How to Accelerate Computing with GPU #481
Closed
fanyimeng0
started this conversation in
General
Replies: 1 comment 2 replies
-
Hi @fanyimeng0, |
Beta Was this translation helpful? Give feedback.
2 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
I've been using Sionna for some simulations lately, but it seems that the computational power of the GPU is not being used to its full potential. This seems to be due to the fact that I'm using some other libraries to do the calculations, e.g., from skimage.feature import peak_local_max, and when using these, I'm converting the TensorFlow variables to NumPy and using a for loop instead of batch processing.
I have learned a lot from the Sionna discussion on GitHub and the Discover Sionna guide. However, I still can't solve my problem.
Does this mean that I can't use the GPU efficiently unless I completely rewrite some of my NumPy-related functions to TensorFlow's functions? If you have some development experience in this area, I hope you can provide some ideas. Thanks a lot!
Beta Was this translation helpful? Give feedback.
All reactions