Can't locate InstallUtils.pm in @INC

This is like the 10th time I have downloaded Cuda 8.0 Toolkit and tried to install it. It fails every time. I say yes for everything except the nvidia driver and install everything in the default location.

FULL LOG :

Using more to view the EULA.

Installing the CUDA Toolkit in /usr/local/cuda-8.0 ...
Verifying archive integrity... All good.

Uncompressing NVIDIA CUDA................................................................................................

Can't locate InstallUtils.pm in @INC (you may need to install the InstallUtils module) (@INC contains: /etc/perl /usr/local/lib/x86_64-linux-gnu/perl/5.24.1 /usr/local/share/perl/5.24.1 /usr/lib/x86_64-linux-gnu/perl5/5.24 /usr/share/perl5 /usr/lib/x86_64-linux-gnu/perl/5.24 /usr/share/perl/5.24 /usr/local/lib/site_perl /usr/lib/x86_64-linux-gnu/perl-base) at ./install-linux.pl line 6.

BEGIN failed--compilation aborted at ./install-linux.pl line 6.

Verifying archive integrity... All good.

Uncompressing NVIDIA CUDA Samples................................................................................................

Can't locate InstallUtils.pm in @INC (you may need to install the InstallUtils module) (@INC contains: /etc/perl /usr/local/lib/x86_64-linux-gnu/perl/5.24.1 /usr/local/share/perl/5.24.1 /usr/lib/x86_64-linux-gnu/perl5/5.24 /usr/share/perl5 /usr/lib/x86_64-linux-gnu/perl/5.24 /usr/share/perl/5.24 /usr/local/lib/site_perl /usr/lib/x86_64-linux-gnu/perl-base) at ./install-sdk-linux.pl line 6.

BEGIN failed--compilation aborted at ./install-sdk-linux.pl line 6.

'uninstall_cuda_8.0.pl' -> '/usr/local/cuda-8.0/bin/uninstall_cuda_8.0.pl'

Installing the CUDA Samples in /home/dollarakshay ...

===========
= Summary =
===========

Driver:   Not Selected
Toolkit:  Installation Failed
Samples:  Installation Failed

Logfile is /tmp/cuda_install_6522.log

Please help.

What linux distro is this?

Ubuntu 16.04

maybe try setting a perl environment variable as mentioned here:

https://devtalk.nvidia.com/default/topic/968656/cuda-8-0-on-debian/

@txbob Yup Tried that before posting this. Tried it again just in case. Still dint work

I have printed the value of $PERL5LIB

Screenshot : https://i.imgur.com/cg71Ujd.png

These are the Nvidia and CUDA Packages that I have installed. The nvidia driver was installed using

sudo apt-get install nvidia-370

Screenshot : https://i.imgur.com/VegkdSM.png

Okay i just noticed that whatever I do to PERL5LIB it dosent change at all when it prints the log file, but I can actually see that it has changed in perl

Hello! i have same problem and decide theese next decisions:
unfortunately try export PERL5LIB and see in log whatever installer not seem file InstallUtils.pm

  1. unpack .run file ./cuda*.run --tar mxvf
  2. copy InstallUtils.pm to /usr/lib/x86_64-linux-gnu/perl-base
  3. export $PERL5LIB
    enjoy

Thank you soo much. Really you have solved a problem that I had for months. I even gave up and moved to windows. I have spend over 250+ hours over a period of 3 months trying to figure out a way around this solution.

Your solution works flawlessly. I really cant thank you enough. Thanks to you I can finally work on Tensorflow GPU in peace.

<3 Thank You

1 Like

It works like a charm! Thanks a lot.

OMG. OMG. OMG
You are the REAL MVP.
Lost dozens of hours on it too. Finally a solution, after so many tries, in different ways…

Total awesomeness! I’ve been trying to get the run-file working on Kubuntu 17.10 beta for the last 3 hours - your post solved my problem!

Thank you so much!!!

For the benefit of Fedora users (I’m working on Fedora 27 with GCC 5.4.0), these commands are:

sh ./cuda_8.0.61_375.26_linux.run --tar mxvf
sudo cp InstallUtils.pm /usr/lib64/perl5/
export $PERL5LIB
sh ./cuda_8.0.61_375.26_linux.run
sudo rm /usr/lib64/perl5/InstallUtils.pm

I had this problem with Cuda 9.2 on an Ubuntu 16.04 system, and had to run:

apt -y install libmodule-install-perl

To install the perl install modules.

I am curious…how were you able to dual install gcc5 on Fedora 27? Did you have to build this yourself with an alternative linker?