-
Notifications
You must be signed in to change notification settings - Fork 218
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
OutOfGPUMemoryError With Available Memory #1346
Comments
Same behavior is also experienced for v1.7
|
Not really something we can do much about. |
If my memory was actually that full would
|
I also ran Below is a terminal call assessing my available memory on GPU.
I'm not certain what other tests I can use to determine GPU usage at runtime, or what is using all of the memory. As far as I can tell I should have enough memory for these functions to run (assuming they do not need over 7 GiB of memory). So if anything seems missing, or you'd like more machine information just ask. |
I've never seen glxinfo used for checking available device memory before... |
Here is the output from
|
I was able to run a glmark2 test (results show below). This shows that the GPU is operational and has enough available memory to run the test.
|
Also was able to use the GPU with a TesorFlow tutorial. |
Could this be a cause for the
|
Apparently this is an unsolved NVIDIA issue where, currently, the only fix is to reset the GPU. From stm:
In my case the recommended code could not be completed because I only have 1 GPU device. In my case I just rebooted the computer. |
Describe the bug
I get an OutOfGPUMemoryError by running
CUDA.memory_status()
.To reproduce
The Minimal Working Example (MWE) for this bug:
Manifest.toml
Expected behavior
I expected to get a status report on the memory of the GPU.
Version info
Details on Julia:
Details on CUDA:
Additional context
Honestly I doubt this is something that has flown underneath CUDA's radar and is probably machine spacific, but I'm at a loss as to what is going on.
The text was updated successfully, but these errors were encountered: