NAMD 2.7b1 on an i7 965
We compare eight threads vs. four threads vs. the performance obtained from a Q6600-based quad. All measurements are in nanoseconds per day and were made with an SMP-aware executable of NAMD 2.7b1 using the flags '+setcpuaffinity +LBSameCpus'. Details of the examples used can be found here.
i7 with 1033 MHz memory
| i7, eight threads | i7, four threads | Q6600-based quad | Percent faster | |
|---|---|---|---|---|
| 100K (ApoA1) | 0.65 | 0.51 | 0.37 | 75% |
| 60K atoms | 1.61 | 1.45 | 0.95 | 69% |
| 25K atoms | 4.35 | 3.85 | 2.53 | 72% |
| 6.5K atoms | 15.4 | 15.4 | 9.8 | 57% |
| 1.6K atoms | 79.4 | 75.2 | 48 | 65% |
| 0.9K atoms | 145 | 125 | 82 | 76% |
Changing the memory to triple-channel at 1333 MHz didn't make much difference unfortunately (approximately 5% faster) as I should have expected given the small memory footprint of MD jobs.
