sgemm_wg: Implement 2D threadtiling

This commit is contained in:
Hansung Kim
2024-02-29 14:40:54 -08:00
parent a06b2dd20e
commit 6f4dfe5a0e
2 changed files with 66 additions and 28 deletions

View File

@@ -147,9 +147,9 @@ int main(int argc, char *argv[]) {
RT_CHECK(vx_dev_open(&device));
// FIXME: hardcoded
uint32_t dim_m = 32;
uint32_t dim_n = 32;
uint32_t dim_k = 32;
uint32_t dim_m = 64;
uint32_t dim_n = 64;
uint32_t dim_k = 64;
generate_source_matrix(dim_m, dim_n, dim_k);
generate_reference_matmul(dim_m, dim_n, dim_k);