Table 2 Average running time for ninitial = 1 on Intel(R) Xeon(R) Gold 6278C CPU @ 2.60GHz and Nvidia(R) Tesla(R) V100
From: Reply to: Deep reinforced learning heuristic tested on spin-glass ground states: The larger picture
3D, L = 10 | 3D, L = 20 | 3D, L = 30 | |
|---|---|---|---|
DIRAC1(Python,C++;GPU) | ~2.8s | ~49s | ~330s |
DIRAC1(MATLAB;GPU) | ~0.5s | ~1.1s | ~3.2s |
SA sweep(Python,C++) | ~0.015s | ~0.18s | ~0.61s |
SA sweep(MATLAB) | ~0.01s | ~0.09s | ~0.41s |
SA sweep(C++) | ~0.003s | ~0.034s | ~0.13s |
