Folding Forum

I'm folding two of these now -

GTX660Ti 1344 cores at boosted clock 1241 MHz (stock behavior for this EVGA FTW card): Project: 7623 (Run 570, Clone 0, Gen 50) 10 frames so far - 5:09 TPF; 14093 credit; Est 38902 ppd (never seen one of these on this card before - but seems over 2x ppd from what I've seen on P8054s)
GT430 - 96 cores at core OC to 900 MHz and Memory OC to 750 MHz: Project: 7623 (Run 567, Clone 1, Gen 4) Based on only 1 frame - 30:50 TPF; 14093 credit; Est 6661 ppd (vs 2448 ppd before rebaselining)

[Edit] I noticed that I was also folding a 7625 (174, 1, 3) on my GTX560M laptop (192 cores - underclocked to 700 MHz): The FAH GUI shows 5187 points 3084 ppd; HFM.net shows updated values - 18:18 TPF - 14093 credit; Est 11089 ppd - after 18 frames.

Two hours ago I got p7624 (495, 1, 0). I rebooted the client V7.2.9. Still shows the old data. Can I do something to show the new credit?

GTX680SC+ Project 7623 (Run 489, Clone 2, Gen 0) TPF 4 minutes 30 seconds 45097 PPD.

GreyWhiskers wrote:I'm folding two of these now -

GTX660Ti 1344 cores at boosted clock 1241 MHz (stock behavior for this EVGA FTW card): Project: 7623 (Run 570, Clone 0, Gen 50) 10 frames so far - 5:09 TPF; 14093 credit; Est 38902 ppd (never seen one of these on this card before - but seems over 2x ppd from what I've seen on P8054s)

Wow. I have needed some arm-twisting to get a 660 Ti, and this may be it.

I just got a 7623 (R634,C0,G66) on my 560Ti fps stock, and wow they run hot., @63% fan they're running 96C!

tpf is 7min 15sec for a ppd of 28000 roughly.

Oh, i'm not on adv flag either btw.

JimF wrote:

GreyWhiskers wrote:I'm folding two of these now -

GTX660Ti 1344 cores at boosted clock 1241 MHz (stock behavior for this EVGA FTW card): Project: 7623 (Run 570, Clone 0, Gen 50) 10 frames so far - 5:09 TPF; 14093 credit; Est 38902 ppd (never seen one of these on this card before - but seems over 2x ppd from what I've seen on P8054s)

Wow. I have needed some arm-twisting to get a 660 Ti, and this may be it.

If i had a chance to rebuy my 680 over again I'd have bought one with 2 x 8 wire power instead of of the 2 x 6. I had to do a lot of tweaks to get my system to be stable under some games (borderlands 2 was the one I could test stability with most easily, full system restart, and nothing wrong with PSU or other components) I had to disable some of the power saving features in my I7 3770 (c3 and c6), and several other voltage tweaks (slightly overvolted 4 x 8gb 1600 ram and vcssa), and now it's rock solid. I have a gut feeling that if the 8 pin power connectors were there instead of 6 pin, it wouldn't have those issues. I dropped about $2500 on my latest build, and I had to RMA parts because of instability. Definitely not very happy with how touchy things have gotten with both ivy bridge and the keppler cards.

dawhippersnapper wrote:I dropped about $2500 on my latest build, and I had to RMA parts because of instability. Definitely not very happy with how touchy things have gotten with both ivy bridge and the keppler cards.

Thus far I have not had any problems with Ivy Bridge (i5-3550) on Folding. But that is probably because it is a dedicated machine, and I don't do any games with it. A GTX 660 Ti should consume no more power than a GTX 560, so it is just a question of whether it is worth the upgrade. Thus far, the answer has been a resounding NO, but with a few more work units like this, the picture could change. I am really wondering about Gromacs 4.6 though, but we won't know the answer to that for some time.

GreyWhiskers wrote:[Edit] I noticed that I was also folding a 7625 (174, 1, 3) on my GTX560M laptop (192 cores - underclocked to 700 MHz): The FAH GUI shows 5187 points 3084 ppd; HFM.net shows updated values - 18:18 TPF - 14093 credit; Est 11089 ppd - after 18 frames.

There is something strange going on with the credits. On P7624, HFM.net shows my GTX 560 is getting 8651 PPD, while the FAH GUI shows 23,461. I think I will wait until the real credits stand up.

The credits from FAH are correct. I don't think HFM.net has the latest calculations. Here is what the server spit out at me.

16:19:55:WU02:FS01:Upload complete
16:19:55:WU02:FS01:Server responded WORK_ACK (400)
16:19:55:WU02:FS01:Final credit estimate, 14093.00 points

JimF wrote:

GreyWhiskers wrote:[Edit] I noticed that I was also folding a 7625 (174, 1, 3) on my GTX560M laptop (192 cores - underclocked to 700 MHz): The FAH GUI shows 5187 points 3084 ppd; HFM.net shows updated values - 18:18 TPF - 14093 credit; Est 11089 ppd - after 18 frames.

There is something strange going on with the credits. On P7624, HFM.net shows my GTX 560 is getting 8651 PPD, while the FAH GUI shows 23,461. I think I will wait until the real credits stand up.

Did you remember to tell HFM to Download Projects from Stanford? It won't update its knowledge of the point/deadline figures until you do.

Zagen30 wrote:Did you remember to tell HFM to Download Projects from Stanford? It won't update its knowledge of the point/deadline figures until you do.

Thanks, I had not done that. It fixed it right up. (Fortunately, it went in the right direction.)

I was assigned a WU from Project 7626, and noticed I went up to 5300 PPD from about 3800 before, although I noticed that the card had do underclock itself by half (700 to 350 MHz) in brief spurts because the heat was between 99-100 degrees Celsius. I underclocked to 600 and haven't had problems as temperatures are down to 97 and the GPU clock is steady. I thought I'd mention it as it may be too strenuous on certain cards, more so than 803x was even.

It's a good idea to manage the heat generated so that the most strenuous WU you might get doesn't cause anything to overheat. (Maybe 7626 is a candidate for "most strenuous" on your GPU.)

Downclocking is one reasonable approach to overheating. There are other options that might help. If your fan isn't getting to 100% (and if the noise generated at 100% is not considered "too loud" in your environment) it may be time to adjust your fan profile so that the fan does a better job of removing the heat when the temperature goes up.

If you have a 1-slot GPU that just dumps the heat inside your case, you can also look for better ways to get the heat out of the case. If you have a 2-slot GPU what exhausts the hot air directly out of the back of the case, this probably doesn't apply to you.

Just as a side note, there are two slot GPU's that still exhaust air inside the case...

All this heat has me pondering.........

Have we reached an upper platuae for gpuy folding because of hardwares limited ability to cool itself?

I'm running stock settings on mny 560 ti and at 70% fan I'm barely able to keep it under 90C, so if this trend continues, WU are out pacing hardware now.