Skip to content
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

(unhandled): CUDA Exception: invalid argument #80

Closed
rrnewton opened this issue Feb 27, 2013 · 4 comments
Closed

(unhandled): CUDA Exception: invalid argument #80

rrnewton opened this issue Feb 27, 2013 · 4 comments

Comments

@rrnewton
Copy link
Member

Hi there,

I just started using accelerate HEAD (accelerate 59e0257, accelerate-cuda d60e098baa1c492da39a2) on a new machine (Mac os) and immediately ran into some errors.

I was in the middle of modifying an nbody benchmark, here:

https://github.com/iu-parfunc/array-dsl-benchmarks/commit/fa9aa33c88c6c323b6926bea22adc678447f9473

First thing was an error on the non-existent ~/.accelerate dir:

naive_cuda.exe: /Users/rrnewton/.accelerate: canonicalizePath: does not exist (No such file or directory)

I created that directory by hand, but then I got this error:

naive_cuda.exe: 
*** Internal error in package accelerate-cuda ***
*** Please submit a bug report at https://github.com/AccelerateHS/accelerate/issues
./Data/Array/Accelerate/CUDA.hs:145 (unhandled): CUDA Exception: invalid argument
@tmcdonell
Copy link
Member

Huh, canonicalizePath doesn't throw an exception for me when the directory doesn't exist? I'll add a mkdir.

@tranma got some exceptions recently using ghc-7.6.2-x86_64. This might be the same problem; I'll take a look. (I'm still on 7.6.1 and haven't come across it, and I think we all have the same machine)

@tranma
Copy link

tranma commented Feb 27, 2013

I also simply created a directory for canonicalizePath, apparently the behaviour changed at some point - I don't think it matters.

However running the CUDA backend on anything (regression, manually constructed acc's in ghci) results in the same error as above. This happens on 7.6.* on OSX 10.8.

@tmcdonell
Copy link
Member

Finally tracked this down. Please update to the latest CUDA bindings package: http://hackage.haskell.org/package/cuda

@rrnewton
Copy link
Member Author

rrnewton commented Mar 4, 2013

Thanks!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants