CineBench R11.5 64bit | ||||
| 处理器 | 双路Intel Westmere-EP Xeon X5680 | DELL PowerEdge M910 四路Intel Nehalem-EX Xeon E7540 | 四路Intel Nehalem-EX Xeon X7560 超线程关闭 | 四路Intel Nehalem-EX Xeon X7560 |
| 显卡 | - | - | - | - |
CPU Benchmark | ||||
| Rendering (1 CPU) | 1.16 pts | 0.78 pts | 0.92 pts | 0.92 pts |
| Rendering (x CPU) | 16.40 pts | 19.26 pts | 15.70 pts | 27.08 pts |
| Threads | 24 | 48 | 32 | 64 |
| Multiprocessor Speedup | 14.18x | 24.65x | 17.14x | 29.43x |
CineBench受超线程影响很大，降低了42%的性能。
MMM - Matrix-Matrix Multiplicaion Benchmark | ||||
| 处理器 | 双路Intel Westmere-EP Xeon X5680 | DELL PowerEdge M910 四路Intel Nehalem-EX Xeon E7540 | 四路Intel Nehalem-EX Xeon X7560 超线程关闭 | 四路Intel Nehalem-EX Xeon X7560 |
| 单位 | GFLOPS | GFLOPS | GFLOPS | GFLOPS |
Threads 1 | ||||
| 5000 step | 8.877563 | 5.867645 | 6.965506 | 7.036748 |
| 10000 step | 8.883291 | 5.865347 | 6.974666 | 7.034792 |
| 15000 step | 8.881528 | 5.826604 | 6.935187 | 6.960592 |
Threads 2 | ||||
| 5000 step | 17.5891 | 11.570261 | 13.944669 | 13.979099 |
| 10000 step | 17.73566 | 11.689317 | 13.984364 | 14.032266 |
| 15000 step | 17.67208 | 11.602247 | 13.920488 | 13.89951 |
Threads 4 | ||||
| 5000 step | 34.85343 | 21.788607 | 27.715225 | 27.555005 |
| 10000 step | 34.90105 | 22.21115 | 27.846714 | 27.685804 |
| 15000 step | 34.92557 | 22.073702 | 27.702342 | 27.297404 |
Threads 8 | ||||
| 5000 step | 45.99856 | 41.228878 | 49.714209 | 49.518835 |
| 10000 step | 45.99856 | 43.472432 | 50.227199 | 49.767277 |
| 15000 step | 47.74417 | 43.37777 | 50.514463 | 49.750117 |
Threads 16 | ||||
| 5000 step | 66.10022 | 73.379889 | 96.880558 | 47.855051 |
| 10000 step | 72.38159 | 78.596851 | 99.776303 | 48.320744 |
| 15000 step | 73.2495 | 79.099092 | 99.385572 | 48.421492 |
Threads 24 | ||||
| 5000 step | 66.10022 | 94.000418 | 138.717733 | 70.233111 |
| 10000 step | 72.38159 | 124.028823 | 148.818098 | 71.502532 |
| 15000 step | 73.2495 | 124.574801 | 148.054922 | 71.685326 |
Threads 32 | ||||
| 5000 step | 151.664487 | 96.076302 | ||
| 10000 step | 184.741778 | 97.64478 | ||
| 15000 step | 183.894563 | 98.195937 | ||
Threads 48 | ||||
| 5000 step | 97.335138 | 136.623189 | ||
| 10000 step | 119.780984 | 142.50301 | ||
| 15000 step | 121.637469 | 145.534244 | ||
Threads 64 | ||||
| 5000 step | 137.141474 | |||
| 10000 step | 182.49902 | |||
| 15000 step | 185.098571 | |||
打开关闭的峰值计算性能都差不多，表明了这些程序对运算单元的利用率高，超线程起到的作用不太大。
SunGard Adaptiv Analytics Benchmark v4.0 | ||||
| 处理器 | 双路Intel Westmere-EP Xeon X5680 | DELL PowerEdge M910 四路Intel Nehalem-EX Xeon E7540 | 四路Intel Nehalem-EX Xeon X7560 超线程关闭 | 四路Intel Nehalem-EX Xeon X7560 |
| Threads | 24 | 48 | 32 | 64 |
| Time (lower is better) | 94.911s | 139.512s | 66.986s | 104.925s |
这个程序本身就推荐关闭超线程运行。
black_scholes | ||||
| 处理器 | 双路Intel Westmere-EP Xeon X5680 | DELL PowerEdge M910 四路Intel Nehalem-EX Xeon E7540 | 四路Intel Nehalem-EX Xeon X7560 超线程关闭 | 四路Intel Nehalem-EX Xeon X7560 |
| Threads | 24 | 48 | 32 | 64 |
| Time (lower is better) | 5.51s | 4.40s | 3.21s | 2.78s |
这个程序可以有效地利用多线程。
