

Co-authored-by: yuzhai <yuzhai@nvidia.com>
Co-authored-by: Haicheng Wu <haichengw@nvidia.com>
Co-authored-by: Haicheng Wu <57973641+hwu36@users.noreply.github.com>
This commit is contained in:
Yujia Zhai
2025-03-21 01:52:23 -04:00
committed by GitHub
co-authored by yuzhai Haicheng Wu Haicheng Wu
parent 8c4d1dc47d
commit 62750a2b75
334 changed files with 91517 additions and 2656 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tools/library/include/cutlass/library/arch_mappings.h
+5 -1
View File
@@ -127,9 +127,13 @@ template <typename OperatorClass> struct ArchMap<arch::Sm100, OperatorClass> {
template <> struct ArchMap<arch::Sm100, arch::OpClassTensorOp> {
static int const kMin = 100;
static int const kMax = 100;
static int const kMax = 101;
};
template <typename OperatorClass> struct ArchMap<arch::Sm120, OperatorClass> {
static int const kMin = 120;
static int const kMax = 120;
};
/////////////////////////////////////////////////////////////////////////////////////////////////
tools/library/include/cutlass/library/library.h
+2 -3
View File
@@ -323,12 +323,11 @@ struct GemmUniversalArguments {
int swizzle_size{1};
int split_k_slices{1};
// For mixed input dtype kernels
bool is_mixed_dtype{false};
// For SM90 mixed input dtype kernels
bool is_sm90_mixed_dtype{false};
Sm90MixedInputWiderOperand wider_operand{Sm90MixedInputWiderOperand::B};
bool generate_scale_and_zero{false};
bool generate_dequantized_AB{false};
bool *dequantized_AB_ready{nullptr}; // Carry the info back to gemm_operation_profiler.cu
void *Scale{nullptr}; // Scale tensor
void *Zero{nullptr}; // Zero tensor
void *dequantized_AB{nullptr}; // Dequantized A or B tensor for verification