Performance of the H2O-DFT-LS Benchmark on Piz Daint
| Nodes | Cores | Best time (No GPU) (s) | Config | Time with GPU (8 OMP threads per MPI task) (s) |
|---|---|---|---|---|
| 64 | 512 | 888.262 | 1 OMP threads per MPI task | 632.335 |
| 128 | 1024 | 450.648 | 1 OMP threads per MPI task | 338.648 |
| 256 | 2048 | 234.616 | 1 OMP threads per MPI task | 175.674 |
| 512 | 4096 | 125.772 | 1 OMP threads per MPI task | 98.741 |
| 1024 | 8192 | 66.73 | 1 OMP threads per MPI task | 55.828 |
| 2048 | 16384 | 39.227 | 1 OMP threads per MPI task | 43.089 |
| 4096 | 32768 | 27.9 | 2 OMP threads per MPI task | 34.669 |
