Folding Forum

Hello everyone, sorry if this post is not on the best place or already answered but i just joined the forums

So my problem is, i have a 680 and a 560 TI folding on the 7.1.52 version.

I have copied the GPUs.txt and all and my 680 kinda folds and so does my 560 but it shows:

gpu:0:GK104 ( Geforce GTX680 ) [which is normal]
gpu:1:GF114 (Geforce GTX680) [which is wrong since that should be my 560]

So if anyone has a fix for this, cause i dont know if the 560 is getting the proper WU that it should or whatever.

Thanks for your time

From the log in your other post, the V7 client sees your 2 GPUs like this...

16:18:29: GPU 0: FERMI:1 GK104 [Geforce GTX680]
16:18:29: GPU 1: FERMI:1 GF114 [GeForce GTX 680]
16:18:29: GPU 2: FERMI:1 GF114 [GeForce GTX 560 Ti]

We have a bug ticket open to get the 680 double entry resolved.

And is it normal that both of them get the same WU? ( 8010 )

Since in order for the 680 to fold i had to add the flag "client-type = beta" which im guessing gives it beta WU? ( i am probably wrong on this )

Yes, normal. 680 is so new, nothing has been optimized for it yet. It will fold like a previous gen NV card until that happens.

Thanks again for all your help!!

Last question, which has probably been asked a million times and i dont know if you know though, any date for when the 680 will be optimized? Im guessing itll be pretty good for folding.

@7im
Next to last gen NV card!

@Theo1993
It might be that you ´ll get some problems to manage the slots correctly in the F@ Client Control. At least I ran into this problem with a GTX680 and a GTX460 in the same system. I removed the GTX460 to get rid of the "slot machine" that randomly choosed a slot to start or stop regardless of which slot I marked before I gave the command.

Heiko

@heikosch
Will see how it goes after a full day folding tomorrow, since i only put the 560 couple of hours ago, but so far i havent gotten any issues which prevent it from folding, so im hoping everything works well!

Please watch the PPD of the slots. It seems that sometimes - still unclear whether this is related to tools like EVGA precision X - the gpu clock of the GTX680 decreases.
I didn´t observe this with only the GTX680 but have seen this with a GTX680 and a GTX460 in the same system.

PPD for GTX680 should be only 15.000-16.000 (2 * half if you use both "slots")
PPD for GTX560 Ti should be 18.000 (with OC up to 24.000)

Heiko

heikosch wrote:@7im
Next to last gen NV card!

Heiko

So you were getting 20K PPD from your 4xx GPUs? This guy was getting 20K from his 680... http://www.evga.com/forums/fb.ashx?m=1572957

Like I said, I have seen many posts where the 680 folds like a 560.

7im wrote:

heikosch wrote:@7im
Next to last gen NV card!

Heiko

So you were getting 20K PPD from your 4xx GPUs? This guy was getting 20K from his 680...

Like I said, I have seen many posts where the 680 folds like a 560.

Well since i have both i can realistically say that the 560 for me gets around 15,500 PPD on average and the 680 around 16k, and just for a bonus, my AMD 1090T around 12k ( on the 8042 WU ) so im getting a nice 42,000 PPD

(Although i wont really run the 680 and the 560 at the same time for power consumption reasons and for heat reasons( case becomes an oven lol ))

7im wrote:

heikosch wrote:@7im
Next to last gen NV card!

So you were getting 20K PPD from your 4xx GPUs? This guy was getting 20K from his 680... http://www.evga.com/forums/fb.ashx?m=1572957

Like I said, I have seen many posts where the 680 folds like a 560.

My GTX460 actually gets 15.5k PPD.

These guys with the 18k or 20k PPD use EVGA GTX Hydro Copper or other factory OC versions! At stock clock you only get 15.5k to 16k PPD.
And kram36 (the guy with 20k) mentions that he "was getting" 20k "prior to the 2510 point wu´s" - probably project 8010 WUs.

And we should compare the single GPU flagship GTX680 to a GTX580 or GTX480 and not the GTX460 or GTX560. And from that perspective the GTX680 is folding slow.

Heiko