#include <cstdio>
#include <cstdlib>
#include <vector>
#include "realtype.h"
#include "utilities.h"
#include "gblas.h"
Functions | |
| static void | fill_matrix_with_random_numbers (int n, int m, std::vector< ergo_real > &A) |
| static void | do_naive_mmul (std::vector< ergo_real > &C, const std::vector< ergo_real > &A, const std::vector< ergo_real > &B, int n, int m, int p) |
| static void | verify_mmul_result (const std::vector< ergo_real > &A, const std::vector< ergo_real > &B, const std::vector< ergo_real > &C, int n, int m, int p) |
| ergo_real | compare_matrices (const std::vector< ergo_real > &A, const std::vector< ergo_real > &B, int n, int m) |
| int | main (int argc, char *argv[]) |
The idea is to run this linking to different BLAS variants with and without threading inside the BLAS gemm routine, to see how much speedup can be achieved from threading.
| ergo_real compare_matrices | ( | const std::vector< ergo_real > & | A, | |
| const std::vector< ergo_real > & | B, | |||
| int | n, | |||
| int | m | |||
| ) |
| static void do_naive_mmul | ( | std::vector< ergo_real > & | C, | |
| const std::vector< ergo_real > & | A, | |||
| const std::vector< ergo_real > & | B, | |||
| int | n, | |||
| int | m, | |||
| int | p | |||
| ) | [static] |
| static void fill_matrix_with_random_numbers | ( | int | n, | |
| int | m, | |||
| std::vector< ergo_real > & | A | |||
| ) | [static] |
| int main | ( | int | argc, | |
| char * | argv[] | |||
| ) |
| static void verify_mmul_result | ( | const std::vector< ergo_real > & | A, | |
| const std::vector< ergo_real > & | B, | |||
| const std::vector< ergo_real > & | C, | |||
| int | n, | |||
| int | m, | |||
| int | p | |||
| ) | [static] |
1.4.7