Repository navigation
Replies: 2 comments 2 replies
|
Hello @gdlcmf, Are you holding on to anything between iteration, or are the results and per-iteration variables fully scoped to your Otherwise, have you tried calling the Python garbage collector at some regular intervals, e.g. every 10 iterations? import gc
...
paths = p_solver(...)
gc.collect() |
2 replies
|
Thank You,I tested versions 1.2.0 and 1.0.1. This issue occurred in version 1.0.1, but it no longer exists in version 1.2.0. Excellent! |
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Dear Sionna Developers,
I am writing to report a potential issue regarding VRAM management when running the Sionna solver multiple times in sequence.
I have designed a script to simulate propagation paths for a moving radiation source. The core computation part is implemented as a function that leverages the Sionna solver. This function is called thousands of times in a loop, each time with a different source position.
I have observed that the GPU VRAM usage increases rapidly and becomes fully occupied after a number of iterations. Consequently, the execution of my loop becomes progressively slower, which strongly suggests that the GPU memory allocated during each call is not being properly released.
My main questions are:
Could this be caused by an issue in my code (e.g., how I'm calling the Sionna functions), or does it indicate a possible memory leak within the Sionna codebase itself?
What are the recommended best practices to ensure proper VRAM cleanup between consecutive runs of Sionna simulations in a loop?
I would greatly appreciate any guidance on how to diagnose and resolve this issue. Thank you for your time and for developing this excellent tool.
Best regards,
Li
All reactions