-
Notifications
You must be signed in to change notification settings - Fork 254
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
ERROR: LoadError: ccall: could not find function cublasCreate_v2 in library #172
Comments
The issue was with my cudnn install - things where not linking at run time so I had to move my libraries around (to the cuda main folders) to allow Mocha to see them. |
How did you do this? I'm facing the same problems, but I have no idea how to go about fixing it. Which files need to be moved, and where do you move them to? |
I moved to MXNet. |
What version of the cuda libraries were you using, I use Cuda 7.5 libraries with CUDNN libraries v4. Also, some troubleshooting steps in this issue might be helpful to you: #199 |
I use Cuda 7.5 libraries with CUDNN libraries v4, also i face this problem, and how do you fix it? |
Check that the correct versions of cuda/cudnn are listed first in LD_LIBRARY_PATH. You might also search the code for LD_LIBRARY_PATH and see what it's searching for and verify on your system that it's there. I dealt with some similar issues when I got it working, however nowadays I'm using tensorflow, so I can't offer much more specific advice. |
Hi,
Pkg.test("Mocha") works.
But now I want to use my GPU backend; I set ENV["MOCHA_USE_CUDA"] = "true"
I find the device no issues
* CUDA enabled [DEV=0] (MOCHA_USE_CUDA environment variable detected)
But I get the error: ERROR: LoadError: ccall: could not find function cublasCreate_v2 in library from here: Mocha/src/cuda/cublas.jl:50
EDIT: after a couple days of Pkg.update() I now have the error:
ERROR: LoadError: ccall: could not find function cudnnCreate in library
Should I be linking libraries at runtime? how do I do that using Julia?
Thx
The text was updated successfully, but these errors were encountered: