====== Performance of the Fayalite-FIST Benchmark on Piz Daint ======

^ Nodes ^ Cores ^ Best time (s) ^ Config ^ Time with GPU, 8 OMP threads per MPI task (s) ^
| 1 | 8 | 405.389| 1 OMP thread per MPI task, no GPU | 978.921 |
| 2 | 16 | 328.731 | 1 OMP thread per MPI task, no GPU | 649.531 |
| 4 | 32 | 260.75 | 1 OMP threads per MPI task, no GPU | 529.224 |
| 8 | 64 | 273.003 | 1 OMP threads per MPI task, no GPU | 480.87 |
| 16 | 128 | 229.692 | 2 OMP threads per MPI task, no GPU | 339.21 |
| 32 | 256 | 219.859 | 4 OMP threads per MPI task, no GPU | 321.637 |
| 64 | 512 | 207.972 | 2 OMP threads per MPI task, no GPU | 323.565 |

{{:performance:fayalite-fist-comparison-piz-daint-piz-daint-gpu.png?direct|}}