Unsupported Configurations with CUDA 5.5?

Accelerated Computing CUDA CUDA Setup and Installation
1

Hi,

I have a few different machines ranging from three K20s to two 2050s and two K10s to a single K20. On all of these setups, when I go to install CUDA 5.5 I receive the following error:

You are attempting to install on an unsupported configuration. Do you wish to continue? ((y)es/(n)o) [ default is no ]:

All of the machines currently have CUDA 5.0 installed on them. The operating systems vary from CentOS 5.9, CentOS 6.4., and RHEL 5.9. I am installing via the run file and not the rpm.

I also noticed that on the download page, RHEL 5 and RHEL 6 point to the same file as of today. However, this may be on purpose.

To install CUDA 5.5, I first bring the machine into runlevel 3. Next I execute the .run executable. Am I missing anything else?

Any idea why I am getting this error?

##Setup A##
OS: CentOS 6.4

Kernel: 2.6.32-358.11.1.el6.x86_64

Output of cat /proc/driver/nvidia/version:

NVRM version: NVIDIA UNIX x86_64 Kernel Module  304.54  Sat Sep 29 00:05:49 PDT 2012
GCC version:  gcc version 4.4.7 20120313 (Red Hat 4.4.7-3) (GCC)

Output of nvidia-smi:

Wed Aug  7 13:17:16 2013
+------------------------------------------------------+                       
| NVIDIA-SMI 4.304.54   Driver Version: 304.54         |                       
|-------------------------------+----------------------+----------------------+
| GPU  Name                     | Bus-Id        Disp.  | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap| Memory-Usage         | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  Tesla K20m               | 0000:02:00.0     Off |                    0 |
| N/A   29C    P8    25W / 225W |   0%   11MB / 4799MB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   1  Tesla K20m               | 0000:03:00.0     Off |                    0 |
| N/A   34C    P8    25W / 225W |   0%   11MB / 4799MB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   2  Tesla K20m               | 0000:84:00.0     Off |                    0 |
| N/A   31C    P8    25W / 225W |   0%   11MB / 4799MB |      0%      Default |
+-------------------------------+----------------------+----------------------+
                                                                               
+-----------------------------------------------------------------------------+
| Compute processes:                                               GPU Memory |
|  GPU       PID  Process name                                     Usage      |
|=============================================================================|
|  No running compute processes found                                         |
+-----------------------------------------------------------------------------+

##Setup B##
OS: CentOS 5.9

Kernel: 2.6.18-194.el5

Output of cat /proc/driver/nvidia/version:

NVRM version: NVIDIA UNIX x86_64 Kernel Module  304.54  Sat Sep 29 00:05:49 PDT 2012
GCC version:  gcc version 4.1.2 20080704 (Red Hat 4.1.2-48)

Output of nvidia-smi:

Wed Aug  7 13:23:59 2013       
+------------------------------------------------------+                       
| NVIDIA-SMI 4.304.54   Driver Version: 304.54         |                       
|-------------------------------+----------------------+----------------------+
| GPU  Name                     | Bus-Id        Disp.  | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap| Memory-Usage         | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  Tesla C2050              | 0000:02:00.0     Off |                    0 |
| 30%   35C   P12    N/A /  N/A |   0%    6MB / 2687MB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   1  Tesla K10.G2.8GB         | 0000:05:00.0     Off |                    0 |
| N/A   21C    P8    15W / 117W |   0%    9MB / 3583MB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   2  Tesla K10.G2.8GB         | 0000:06:00.0     Off |                    0 |
| N/A   19C    P8    15W / 117W |   0%    9MB / 3583MB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   3  Tesla C2050              | 0000:83:00.0     Off |                    0 |
| 30%   34C   P12    N/A /  N/A |   0%    6MB / 2687MB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   4  Tesla K10.G2.8GB         | 0000:86:00.0     Off |                    0 |
| N/A   20C    P8    15W / 117W |   0%    9MB / 3583MB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   5  Tesla K10.G2.8GB         | 0000:87:00.0     Off |                    0 |
| N/A   18C    P8    15W / 117W |   0%    9MB / 3583MB |      0%      Default |
+-------------------------------+----------------------+----------------------+
                                                                               
+-----------------------------------------------------------------------------+
| Compute processes:                                               GPU Memory |
|  GPU       PID  Process name                                     Usage      |
|=============================================================================|
|  No running compute processes found                                         |
+-----------------------------------------------------------------------------+

##Setup C##
Note: I went ahead an installed CUDA 5.5 anyway for testing on this system. Also the Quadro FX 4800 is not being used for CUDA, just to drive a monitor.

OS: Red Hat Enterprise Linux Server release 5.9

Kernel: 2.6.18-194.26.1.el5

Output of cat /proc/driver/nvidia/version:

NVRM version: NVIDIA UNIX x86_64 Kernel Module  319.37  Wed Jul  3 17:08:50 PDT 2013
GCC version:  gcc version 4.1.2 20080704 (Red Hat 4.1.2-54)

Output of nvidia-smi:

Wed Aug  7 13:25:41 2013       
+------------------------------------------------------+                       
| NVIDIA-SMI 5.319.37   Driver Version: 319.37         |                       
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  Quadro FX 4800      Off  | 0000:02:00.0     N/A |                  N/A |
| 35%   56C  N/A     N/A /  N/A |       17MB /  1535MB |     N/A      Default |
+-------------------------------+----------------------+----------------------+
|   1  Tesla K20c          Off  | 0000:03:00.0     Off |                    0 |
| 30%   36C    P8    14W / 225W |       13MB /  4799MB |      0%      Default |
+-------------------------------+----------------------+----------------------+
                                                                               
+-----------------------------------------------------------------------------+
| Compute processes:                                               GPU Memory |
|  GPU       PID  Process name                                     Usage      |
|=============================================================================|
|    0            Not Supported                                               |
+-----------------------------------------------------------------------------+
2

CentOS is not a supported distro for CUDA 5.5.

It looks like your RHEL 5.5 machine is 32-bit (I base this on your kernel). RHEL 5.5 x86_64 is supported, but not x86.

3

The RHEL 5.9 machine is 64 bit.

Linux hostname 2.6.18-194.26.1.el5 #1 SMP Fri Oct 29 14:21:16 EDT 2010 x86_64 x86_64 x86_64 GNU/Linux
4

Interesting, then it should be marked as supported.

Please post the output of the following:
cat /etc/issue
uname -i

5

/etc/issue is customized per company policy but /etc/issue.rpmnew is

Red Hat Enterprise Linux Server release 5.9 (Tikanga)
Kernel \r on an \m

Where \r is the 2.6.18-194.26.1.el5 and \m is x86_64 I am assuming.

uname -i:

x86_64

Could the alteration of /etc/issue be the problem?

6

I think that is indeed the issue Andy. All of our machines must have a specific /etc/issue due to policy. However, when I temporarily copied /etc/issue.rpmnew to /etc/issue, the error went away.

Is this a bug since /etc/issue is commonly customized?

7

Yes, I’ve filed a bug to track this to use another source (possibly lsb_release) before attempting to use /etc/issue.

8

Having a similar problem, I think with cuda 6.0… my first time installing or using Cuda:

$ ./cuda_6.0.37_linux_64.run
Do you accept the previously read EULA? (accept/decline/quit): accept
You are attempting to install on an unsupported configuration. Do you wish to continue? ((y)es/(n)o) [ default is no ]: no

$ cat /etc/ issue
Red Hat Enterprise Linux Client release 5.10 (Tikanga)
Kernel \r on an \m

$ lsb_release
LSB Version: :core-4.0-amd64:core-4.0-ia32:core-4.0-noarch:graphics-4.0-amd64:graphics-4.0-ia32:graphics-4.0-noarch:printing-4.0-amd64:printing-4.0-ia32:printing-4.0-noarch

$ /sbin/lspci | grep -i nvi
05:00.0 VGA compatible controller: NVIDIA Corporation GK107GL [Quadro K2000] (rev a1)

$ uname -i
x86_64

Any help would be appreciated.

9

The Linux platforms supported by CUDA 6.0 are listed here:

[url]http://docs.nvidia.com/cuda/cuda-getting-started-guide-for-linux/index.html[/url]

It seems only RHEL versions more recent than 5.1 are supported. However, that does not necessarily mean it will not work. “unsupported” just means it has not been tested, you are proceeding at your own risk, and it would not be considered a bug should it not work. That is why the installer gives you the choice of overriding the default choice of “no”.

[much later:]

I have a 64-bit Linux machine here running “Red Hat Enterprise Linux Client release 5.2 (Tikanga)” that seems to run CUDA 6.0 just fine, so you may want to consider using the override option provided by the installer.