Dear engineer,
I had a similar question as wlugmayr had before. And the workstation was equiped with Tesla K20C and CUDA 8.0.
I tried install SCIPION software, and the installation process seems to update my CUDA or something. And then the error exist when I run cryoSPARC. However, it still did not work after updating the pycuda.
Best wishes,
Ming
Hi Ming,
We’re actively looking into this nvcc error and will update you.
As for the account issue, could you email us the contents of cryosparc/run/meteor.log to feedback@structura.bio?
Thanks,
Suhail
hi ming,
we solved the nvcc error by explicitly adding the following to .bashrc (even when nvcc is already callable when you login in bash):
so find out where your nvcc is located (e.g. which nvcc) and add the directory to PATH:
export PATH=/opt/cryosparc/bin:/usr/local/cuda-8.0/bin:$PATH
export CFLAGS="-I/usr/local/cuda-8.0/targets/x86_64-linux/include"
export LDFLAGS="-L/usr/local/cuda-8.0/targets/x86_64-linux/lib"
cheers,
wolfgang
hi ming,
i forgot. we also have scipion on this machine. but we did not put it into the environment - we just made an alias
e.g. alias scipion=’/opt/scipion-1.1/scipion’
this is enough to run scipion without problems.
cheers,
wolfgang
Dear Wolfgang,
Thank you very much for your kindly help and information about debug of my cryosparc running. I used the method you suggested but the problem still there even after I re-installled the whole cryosparc. I noticed that the installation is a litter different from my first installation of cryosparc in this machine. I attached the difference part of the installation recording.
I know that you are very busy recently. So, please have a look into my issue only when you have time. Thank you for your time!
Best wishes,
Ming
Dear Wolfgang and Ali,
My Nvcc processing problem has been solved by re-installation of CUDA-toolkit 8.0. I it should be exactly as you mentioned that the multiple version of CUDA is the major reason for this issue.
Thank you so much for your time!
Best wishes,
Ming
