CUTLASS
CUDA Templates for Linear Algebra Subroutines and Solvers

threadblock → layout Relation

File in include/cutlass/gemm/threadblockIncludes file in include/cutlass/layout
default_gemv_core.hlayout/matrix.h
default_mma_core_sm50.hlayout/matrix.h
default_mma_core_sm70.htensor_op_multiplicand_sm70.h
default_mma_core_sm75.htensor_op_multiplicand_sm75.h