How to get 20k PPD without bigadv :p

Status
Not open for further replies.
Veeb0rg said:
I tried to install gpu3 on my g80 card, did a fresh client install and got the same results as above, it would download core15 and then nothing.
Click to expand...
I guess that pretty much confirms it then. What I saw was only the "no appropriate server was available" message along with other similar messages about not connecting being repeated at regular intervals, however nothing was downloaded, so I see now it's not the exact same issue. :(
 
I just testing it on one of my 9800gt´s,shader 1620,xp32,and it looks promising so far,fingers crossed.
24 % as is,estimated ppd in hfm.net is 6700,while that card normally are around 5300 with 353 wu´s.
Run it with oneunit so far.

If I understand it correctly it is pointless on gtx 260 but apparently gts 250 is a good candidate ?
Gonna try those gts 250 next,could be even more fun (higher speed on shaders on those :)
 
folding_rejects said:
If I understand it correctly it is pointless on gtx 260 but apparently gts 250 is a good candidate ?
Gonna try those gts 250 next,could be even more fun (higher speed on shaders on those :)
Click to expand...
250 should be fine. Only way of knowing for sure is to try it out but I'm certain it will work. Let us know.
 
Um, I have a P11257 WU that has gone over 100% completion and is not returning results, just continuing to process. I don't think I ever saw a WU go above 100%, LOL. The progress bar for that client in FahSpy has turned maroon in color, and I never saw that either. What should I do, wait till it reaches 200%...?! :eek:
 
are you going by what fahspy is telling you or have you checked the client/log?
 
Veeb0rg said:
are you going by what fahspy is telling you or have you checked the client/log?
Click to expand...
I'm looking right at the client. It's currently at 112%. If it wasn't so amusing I'd probably have deleted the WU. It's so freaky I think I'm going to let continue...:D
 
Hmm, I've been doing some more research and thinking on this and I don't know if I can support using the fermi flag on non-Fermi hardware. We all know that Core 15 WUs can run on non-Fermi hardware and there is specific Core 15 work available for both cards but most of the WUs people are getting by doing this are of the 10927-10978 and 11214-11265 variety.

These WUs were released by PG, restricted by the -forcegpu nvidia_fermi flag, to evaluate the performance of Core 15 on Fermi boards with Generalized Born (GB) model used as implicit solvent. It is obvious we can run these WUs on non-Fermi cards but we are not providing PG with the accurate performance data they are looking for by doing so. Therefore, although we are getting nice points, we are sending bad data back to Stanford.
 
Veeb0rg said:
I tried to install gpu3 on my g80 card, did a fresh client install and got the same results as above, it would download core15 and then nothing.
Click to expand...

I tried it last night on a 8800gts 640 and it would do the same thing.
 
Glad I saw this thread...didn't know I was supposed to be running -advmethods on my GTX 460, boosted my ppd by 2500k
 
Tobit said:
Hmm, I've been doing some more research and thinking on this and I don't know if I can support using the fermi flag on non-Fermi hardware. We all know that Core 15 WUs can run on non-Fermi hardware and there is specific Core 15 work available for both cards but most of the WUs people are getting by doing this are of the 10927-10978 and 11214-11265 variety.

These WUs were released by PG, restricted by the -forcegpu nvidia_fermi flag, to evaluate the performance of Core 15 on Fermi boards with Generalized Born (GB) model used as implicit solvent. It is obvious we can run these WUs on non-Fermi cards but we are not providing PG with the accurate performance data they are looking for by doing so. Therefore, although we are getting nice points, we are sending bad data back to Stanford.
Click to expand...

Good head's up and advice.
 
Thanks for the tip, I'm over 14k on my GTX460 now. Nice. :cool:
[Edit] Between a GT240 and GTX460 I'm getting 19,719 ppd.
 
Last edited:
Tobit said:
Hmm, I've been doing some more research and thinking on this and I don't know if I can support using the fermi flag on non-Fermi hardware. We all know that Core 15 WUs can run on non-Fermi hardware and there is specific Core 15 work available for both cards but most of the WUs people are getting by doing this are of the 10927-10978 and 11214-11265 variety.

These WUs were released by PG, restricted by the -forcegpu nvidia_fermi flag, to evaluate the performance of Core 15 on Fermi boards with Generalized Born (GB) model used as implicit solvent. It is obvious we can run these WUs on non-Fermi cards but we are not providing PG with the accurate performance data they are looking for by doing so. Therefore, although we are getting nice points, we are sending bad data back to Stanford.
Click to expand...

Thanks for the info Tobit.

I tried it and I gained about 2-3K PPD using a 9800GT with 1836Mhz shaders. As soon as this WU is done I'll switch back to my previous flag settings. Science and inaccurate results = problems!
 
Well,I have ran it all day now with gpu 3 on two gts 250 and one 9800 gt and it looks nice.
those gts 250 is two versions,one asus dark knight: shader 1890,and one gainward green: shader 1620,huge improvement so big thanks from me :D

foldingfermiswitch.jpg



on gpu 2 the 9800 gt usually got 5300 with 353 wu´s.
asus gts 250 was around 6600 with 353 wu.
gainward gts 250 in the higher of 5000, 5700 aroundish.

have shut down one pc but it is 9800 gt on that one too,didn´t bothered with gtx 260.
So again, thanks a bunch :)

edit: just wake up here, may switch back but it had been nice with someone in the know could say something,so I will check at the officials if there is anything there :?
 
Last edited:
until PG comes out and says to stop doing it (which they wont because its the PG way) ill continue to run the fermi flag.. who knows maybe it will actually force them to fix their dumb points system thats wishful thinking..
 
sirmonkey1985 said:
until PG comes out and says to stop doing it (which they wont because its the PG way) ill continue to run the fermi flag.. who knows maybe it will actually force them to fix their dumb points system thats wishful thinking..
Click to expand...

And you want them to officially mention more than once to draw attention to it so more can mess with the testing that they want to do??
 
I´m shutting my fermi switches down.

After looking at the officials the only thing I have seen there is none PG members talking about lifetime bans,zero out points and more in that style,and speculations that may be correct or not,still speculations until someone with accountability say something.
That means a PG member or people who speaks on their behalf.
To be on the safe side I quit using it anyway.
 
Status
Not open for further replies.
Back
Top