Richard Graham and aLTeReGo had quite a brilliant exchange on the subject but both are quite ahead of the curve for me, i.e. most is over my head. Would anyone be willing to point me in the right direction in terms what's required to get it going?
We just deployed three Telestream Lightspeed servers with Tesla K20 cards as part of our transcoding farm, and need to see how busy they get in terms of GPU utilization. Richard Graham, looks like, got it going using (1) NVidia WMI service, (2) Windows performance counters, (3) Component Monitor Wizard, with the help of WMI Explorer and Powershell.
Is there a list of steps required to get GPU metrics in SW?
Thanks!
