[RC5] installing CUDA client on Mac OS

Jeff Lawson bovine at distributed.net
Wed Jun 8 11:25:34 EDT 2011


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.

Unfortunately, there are no Stream API libraries on Mac so the current
cores cannot be directly compiled there.  Apple is instead
recommending that people rewrite their Stream code to use OpenCL
instead, and unfortunately we don't have anyone with the time and
experience to do that yet.  If you are able to, you are welcome to
help us by working on this.  The reference source code at
http://www.distributed.net/Source can be used as a starting point.



On Tue, Jun 7, 2011 at 5:51 PM, Andrew Hime <andrewhime at verizon.net> wrote:
> I have come into possession of a Mac recently, and I'm looking to
> install the CUDA client on it, but it is constantly giving me errors
> about libcudart.dylib.
>
> I've read the text file that comes with. I've read some things on Google
> about DYLD_something_PATH.
>
> What's the easiest way to fix this?
>
> Also, I notice there's no Stream client for Mac OS, which there should
> be, since the latest Macs come with Radeons now...
> _______________________________________________
> rc5 mailing list
> rc5 at lists.distributed.net
> http://lists.distributed.net/mailman/listinfo/rc5
>
>


More information about the rc5 mailing list