-
Notifications
You must be signed in to change notification settings - Fork 0
Gromacs Benchmark
Zhe Shen edited this page Sep 15, 2020
·
1 revision
Note: The experiments use data provided by SJTU Biology research group
Parameters:
- r: MPI rank
- t: OpenMP thread
- CPU: CPU usage of
gmxprocess - MEM: Memory usage of
gmxprocess - GPU: Overall GPU usage
- US: CPU user time
- SY: CPU system time
- ID: CPU idle time
- Rate: Gromacs processing efficiency, unit: ns/hour
- CPU: Intel(R) Xeon(R) Gold 5120 CPU @ 2.20GHz
- GPU: Nvidia Tesla P4
| 2 CPU, 1 GPU | CPU | MEM | GPU | US | SY | ID | Rate |
|---|---|---|---|---|---|---|---|
| 1 r, 2 t | 195% | 1% | 73% | 92.4% | 7.5% | 0.2% | 3.521 |
| 2 r, 1 t | 195.3% | 0.6% | 54% | 96.7% | 1.5% | 1.8% | 2.165 |
| \ | 199.7% | 0.2% | \ | 100% | 0% | 0% | 0.300 |
| 4 CPU, 1 GPU | CPU | MEM | GPU | US | SY | ID | Rate |
|---|---|---|---|---|---|---|---|
| 1 r, 4 t | 394% | 1% | 79% | 95 % | 4.7% | 0.3% | 3.968 |
| 2 r, 2 t | 389% | 0.7% | 75% | 96.1% | 2.7% | 1.3% | 3.367 |
| 4 r, 1 t | 366% | 0.6% | 75% | 89.3% | 3.9% | 6.8% | 3.195 |
| \ | 399% | 0.2% | \ | 99.8% | 0.2% | 0% | 0.553 |
| 8 CPU, 1 GPU | CPU | MEM | GPU | US | SY | ID | Rate |
|---|---|---|---|---|---|---|---|
| 1 r, 8 t | 791% | 1% | 84% | 97.1% | 2.3% | 0.6% | 4.149 |
| 2 r, 4 t | 786% | 0.7% | 76% | 97.5% | 1.6% | 0.8% | 4.149 |
| 4 r, 2 t | 756% | 0.7% | 79% | 82.7% | 2.8% | 4.5% | 3.704 |
| 8 r, 1 t | 576% | 0.7% | 73% | 69.3% | 4.6% | 26.1% | 2.762 |
| \ | 799% | 0.3% | \ | 99.8% | 0% | 0.2% | 0.962 |
| 16 CPU, 1 GPU | CPU | MEM | GPU | US | SY | ID | Rate |
|---|---|---|---|---|---|---|---|
| 1 r, 16 t | 1588% | 1.1% | 83% | 97.9% | 1.5% | 0.5% | 4.016 |
| 2 r, 8 t | 1580% | 0.9% | 83% | 98.2% | 1% | 0.8% | 4.444 |
| 4 r, 4 t | 1545% | 0.8% | 83% | 95.2% | 1.9% | 3% | 4.310 |
| 8 r, 2 t | 1347% | 0.8% | 65% | 81.7% | 3% | 15.3% | 2.907 |
| 16 r, 1 t | 830% | 0.8% | 61% | 47.6% | 4.4% | 48% | 1.577 |
| \ | 1600% | 0.5% | \ | 99.9% | 0.1% | 0% | 1.656 |
| 24 CPU, 1 GPU | CPU | MEM | GPU | US | SY | ID | Rate |
|---|---|---|---|---|---|---|---|
| 1 r, 24 t | 2385% | 1.1% | 83% | 98.5% | 0.9% | 0.6% | 4.032 |
| 2 r, 12 t | 2378% | 1.1% | 82% | 98.6% | 0.8% | 0.5% | 4.444 |
| 3 r, 8 t | 2360% | 1% | 81% | 97.5% | 1.1% | 1.4% | 4.274 |
| 4 r, 6 t | 2335% | 0.9% | 83% | 96.4% | 1.3% | 2.2% | 4.329 |
| 6 r, 4 t | 2245% | 0.9% | 77% | 92.3% | 1.7% | 5.9% | 3.704 |
| 8 r, 3 t | 2120% | 0.9% | 66% | 86.1% | 2.4% | 11.5% | 2.915 |
| 12 r, 2 t | 1840% | 0.8% | 64% | 74.3% | 2.7% | 23% | 2.151 |
| 24 r, 1 t | 1050% | 1% | 41% | 40% | 3.2% | 56.8% | 0.905 |
| \ | 2400% | 0.5% | \ | 99.9% | 0.1% | 0% | 2.273 |
| 32 CPU, 1 GPU | CPU | MEM | GPU | US | SY | ID | Rate |
|---|---|---|---|---|---|---|---|
| 1 r, 32 t | 3180% | 1.2% | 80% | 99.2% | 0.5% | 0.3% | 3.876 |
| 2 r, 16 t | 3170% | 1.3% | 79% | 98.9% | 0.6% | 0.6% | 4.310 |
| 4 r, 8 t | 3120% | 1.1% | 82% | 96.9% | 0.9% | 2.2% | 3.731 |
| 8 r, 4 t | 2920% | 1% | 67% | 90.2% | 1.4% | 8.3% | 2.577 |
| 16 r, 2 t | 2480% | 0.9% | 64% | 76.2% | 1.9% | 21.9% | 1.639 |
| 32 r, 1 t | 1050% | 1% | 30% | 29.1% | 2.7% | 68.2% | 0.455 |
| \ | 3200% | 0.6% | \ | 99.9% | 0.1% | 0% | 1.988 |