CUTLASS
CUDA Templates for Linear Algebra Subroutines and Solvers
|
Configuration for batched GEMM in which multiple matrix products are computed.
#include <library.h>
Public Attributes | |
gemm::GemmCoord | problem_size |
GEMM problem size. More... | |
int64_t | lda |
Leading dimension of A matrix. More... | |
int64_t | ldb |
Leading dimension of B matrix. More... | |
int64_t | ldc |
Leading dimension of C matrix. More... | |
int64_t | ldd |
Leading dimension of D matrix. More... | |
int64_t | batch_stride_A |
Stride between instances of the A matrix in memory. More... | |
int64_t | batch_stride_B |
Stride between instances of the B matrix in memory. More... | |
int64_t | batch_stride_C |
Stride between instances of the C matrix in memory. More... | |
int64_t | batch_stride_D |
Stride between instances of the D matrix in memory. More... | |
int | batch_count |
Number of GEMMs in batch. More... | |
int cutlass::library::GemmBatchedConfiguration::batch_count |
int64_t cutlass::library::GemmBatchedConfiguration::batch_stride_A |
int64_t cutlass::library::GemmBatchedConfiguration::batch_stride_B |
int64_t cutlass::library::GemmBatchedConfiguration::batch_stride_C |
int64_t cutlass::library::GemmBatchedConfiguration::batch_stride_D |
int64_t cutlass::library::GemmBatchedConfiguration::lda |
int64_t cutlass::library::GemmBatchedConfiguration::ldb |
int64_t cutlass::library::GemmBatchedConfiguration::ldc |
int64_t cutlass::library::GemmBatchedConfiguration::ldd |
gemm::GemmCoord cutlass::library::GemmBatchedConfiguration::problem_size |