-
Notifications
You must be signed in to change notification settings - Fork 862
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
NCCL fails to build on mac #61
Comments
That usually means that your CUDA_HOME is not set properly. CUDA_HOME defaults to /usr/local/cuda ; you should set it to your CUDA installation path. Agreed, we could return a more explicit message. Let me know if that fixes the problem. |
Thanks @sjeaugey! I tried adding export CUDA_HOME=/usr/local/cuda; make and it seems to not work either. |
Setting CUDA_HOME to /usr/local/cuda should not change anything (it is already the default value). Given the output, it looks like "ls /usr/local/cuda/lib64/libcudart.so.*" fails. If you are sure the path is correct, then it may be a mac-specific issue. Be aware that you may be the first person trying to compile NCCL on mac, so you are in unknown territory here. |
I think the problem is that on mac the directory should be lib not lib64. |
Thanks guys for the pointers - I think I have a solution by fixing the Makefile, will send a PR. |
#62 filed. |
I am building the master version (at commit 2a974f5), with Mac OS 10.12 and got the following error:
@slayton58 recommended me opening an issue - happy to provide more details :)
My build environment is nvcc 8.0.54 and Apple LLVM version 8.0.0 (clang-800.0.42.1).
The text was updated successfully, but these errors were encountered: