Re: [HTCondor-users] matching gpu devicename

[Michael Di Domenico <mdidomenico4@xxxxxxxxx>]

On Tue, Jun 2, 2015 at 2:19 PM, John (TJ) Knoeller <johnkn@xxxxxxxxxxx> wrote:
> You should be getting back a CUDACapability attribute, something like this:
>
> CUDACapability=2.0
> CUDADeviceName="GeForce GTX 480"
> CUDADriverVersion=4.20
> CUDAECCEnabled=false
> CUDAGlobalMemoryMb=1536
> CUDARuntimeVersion=4.10
>
> if you pass the -extra flag, then you should also be getting attributes
> CUDAClockMhz, CUDAComputeUnits and CUDACoresPerCU
> like this
>
> CUDACapability=2.0
> CUDAClockMhz=1400.00
> CUDAComputeUnits=15
> CUDACoresPerCU=32
> CUDADeviceName="GeForce GTX 480"
> CUDADriverVersion=4.20
> CUDAECCEnabled=false
> CUDAGlobalMemoryMb=1536
> CUDARuntimeVersion=4.10
>
> If you are using the GPUs metaknob, you can set the -extra flag by setting
> GPU_DISCOVERY_EXTRA like this:
>
> GPU_DISCOVERY_EXTRA = -extra
> use FEATURE : GPUs

Thanks, i am getting the info now.  Turns out if condor_gpu_discovery
doesn't find the cuda libs it falls back to opencl.  it wasn't overly
clear that was what was happening, but it's straightened out now

it would be nice if there was a -cuda option like the -opencl option,
but if it doesn't find the cuda libs it bombs with some error.  that
would have at least helped me