RFIC performance metric prediction on RF circuit dataset (test)
25.6PDCFALCON
Evaluation results
| Method | Links | |||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| FALCONTraining Dataset configuration=Reduced, Number of Models=1, Training Samples=60k×5, Training Hours=∼1002025.08 | 25.6 | 9.8 | 25.2 | 10.2 | 14.3 | 4.4 | 5.4 | 10.6 | 0.9 | 7.6 | 5.6 | 11.2 | 10.6 | 27.3 | 1.6 | 3.3 | 13.72 | |
| FALCONTraining Dataset configuration=Per-class, Number of Models=5, Training Samples=60k×5, Training Hours=∼502025.08 | 15.6 | 25.7 | 163.8 | 11.1 | 25.1 | 22.6 | 9.7 | 30.8 | 1.1 | 6.9 | 5.1 | 31.9 | 28.2 | 64.7 | 1.6 | 4 | 31.74 | |
| FALCONTraining Dataset configuration=Complete, Number of Models=1, Training Samples=684k×1, Training Hours=∼1002025.08 | 11.2 | 2.6 | 19 | 6.1 | 11.4 | 1.9 | 4.5 | 6.5 | 0.6 | 6.5 | 4.6 | 4.4 | 4.6 | 11 | 1.3 | 1.4 | 9.09 | |
| RF-Informed GNNTraining Dataset configuration=Per-class, Number of Models=5, Training Samples=60k×5, Training Hours=22025.08 | 0.7 | 11.5 | 8.1 | 14.9 | 3.9 | 1.5 | 1.2 | 3 | 0.7 | 5.4 | 4.4 | 2.7 | 1.5 | 5.7 | 1.2 | 2 | 3.45 |