Needed: More Info About GPU Folders

[dicecca112]

I know that there is a least one person folding GPU WUs.

So list what your times are, what card your using.

Also would help what credit you did without the GPU and with.

[kenc51]

X1900XTX 512mb running 3D clocks (650/775)

6mins 19sec per frame | WU = p2733_GPUlambda
~10hrs per WU


Checking FAH server stats (link) shows all GPU WU's are worth 330pts
I don't know what my PPD was before folding, as I've only got this rig setup recently and Bubba has kinda sidetracked me (eat my dust btw)

See sig for more specs

:edited:

[dicecca112]

please add memory as well on the Vid Card

[Apoptosis]

they are 330 points and my X1900 completes a step every 6 minutes or so... I run rivatuner and manually run 3d clock speeds and now i get one step every 5 minutes.

[cadaveca]

Running one cpu instance:

XTX 512mb 756/900 5:16/frame(sweetspot on stock volts under water for my card )
XTX 512mb 648/774 5:46/frame
CrossfireXT 512mb 620/720 5:52/frame

XTX and Crossfire @ 690/765(craptastic mastercard!!!)running multi-gpu two instances, no cpu instance 5:58/frame AVG(times differ due to multi-gpu client only using one cpu for data shuffling, IMHO, no confirmation form Stanford yet)

XTX and Crossfire @ 690/765 running multi-gpu 2 instances, 1 cpu instance 6:15/frame to 7:45.frame, dependant on what unit the cpu is crunching.

XTX and crossfire @ 500/600(stock 2D) running 2 multi-gpu instances, and 1 cpu isntance 9:39 to 12:06/frame, once again dependant on cpu WU.








All tests done on D975XBX, 400FSB/1066 strap, Conroe @ 400x9, ram 1:1 4-4-3-4-28-6-8-5-4-6(memset timings). I'll fill in missing configs and such as time allows.

[Jeff_Grant]

x1600 Pro AGP 575 core/805 mem. 21min a frame
x1900 GT PCIe 540 core /1315 mem 11min a frame

[werewolfdaddy]

When is nVidia gonna get off their duff and help F@H work out the bugs with the 8800gtx's?

[dicecca112]

GPU folding with ATI was years in the works, so its gonna take a while. Vijay is more concerned on working the bugs out of the current client before embarking on a whole new client with all its issues.