Comments (12)
I am going to put this down to a Jenkins configuration issue, since installing on the machine interactively works fine. Please reopen though if you find further information.
from cuda.
I got this error, too, with accelerate-cuda-0.15.0.0 from github/head.
from cuda.
However, when compiling with 'cabal configure --enable-executable-dynamic' the linker errors disappear.
from cuda.
@amigalemming @svenssonjoel what version of ghc are you using? I am wondering now if this is an issue introduced with 7.8, which started using dynamic linking.
from cuda.
(and I would like to say "does it happen with 7.6", but I don't think getting HEAD to work with 7.6 will be straightforward for you... I'll see if I can dig up a linux box to try it on)
from cuda.
This custom 'configure' stuff usually seems to be quite fragile. I have much better experience with Cabal's support of pkg-config. Unfortunately, not all packages support that, e.g. llvm. Currently I try to generate my own '.pc' file for 'llvm'. Maybe this approach is also reasonable for 'cuda'?
from cuda.
I can confirm that I still have the same problem. Just tried it fresh in a sandbox locally on my laptop with accelerate and accelerate-cuda from HEAD. Using GHC 7.8.2 (with the patch - not that it influences this).
from cuda.
@amigalemming I've never tried writing my own pgk-config scripts. Could you provide an example script demonstrating what you mean? To my naive eyes, it would just seem like shifting the problem somewhere else...?
@svenssonjoel okay, thanks for the update.
Since I can't reproduce this on my machine, I'm going to need you both to try and provide more information so we can narrow down the problem. Since there are two of you now experiencing this problem, hopefully the differential diagnosis will be easier! For a start, please grab the latest github version of the package and try installing with verbose options enabled, and just copy/paste that into a gist for me.
It might also be good to make sure you have the latest cabal
installed, and maybe a clean ~/.cabal/config
.
from cuda.
@tmcdonell The pkg-config pc files are simple configuration text files, no scripts, i.e. no running programs. Creating such configuration files certainly shifts the problem - in the long run, we can hopefully shift that to the CUDA maintainers. :-)
Here is an example for a pc file generator for LLVM, unfortunately only tested on Linux:
http://hackage.haskell.org/package/llvm-pkg-config
http://hackage.haskell.org/package/llvm-pkg-config-0.0/src/make-pkg-config.sh
http://llvm.org/bugs/show_bug.cgi?id=9405
from cuda.
To clarify, are you getting this on git HEAD?
from cuda.
@amigalemming I just noticed that my CUDA-6.0 install includes pkg-config files. I'm not sure if they were also there in 5.5, but they are not there with 5.0. It appears that they aren't installed into a default PKG_CONFIG_PATH, however. They also have names prefixed by version, which is a bit annoying (cuda-6.0.pc
instead of cuda.pc
). Anyway, not sure if this will be useful right now, but perhaps future versions will have better integration with pkg-config, and we can transition to that at that time.
from cuda.
I'm closing this because it's been quiet for so long, and the configuration process is now completely rewritten. Please open a new ticket if you still have problems!
from cuda.
Related Issues (20)
- Unknown CUDA device compute capability 6.1 HOT 2
- Minor edits needed for 7.10 HOT 1
- CUDA 8 support [was: Plans to support CUDA 8?] HOT 4
- Build fails with library profiling HOT 4
- Compilation fails for CUDA-8 [was: ghc 7.10.3 fail to install] HOT 2
- On Windows, the Cabal installer is looking in the wrong place. HOT 5
- Problem with module `Foreign.CUDA.Path' when installing on Windows 8.1 with CUDA-8.0 HOT 38
- Issues with dynamic parallelism? HOT 3
- GHCi only works with Stack with a custom cabal.buildinfo file. HOT 1
- Structure of nvidia-cuda-toolkit seems to have changed in Ubuntu 18.04 HOT 3
- Driver failure on AWS p3.2xlarge instance HOT 18
- Build failure in cuda-0.10.0.0 HOT 3
- Won't compile on Arch Linux HOT 3
- Ubuntu 16.04 + nvidia-cuda-toolkit "Found CUDA toolkit at: /usr" but 'Could not find path: ["/usr/lib64"]' HOT 7
- package revision to restrict Cabal to <3.0 HOT 1
- Compile issues for Cabal 2? HOT 2
- Cannot find a definition for `cuDevicePrimaryCtxRelease' in the header file. HOT 3
- CUDA 11.3 compatibility HOT 4
- rpath linking to cuda toolkit HOT 2
- This Library isn't Compatible with Windows HOT 4
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from cuda.