I’m trying to revive an older Santiago RHEL6.5 CUDA install that doesn’t seem to have been completed and would like some advice as to what the best upgrade path would be.
Currently, CUDA 5.5 appears to be installed:
$ /usr/local/cuda-5.5/bin/nvcc -V
nvcc: NVIDIA ® Cuda compiler driver
Copyright © 2005-2013 NVIDIA Corporation
Built on Wed_Jul_17_18:36:13_PDT_2013
Cuda compilation tools, release 5.5, V5.5.0
However, none of the samples were compiled. When I did so and tried to run deviceQuery, I get the following error:
CUDA Device Query (Driver API) statically linked version
FATAL: Module nvidia not found.
cuInit(0) returned 100
-> CUDA_ERROR_NO_DEVICE (no CUDA-capable devices were detected)
Result = FAIL
Although this might not be necessary, I was thinking that I should simply remove v5.5 and do a fresh install of v8.0, but I’m not sure how to properly uninstall v5.5.
It doesn’t seem to have been installed locally because I do not have the
executable and I don’t think it was installed via YUM because, when I query the YUM database, I get only the following package listed:
$ yum list installed “cuda”
cuda-repo-rhel6.x86_64 8.0.61-1 installed
(which I find odd… did one of my prior yum updates update this file without me knowing?)
I have downloaded cuda-repo-rhel6-8.0.61-1.x86_64.rpm but I’m waiting on a clean uninstall before I “rpm -i”
Any hints are greatly appreciated. Much thanks.
Edit: Wrong RHEL sub-version.