Using Multiple GPUs does not improve simulation time
-
Hello!
I'm using a setup with two GPUs (RTX 2080 Super) and attempting to analyze if it's worth it to purchase another with two GPUs, or just go for a single. Currently what I'm seeing in the two log files is this:
Single GPU
The following hardware configuration has been chosen for the simulation:
Device (0): GeForce RTX 2080 SUPER using 824.146 MB
Elapsed time for 'Preparing for time update' was 00:00:17 wall clock time.
Starting solver CUDA (hardware accelerated).
Time Update
Simulation performed 66994 iterations.Dual GPU
Device (0): GeForce RTX 2080 SUPER using 414.731 MB
Device (1): GeForce RTX 2080 SUPER using 414.731 MB
Elapsed time for 'Preparing for time update' was 00:00:17 wall clock time.
Starting solver CUDA (hardware accelerated).
Time Update
Simulation performed 66994 iterations.The resulting completion times have minimal differences (27 m 14 s vs. 22 m 23 s) and is likely due to the reduction in memory usage on each GPU. Am I missing something or is there negligible difference between 1 and 2 GPUs?
Thanks
cmcc
Hello! It looks like you're interested in this conversation, but you don't have an account yet.
Getting fed up of having to scroll through the same posts each visit? When you register for an account, you'll always come back to exactly where you were before, and choose to be notified of new replies (either via email, or push notification). You'll also be able to save bookmarks and upvote posts to show your appreciation to other community members.
With your input, this post could be even better 💗
Register Login