[torquedev] GPU Hardware

Chris Reynolds C.Reynolds at westminster.ac.uk
Thu Oct 21 10:15:39 MDT 2010


Hi guys,
The software and hardware for our GPU cluster is as follows (for 10
nodes):

-Intel I7 (920 quad core) with 6GB 1600Mhz DDR3 RAM.
-2x Nvidia GTX 470 (Gainward) per node.
-OS: Debian (Lenny) Linux.

Restricting a job to use only one GPU might be difficult. An OpenCL or
Cuda capable binary may well attempt to get (through for example
clGetDevices) as many GPU's as are present on the node, and subsequently
try to schedule kernels across each device. To prevent this, maybe
LD_PRELOAD and an interception library could be used, i.e. to always
return one (free) device upon each query?

Cheers

Chris


On Thu, 2010-10-21 at 09:58 -0600, Ken Nielson wrote:
> We would like to know what kind of GPU hardware is deployed in your data 
> centers so we can better understand what we need to do to provide GPU 
> support in TORQUE.
> 
> If you currently have GPUs deployed would you please reply and let us 
> know the details of the hardware and any other information you think is 
> relevant.
> 
> Thanks
> 
> Ken Nielson
> Adaptive Computing
> _______________________________________________
> torquedev mailing list
> torquedev at supercluster.org
> http://www.supercluster.org/mailman/listinfo/torquedev


-- 
The University of Westminster is a charity and a company limited by
guarantee.  Registration number: 977818 England.  Registered Office:
309 Regent Street, London W1B 2UW, UK.


More information about the torquedev mailing list