FIX:CUDA on Debian Jessie

Andrew Bolster

Data Scientist at Sensum Co., Founder/Director at Farset Labs

Hopefully a super quick one (while I’m procrastinating from procrastinating).

Debian Jessie is a lovely operating system until you try and do anything with it. Lots of Package deprecations etc etc etc.

Anyway, I’ve got a history with GPU stuff and I’ve been playing with integrating it into some of my research, but in a bout of insanity I decided while I’m over in Liverpool (for another 4 hours) to wipe my old workstation and bring it over from Ubuntu 15.04 to Jessie (which I’ve been using on my main laptop for a while now).

This was painful. Basically, NVIDIA somewhere along the line deprecated the whole Debian base-family (even if it’s derivitives are supported). But we can rebuild it, we have the technology!

The Problem

[email protected] /usr/local/cuda/samples/2_Graphics/Mandelbrot  
➜   make                                                                                                            
>>> WARNING - libGL.so not found, refer to CUDA Getting Started Guide for how to find and install them. <<<
>>> WARNING - libGLU.so not found, refer to CUDA Getting Started Guide for how to find and install them. <<<
>>> WARNING - libX11.so not found, refer to CUDA Getting Started Guide for how to find and install them. <<<
[@] /usr/local/cuda-7.5/bin/nvcc -ccbin g++ -I../../common/inc -m64 -gencode arch=compute_20,code=sm_20 -gencode arch=compute_30,code
=sm_30 -gencode arch=compute_35,code=sm_35 -gencode arch=compute_37,code=sm_37 -gencode arch=compute_50,code=sm_50 -gencode arch=comp
ute_52,code=sm_52 -gencode arch=compute_52,code=compute_52 -o Mandelbrot.o -c Mandelbrot.cpp

Even though I do have these packages, and have updated my LD_LIBRARY_PATH, etc. etc.

The Guide

Extended from this brilliant answer on Stack Exchange by einpoklum

sudo apt-get install gcc g++ gcc-4.8 g++-4.8 gcc-4.9 g++-4.9 libxi6 libxi-dev libglu1-mesa libglu1-mesa-dev libxmu6 linux-headers-amd64 linux-source freeglut3-dev

Then

  • wget http://developer.download.nvidia.com/compute/cuda/7.5/Prod/local_installers/cuda_7.5.18_linux.run
  • chmod a+x cuda_*.run
  • Drop into a root shell
  • bash cuda_*.run

Yes yes yes and install the things wherever you like them. The problem comes when trying to build the sample projects.

The Fix

The problem is the findgllib.mk “helper” file use by all of the sample projects. Basically, with Debian, some key variables never get assigned, and theres no “catchall” fall-down case in the logic, so they end up blank which, well, screws things up.

YMMV but in my case, the location of the missing libraries was /usr/lib/x86_64-linux-gnu. This can be found using the locate` command

➜   locate libGL.so                                                                                                   
/usr/lib/libGL.so.1
/usr/lib/x86_64-linux-gnu/libGL.so
/usr/lib/x86_64-linux-gnu/libGL.so.1
/usr/lib/x86_64-linux-gnu/libGL.so.352.39
[email protected] ~/NVIDIA_CUDA-7.5_Samples/2_Graphics/Mandelbrot  
➜   locate libGLU.so                                                                                                  
/usr/lib/x86_64-linux-gnu/libGLU.so
/usr/lib/x86_64-linux-gnu/libGLU.so.1
/usr/lib/x86_64-linux-gnu/libGLU.so.1.3.1
[email protected] ~/NVIDIA_CUDA-7.5_Samples/2_Graphics/Mandelbrot  
➜   locate libX11.so                                                                                                  
/usr/lib/x86_64-linux-gnu/libX11.so
/usr/lib/x86_64-linux-gnu/libX11.so.6
/usr/lib/x86_64-linux-gnu/libX11.so.6.3.0

So, take your $LIBRARY_PATH and place it in a copy of the findgllib.mk file somewhere;

cp /usr/local/cuda/samples/common/findgllib.mk /tmp/

Edit it in your favourite editor to add the following lines before the commented line # find libGL, libGLU, libXi,

GLPATH ?= /usr/lib/x86_64-linux-gnu
GLLINK ?= -L/usr/lib/x86_64-linux-gnu
DFLT_PATH ?= /usr/lib

But as I said, there’s a version of this file for every sample project, which is a bit pointless IMO, but we can fix that easily enough by replacing them all

find /usr/local/cuda/samples/ -name "findgllib.mk" -exec cp findgllib.mk '{}' \;

And boom, all the samples should build.

Published: April 15 2016

  • tags:
blog comments powered by Disqus