Cuda 9.0 Device Query failure Win10 Ubuntu 16.04LTS

Accelerated Computing CUDA CUDA Setup and Installation
#1

So im trying to use Tensorflow-GPU, and inorder to use it i have to use Cuda 9.0, as to my current knowledge there isnt a way to get it to work on Cuda 9.1 or 9.2

lolbozo@Jim-Jam:/usr/local/cuda/samples/bin/x86_64/linux/release$ ./deviceQuery
./deviceQuery Starting...

 CUDA Device Query (Runtime API) version (CUDART static linking)

modprobe: ERROR: ../libkmod/libkmod.c:514 lookup_builtin_file() could not open builtin file '/lib/modules/4.4.0-17134-Microsoft/modules.builtin.bin'
modprobe: ERROR: ../libkmod/libkmod.c:586 kmod_search_moddep() could not open moddep file '/lib/modules/4.4.0-17134-Microsoft/modules.dep.bin'
modprobe: ERROR: ../libkmod/libkmod.c:586 kmod_search_moddep() could not open moddep file '/lib/modules/4.4.0-17134-Microsoft/modules.dep.bin'
modprobe: ERROR: ../libkmod/libkmod-module.c:832 kmod_module_insert_module() could not find module by name='nvidia_384_uvm'
modprobe: ERROR: could not insert 'nvidia_384_uvm': Unknown symbol in module, or unknown parameter (see dmesg)
cudaGetDeviceCount returned 30
-> unknown error
Result = FAIL


lolbozo@Jim-Jam:/usr/local/cuda/samples/bin/x86_64/linux/release$ sudo ./deviceQuery
./deviceQuery Starting...

 CUDA Device Query (Runtime API) version (CUDART static linking)

cudaGetDeviceCount returned 30
-> unknown error
Result = FAIL

The only 2 solutions I’ve been able to find are to run with sudo, and as you see that didn’t fix anything, or to do a full shutdown of my computer, which also didn’t help at all

#2

your GPU driver is not installed correctly

not sure why you mention win10 in the title. Are you running this in a vm windows, or something like that?

#3

How would i go about installing it correctly? my graphics card is the GeForce MX150
It says Win10 because im using the 16.04LTS version of Ubuntu that you can download through the windows store, but my OS is Win10

#4

If you’re running linux on top of win10, you won’t be able to use CUDA within that, I don’t believe. It needs to be a baremetal install of linux.

You would go about installing it correctly by reading the linux install guide, loading a fresh copy of a supported linux OS on your machine (not within win10, but on baremetal) and get your installers from http://www.nvidia.com/getcuda and follow the linux install guide carefully:

https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html

Installation on a laptop can have additional requirements, which may vary by laptop.

#5

Well, i can install other versions of cuda using this method (it defaults to 7.5 for some reason), but when i did that i had no problems

#6

You can install CUDA and compile programs. But if you attempt to run those compiled programs, you’re going to hit a roadblock.