-
Notifications
You must be signed in to change notification settings - Fork 221
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Segmentation Fault on Loading CUDA #2453
Comments
A quick update, following this issue I have run
I get the following output:
Does this help with the issue? How do I fix the forward compatability? |
What happens if you run with |
So I ran the following:
and the output was:
|
I meant to try and precompile CUDA.jl with that env var set. |
My bad, I got the following after running
I believe this is what you meant correct? |
I don't know if this helps. but aomething else I just tried was deleting the
I got:
|
You're posting a lot of things here, please take a step back. If I understand your issue correctly:
|
It appears you have a correct understanding of the issue. I went back and tried rerunning the previous steps on a GPU to see if that would prompt the segmentation fault, and after running(note that I did this on a new environment):
I got the following output from trying to add CUDA, and it appears that I am having an issue(both on GPU and CPU) of CUDA_Runtime_jll failing to precompile causing CUDA to undergo a segmentation fault:
Does this help make the issue clearer? |
That confirms my understanding, however, can you try that again now with Lines 19 to 23 in beccab1
|
Sorry for the late response, I just ran that and the output was the following(it seemed to work!):
|
Thank you to everyone who took the time to respond to this issue! I am still unclear on what exactly was driving the error I was getting(I am open to continuing to work on this issue but feel free to close it), but I was able to fix it by creating a new environment in a clean directory and running the following modules on Casper for the UCAR HPC resources:
and then running my script. Thanks again for all the help! |
Great, thanks for following up. I think we can close this in favor of #2446 then, which should resolve the issue by more safely handling the forwards-compatible library. |
I am trying to load CUDA in my Julia script as follows:
using CUDA
and when I do so I get the following error message relating to a segmentation fault:
my versioninfo() for what I am running is:
and the status check is:
Any help would be appreciated! I have been working with @tomchor to fix this issue.
The text was updated successfully, but these errors were encountered: