Wiki source code of Benchmarking
Hide last authors
| author | version | line-number | content |
|---|---|---|---|
 |
1.1 | 1 | == WIP == |
| 2 | |||
| 3 | == TVB-INVERSION 1.0.0 == | ||
| 4 | |||
| 5 | === Sampling priors - Remote execution === | ||
| 6 | |||
| 7 | ==== ==== | ||
| 8 | |||
| 9 | ==== 1. Execution times ==== | ||
| 10 | |||
| 11 | In this section we provide some benchmarks regarding the execution times for the sampling priors step within tvb-inversion workflow, where we need to run a large number of simulations. | ||
| 12 | |||
| 13 | These have been computed on DAINT-CSCS HPC, on a single node, with different configurations for the number of simulations and the number of workers. | ||
| 14 | |||
| |
6.1 | 15 | |=(% scope="row" %)((( |
| 16 | Model | ||
| 17 | )))|=((( | ||
| 18 | Sim length (s) | ||
| 19 | )))|=Regions|=Nr simulations|=Nr workers|=Execution time (hh:mm) | ||
| 20 | |=MontbrioPazoRoxin|30|100|((( | ||
| 21 | 30 | ||
| 22 | )))|30|00:17 | ||
| 23 | |=MontbrioPazoRoxin|30|100|200|20|01:08 | ||
| 24 | |=MontbrioPazoRoxin|30|100|300|30|01:10 | ||
| 25 | |=MontbrioPazoRoxin|30|100|400|40|01:18 | ||
| 26 | |=MontbrioPazoRoxin|30|100|500|50|01:34 | ||
| 27 | |=MontbrioPazoRoxin|30|100|500|55|01:30 | ||
| 28 | |=MontbrioPazoRoxin|30|100|600|55|01:45 | ||
| 29 | |=MontbrioPazoRoxin|30|100|600|60|OOM | ||
| 30 | |=MontbrioPazoRoxin|60|100|500|40|((( | ||
| 31 | 03:07 | ||
| 32 | ))) | ||
| 33 | |(% scope="row" %)**MontbrioPazoRoxin**|60|100|500|55|OOM | ||
| 34 | |= |= |= |= |= |= | ||
| 35 | |= | | | | | | ||
| 36 | |= | | | | | | ||
| 37 | |= | | | | | | ||
| 38 | |= | | | | | | ||
| 39 | |= | | | | | | ||
| 40 | |= | | | | | | ||
| 41 | |= | | | | | | ||
| 42 | |= | | | | | | ||
| 43 | |= | | | | | | ||
| 44 | |= | | | | | | ||
| 45 | |= | | | | | | ||
| |
1.1 | 46 | |
| |
13.1 | 47 | ==== ==== |
| |
9.1 | 48 | |
| |
1.1 | 49 | ==== 2. Limitations ==== |
| 50 | |||
| 51 | * Reaching he memory limit on the CSCS node. | ||
| |
3.1 | 52 | ** For a simulation of 30 seconds, we can fit 55 parallel workers in the available memory |
| 53 | ** For a simulation of 60 seconds, we can fit 40 parallel workers in the available memory | ||
| |
1.1 | 54 | * Reaching the maximum connections on CSCS could happen during a run |
| 55 | ** connect once to launch the job | ||
| 56 | ** connect multiple times during the monitoring step to check the status of the job | ||
| 57 | ** connect once to stage out results |