EXTREMEHW First Annual 96 Hour Folding Challenge, March 17th-20th 00:00 UTC

[The Pook]

M2000 isn't happy, do "particle coordinate is nan" errors = unstable GPU? gets to about 10% and then starts over 

 

flashed a more conservative BIOS on it, hopefully it's fixed. 

  

9 minutes ago, neurotix said:

Competition starts in 3 hours guys!

 

[damric]

1 minute ago, The Pook said:

M2000 isn't happy, do "particle coordinate is nan" errors = unstable GPU? gets to about 10% and then starts over 

 

flashed a more conservative BIOS on it, hopefully it's fixed. 

  

 

Yeah I have to run a bit lower clocks when folding. If it does it again, then you might have a corrupted work unit, which means you might have to uninstall, delete the data folder, and re-install. That would suck because then you lose the credit for the work unit, and it's happened to me before while trying to figure our what clock speed was stable while folding. Also don't bother with VRAM OC. It doesn't help much with folding and can cause problems like the one I just stated.

 

As long as you complete more than 80% of all work units you get the bonus points for quick return.

[elblat]

Exciting to see all the interest. Thanks for hosting this! Came here via the F@H forum post.

[neurotix]

27 minutes ago, damric said:

Yeah I have to run a bit lower clocks when folding. If it does it again, then you might have a corrupted work unit, which means you might have to uninstall, delete the data folder, and re-install. That would suck because then you lose the credit for the work unit, and it's happened to me before while trying to figure our what clock speed was stable while folding. Also don't bother with VRAM OC. It doesn't help much with folding and can cause problems like the one I just stated.

 

As long as you complete more than 80% of all work units you get the bonus points for quick return.

 

Wondering if you guys are overclocking the GPU memory or just the core. A long time ago, folding@home released a GPU memory tester program that works really well finding GPU memory errors. Its good for not just folding, but GPU memory testing in general.

 

https://foldingathome.org/2009/05/05/nvidia-gpu-memory-checker-update/?lng=en

 

There may be a newer version of it as well, dunno.

Edited March 16, 2023 by neurotix

[The Pook]

41 minutes ago, damric said:

Yeah I have to run a bit lower clocks when folding. If it does it again, then you might have a corrupted work unit, which means you might have to uninstall, delete the data folder, and re-install. That would suck because then you lose the credit for the work unit, and it's happened to me before while trying to figure our what clock speed was stable while folding. Also don't bother with VRAM OC. It doesn't help much with folding and can cause problems like the one I just stated.

 

As long as you complete more than 80% of all work units you get the bonus points for quick return.

 

thank friend! I think core clock was to blame, made a BIOS for 1455 instead of 1494 and it's ~20% now without any problems. 

 

guess my M2000 is even more poo than previously thought 

[damric]

7 minutes ago, The Pook said:

 

thank friend! I think core clock was to blame, made a BIOS for 1455 instead of 1494 and it's ~20% now without any problems. 

 

guess my M2000 is even more poo than previously thought 

Stock cooler? Run that fan at 100%

Edited March 16, 2023 by damric

[HaRdLy007]

IN Jeff_Petovello

[Monsterkater]

1 hour ago, The Pook said:

"particle coordinate is nan" errors = unstable GPU? gets to about 10% and then starts over

Yes, I had the same error with my overclock. Strangely, only project 18717 was displaying this behavior. Also, it would eventually finish after some checkpoint restarts, but this same project was always erroring out on the same percentages every time, like clockwork. All other projects were running fine just fine without any hiccups. I was already beginning to suspect a bad project or something, but then I reverted my overclock to stock speeds just to be sure et voilà, it pulled a WU from that project and folded start to finish without one single issue.

F@H just really is that sensitive to every little bit flip, I guess.

 

Luckily I caught it soon enough to be able to dial in a small, but rock stable OC anyway. It's not that much of a difference in PPD tbh (just barely enough to even out the loss from CPU folding), but I'll take what I can get 

[neurotix]

10 minutes ago, HaRdLy007 said:

IN Jeff_Petovello

Welcome to the site!

[Monsterkater]

Um, @damric? I couldn't help but notice that the banner at the top of the forum shows 1h less to start than the countdown at timeanddate.com ... So, which is it? 

[The Pook]

40 minutes ago, damric said:

Stock cooler?

 

ye, just new paste. temps shouldn't really be a problem though, it's mostly hovering in the high 60s. 

 

new BIOS seems to have fixed it, it's been chugging along happily now. 

 

  

12 minutes ago, Monsterkater said:

Um, @damric? I couldn't help but notice that the banner at the top of the forum shows 1h less to start than the countdown at timeanddate.com ... So, which is it? 

 

4 hour difference for me 

 

it's starting at 0:00 UTC (8:00 PM EST), so I'm assuming the timeanddate.com one is right

Edited March 16, 2023 by The Pook

[Monsterkater]

6 minutes ago, The Pook said:

4 hour difference for me 

Ah, so the banner is counting down to midnight in the respective user's time zone. Figures.

[damric]

36 minutes ago, The Pook said:

 

ye, just new paste. temps shouldn't really be a problem though, it's mostly hovering in the high 60s. 

 

new BIOS seems to have fixed it, it's been chugging along happily now. 

 

  

 

4 hour difference for me 

 

it's starting at 0:00 UTC (8:00 PM EST), so I'm assuming the timeanddate.com one is right

I'm looking into this now. That banner just appeared so I got to ask ENTERPRISE.

 

The Start time is 00:00 UTC so ignore that banner. We have about 52 minutes to start time. Did that thing just pop up?

Edited March 16, 2023 by damric

[Monsterkater]

6 minutes ago, damric said:

Did that thing just pop up?

Yes, it did. I refreshed the page and without warning it just hit me 

[damric]

6 minutes ago, Monsterkater said:

Yes, it did. I refreshed the page and without warning it just hit me 

The site dev made an oppsie. He'll fix that ASAP. He did get me a better looking embedded score card though. I'm going to have both score cards for a few hours, then delete the old one if the new one seems to be working fine

 

IGNORE THAT TIME BANNER FOR NOW

 

On top of that my main machine restarted so I have to fix that too

Edited March 16, 2023 by damric

[Monsterkater]

Alright ppl, I just submitted my last WU for LTT and switched teams.

 

Not even 30 minutes to go, so it's time for me to say:
Good luck to everyone in the competition, may the hardware gods be ever in your favor and let your devices survive the stress!
Here's to hoping the assignment servers are nice to y'all, pushing favorable WUs.

It's past my bedtime already, so I'm out at least for today.

Happy folding!

[mx500torid]

Is anyone folding with a 6800 or 6900 radeon? Seems my ppd has dropped 2 million from the older driver.

 

[damric]

Just now, mx500torid said:

Is anyone folding with a 6800 or 6900 radeon? Seems my ppd has dropped 2 million from the older driver.

 

6900XT here. I normally fold at 100W but I just unleashed it to full power for this contest.

 

The points could just be the work units.

[tictoc]

6 minutes ago, mx500torid said:

Is anyone folding with a 6800 or 6900 radeon? Seems my ppd has dropped 2 million from the older driver.

 

 

I am folding on a 6900 XT, but not in Windows. For me there is pretty huge range in ppd on various WUs. I see anywhere from 2.7M-6.4M ppd depending on the project that is running.

 

*Edit* Currently folding at 2875core|2150mem

Edited March 16, 2023 by tictoc

[damric]

I notice my GPU utilization in Windows is only 77% so maybe this new driver is just crap.

[mx500torid]

I was consistant 4.5 million on 22.7.1. Using the newest driver and I do believe it is crap. Will change back after this wu and see.

[damric]

Just now, mx500torid said:

I was consistant 4.5 million on 22.7.1. Using the newest driver and I do believe it is crap. Will change back after this wu and see.

I was going to text you a reminder about the contest starting soon. Glad you made it

[mx500torid]

tictoc nailed it. Dumped the first wu at 2.4 million and second wu jumped to 4.8. Feeling better now.

[The Pook]

M2000 failed another WU at ~60% this time. flashed it back to the stock BIOS and got a 18717 WU that @Monsterkatermentioned and it says it's going to take almost two days 

 

inb4 failure @ 95% 

[damric]

7 minutes ago, The Pook said:

M2000 failed another WU at ~60% this time. flashed it back to the stock BIOS and got a 18717 WU that @Monsterkatermentioned and it says it's going to take almost two days 

 

inb4 failure @ 95% 

Do you have a descent multicore CPU?

 

Edit if I can remember I'll send you my firmware image for m2000 to compare to yours. I might be running different power limits and voltages.

Edited March 17, 2023 by damric