Implementation of matrix multiplication algorithme combining different techniques:
- Naive implementation
- Improve with a temporary variable
- OpenMP parallelization
- Loop tiling
- Matrix transposition
Those techniques are also combined together to measure computation time.
Here's a preview of an execution on a Intel® i7 7500U CPU
| Method | Taken time | Improvement |
|---|---|---|
| Non-optimized | 8.309747 | 0.00% |
| tmp optimization | 5.413826 | 32.33% |
| OpenMP optimization | 5.593702 | 30.08% |
| transpose optimization | 2.852428 | 64.34% |
| transpose_OpenMP optimization | 2.988738 | 62.64% |
| 2D_tiling optimization | 4.061197 | 49.24% |
| 3D_tiling optimization | 4.765365 | 40.43% |
| 2D_tiling_transpose optimization | 2.749773 | 65.63% |
| 3D_tiling_transpose optimization | 2.743929 | 65.70% |
| 2D_tiling_transpose_OpenMP optimization | 2.747584 | 65.66% |