-
Notifications
You must be signed in to change notification settings - Fork 215
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
NVMLError: Not Supported (code 3) #348
Comments
Please include the back trace, as I can't see which test failed. There is no issue though, some NVML operations are not supported and we just need to mark the tests as such. |
Here is more info. The exceptions seem to be the same for all failed tests:
Except of this one:
|
Ah, interesting, it's the use of NVML by the test runner that fails here. I'll have a look. |
Could you test #352? |
Sure, but I have to figure out how to do it. I'm new to Julia and git. |
This should work (you can do this in a fresh depot if you don't want to mess with your global environment):
|
Here are the results: |
Ah, just a typo. Should work now, try |
New results:
|
Thanks, I'll add that test to the list of possibly unsupported ones. Pushed directly to master, so you can |
There are still some issues: |
OK, marked that test as unsupported as well: fedf087. Looks like your NVML doesn't support very much. Note that these are trivial changes, it might be easier if you try yourself. |
I'm using Julia v1.5 with CUDA Toolkit V10.2.89 on Ubuntu 18.04 and GeForce GT 730 with Driver Version: 440.100.
I tried both the toolkit binary installation from Nvidia and binary artifacts pulled by CUDA.jl. In both cases CUDA package test fails:
with
NVMLError: Not Supported (code 3)
in all cases. The simple CUDA examples, e.g. vector addition, seem to work fine, though.The text was updated successfully, but these errors were encountered: