Extraction Failure During SDK Installation

Accelerated Computing CUDA CUDA Setup and Installation
1

Running the SDK script as root results in the following error:
root@Downloads > ./cuda_5.0.35_linux_64_rhel6.x-1.run
Extraction failed.
Signal caught, cleaning up

I’ve tried the -verbose flag when running the script, but it produces no additional verbosity beyond what I pasted above.

My file systems definitely have enough space:
root@~ > df -h
Filesystem Size Used Avail Use% Mounted on
/dev/sda2 29G 15G 15G 51% /
tmpfs 2.9G 84K 2.9G 1% /dev/shm
/dev/sda1 485M 162M 299M 36% /boot
/dev/sda5 426G 199G 206G 50% /home
tmpfs 512M 1016K 512M 1% /tmp

I’m trying to install the SDK on a Dell Precision T5500 running CentOS 6.4. Some more info:

root@~ > lspci | grep VGA
03:00.0 VGA compatible controller: NVIDIA Corporation GT200b [GeForce GTX 285] (rev a1)

root@~ > uname -r
2.6.32-358.11.1.el6.x86_64

Anyone have any insight into why the extraction portion of the installer script fails? To that end, is there a list of system requirements anywhere for installing the SDK?

2

It looks like your /tmp directory is too small. The extracted installer will take more than 512MB.

3

I apologize for the delay. You were 100% correct. I modified /tmp so that it was no longer on a tmpfs. Thanks for the suggestion.