sp_mm_buffer_size

nvmath.bindings.cusparse.sp_mm_buffer_size(

intptr_t handle,

int op_a,

int op_b,

intptr_t alpha,

intptr_t mat_a,

intptr_t mat_b,

intptr_t beta,

intptr_t mat_c,

int compute_type,

int alg,

) → size_t

See cusparseSpMM_bufferSize.