Re: installing CUDA client on Mac OS

On Wed, Jun 8, 2011 at 12:00 PM, Andrew Hime <[email protected]> wrote:
> On 6/8/2011 10:25 AM, Jeff Lawson wrote:
>> The readme.cuda has the following text:
>>
>>      If you are unable to execute the dnetc binary because of a missing
>>      libcudart.so library, you have to set your LD_LIBRARY_PATH environment
>>      variable (on Linux) or DYLD_LIBRARY_PATH (on Mac OS X) to include
>>      the 'lib' subdirectory of the extracted client archive prior to
>>      running dnetc.
>
> A little help with that? It should be somewhat obvious from my post that
> I read the readme.cuda, but I'm coming off a 10+ year break from Unix
> environments, and never really set a bunch of environment variables.
>
> Thanks, though. You've inspired a blog post.


If you're using tcsh try something like:

setenv DYLD_LIBRARY_PATH /path/to/library/

If you're using bash/sh try:

export DYLD_LIBRARY_PATH=/path/to/library/

It might be easiest if you create a simple shell script that sets it
and then just runs the client.  I've been considering bundling such a
shell script with the mac client just to simplify these steps.
_______________________________________________
rc5 mailing list
[email protected]
http://lists.distributed.net/mailman/listinfo/rc5