From: Covariance tracking: architecture optimizations for embedded systems
Sequence | Panda | Pedxing | ||
---|---|---|---|---|
Size | 312×233 | 640×480 | ||
Intel Penryn-M | ||||
Algorithm version | SoA | AoS++ | SoA | AoS++ |
Features computation (cpp) | 128 | 150 | 128 | 150 |
Kernel computation (cpp) | 599 | 87 | 618 | 91 |
Tracking (cpp) | 23 | 23 | 11 | 11 |
Total (cpp) | 738 | 248 | 769 | 264 |
Kernel/total | 81% | 35% | 80% | 34% |
Total speedup | ×2.9 | ×2.8 | ||
1-C execution time (ms) | 45 | 15 | 197 | 68 |
2-C execution time (ms) | 36 | 9 | 158 | 38 |
Intel Haswell-M | ||||
Algorithm version | SoA | AoS++ | SoA | AoS++ |
Features computation (cpp) | 78 | 79 | 88 | 72 |
Kernel computation (cpp) | 190 | 36 | 207 | 40 |
Tracking (cpp) | 13 | 23 | 2 | 3 |
Total (cpp) | 281 | 138 | 297 | 115 |
Kernel/total | 67% | 26% | 69% | 34% |
Total speedup | ×2.0 | ×2.6 | ||
1-C execution time (ms) | 12 | 5 | 54 | 21 |
2-C execution time (ms) | 8 | 3 | 37 | 13 |