Table 3.
Benchmarking results of the FE and FD simulators on a representative case of a cancerous prostate model. Cryoprobe layout is based on a bubble packing method (illustrate d in Fig. 4). For the FE simulator, a time step of the 60 sec represents an upper limit for convergence, whereas the 30 sec time step is deemed practical for concurrent defect region calculations. Time step for the FD simulator is taken at the stability li mit.
| Simulator | FE | FD | ||
|---|---|---|---|---|
| Time step | 60 sec | 30 sec | .02 sec | |
| Minimum defect value | 24.5% | 24.7% | 24.3% | |
| Total simulated time | 255 sec | 255 sec | 255 sec | |
| Simulated time to minimum defect | 180 sec | 199 sec | 255 sec | |
| Actual runtime | 1 core | 613 sec | 966 sec | 144 sec |
| 2 cores | 423 sec | 689 sec | 83 sec | |
| 4 cores | 304 sec | 521 sec | 67 sec | |
| Simulated time ratio = comp. runtime / procedure duration |
1 core | 236% | 379% | 56% |
| 2 cores | 165% | 270% | 32% | |
| 4 cores | 119% | 204% | 26% | |