Performance Optimization on threadfence
24.7Energy (J)LASSI-EE
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| LASSI-EEGPU=A100, Implementation=LASSI-EE2025.05 | 24.7 | 14.6 | 1.3 | 93.3 | 71.4 | 81.8 | 0 | 7 | 18.2 | |
| LASSI-EEGPU=MI100, Implementation=LASSI-EE2025.05 | 40.5 | 27.1 | 1.5 | 79.1 | 51.9 | 56.7 | 0 | 6 | 27.4 | |
| Source codeGPU=MI100, Implementation=Source2025.05 | 194.1 | 56.4 | 3.4 | — | — | — | — | — | — | |
| Source codeGPU=A100, Implementation=Source2025.05 | 365.7 | 50.8 | 7.2 | — | — | — | — | — | — |