Performance of Pytorch backend vs CPU #721
-
|
I see now several OT algorithms have pytorch backend support, I'm wondering if using the pytorch backend on GPU speeds up the convergence of these solvers? |
Beta Was this translation helpful? Give feedback.
Answered by
rflamary
Feb 5, 2025
Replies: 1 comment
-
|
Yes it should be faster on GPU. Still some solvers (exact OT, emd, non regularized GW) use a CPU C++ solvers and those ones will work with GPU array but incur a memory copy overhead. |
Beta Was this translation helpful? Give feedback.
0 replies
Answer selected by
RishalAggarwal
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Yes it should be faster on GPU. Still some solvers (exact OT, emd, non regularized GW) use a CPU C++ solvers and those ones will work with GPU array but incur a memory copy overhead.