GPU Folding killing my PPD
ardichoke
Icrontian
Having problems with my GPU folding on my newly upgraded rig. My PhenomII X4 is folding like a champ. HFM.NET estimates PPD output between 3k and 5k for the SMP client. The estimate seems a bit on the high side but still, for the sake of this problem we will assume it's precise if not accurate. I set up GPU folding for my HD5770, any time I have both SMP and GPU going at the same time though, my estimated PPD for the SMP client drops to around 500-1k range, the GPU estimate is around 2.5K. Is this a known issue when folding on the HD5000 series? Anyone know of a workaround to keep the GPU client from killing my SMP clients performance?
0
Comments
Hooray.
I reduced the priority on the SMP client and knocked it back to 75% usage, and now I'm making more PPD with both clients together than I would with just the GPU. But not by much. It might be a bigger difference if I was running a quad.
This was after I had already tweaked environment variables for the GPU2 client. The CPU usage is 2-8%, and it was still interacting with the SMP client poorly.