• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

No.5 position may come back to haunt us

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

the_cultie

Member
Joined
Feb 3, 2005
I was just checking the folding stats and I noticed that Custom PC & bit-tech has appeared on our threat list. At the time of writing it said overtake in 2 years. So I went to look at their stats and I saw this:
CPCbt1.jpg


I was like :eek: so I investigated further to discover this belonging to a new member of there team called zz9pzza:
CPCbt2.jpg


In one day he produced just over 72% of what our team did in a day!!!

I don't know how long he will be folding at that rate for, but if he keeps it up we're going to be cleaning the number 4 plate for a new team. :cry: :cry:

sfu, the gauntlet has been thrown down, are you willing to take on the challange? :attn: :attn: :salute: :salute:
 
Oh ... crap.

Here's hoping its just a sys admin stress-testing some servers before they hit production. Ouch.
 
Looking closer, he's averaging 6,000 pts/WU, suggesting SMP rather than bigadv.

Looks like 1000+ cores running SMP.

One of his updates intrigues me though: 879 WUs for 1,210 points. A raft of dud WUs?
 
My attitude when folks like this show up is... let's see how long it lasts... but in the mean time, someone should go recruit this person and bring them over here. :)

This is quite impressive. I don't think I've ever seen ANYONE do 4.5 million in one day... and they're on a tear to do even more today. Whew... hope Stanford doesn't run out of WUs. :D
 
Dayum! That is very impressive! It only takes him only 10 minutes to get what I get in a day. I checked their forum and found a thread about him. Here's a quote from him.

So I am breaking in three racks, each rack has 4 chassis ,each chassis has 14 blades. Each blade is a dual processor machine with 6 cores per processor and 36gig of RAM. I will probably stop running the folding first thing Monday as I am just trying to provoke any more failures. About 26 out of the 168 have an issue of some sort and the supplier will be in at some point to fix them. Each chassis is drawing about 3KW right now.
 
I just found that thread on their forum too. Looks like he'll be stopping the folding on Monday or some time soon anyway. So we should be safe on our no.4 spot for now.

Wouldn't mind having one or two of those blades though :D
 
Two thousand cores.... twelve per blade.

Imagine if he'd run bigadv on those beasties ...
 
Yawn....wake me up in two years...I am more concerned with my sudden dive in bigadv and surge in GPU wu's and wondering how it will effect our production...
 
Using the FAH client for stability testing isn't a good thing and should be discouraged by Pande Group. Thanks, but no thanks.
 
Yawn....wake me up in two years

Exactly. If this was meant to be long term, what will our top 100 producers be running in two years?

Still an impressive result from just a weekend run. But nothing to be worried about.
 
Using the FAH client for stability testing isn't a good thing and should be discouraged by Pande Group. Thanks, but no thanks.

I'll agree with this... unless they cut off cleanly and don't drop WUs.
I've stress tested new gear with a -oneunit (is that flag still around?) in the past. I don't feel bad about it as long as I don't leave WUs half-done.
 
I'll agree with this... unless they cut off cleanly and don't drop WUs.
I've stress tested new gear with a -oneunit (is that flag still around?) in the past. I don't feel bad about it as long as I don't leave WUs half-done.
The problem isn't returning units. The problem is returning bad units. If your computer is slightly unstable, enough to not cause crashes, it could be making miscalculations in the work unit and that gets sent back to the server. I believe they wait until x number of people return the unit with the same answer, but it still wastes computing power. If you did return a bad one and someone else returned a good one, it has no way to know which is right, so it may request that more be sent out.

Not to mention, Folding/Rosetta/SETI is no where near as stressful as Prime95 or LinPack.
 
The problem isn't returning units. The problem is returning bad units. If your computer is slightly unstable, enough to not cause crashes, it could be making miscalculations in the work unit and that gets sent back to the server. I believe they wait until x number of people return the unit with the same answer, but it still wastes computing power. If you did return a bad one and someone else returned a good one, it has no way to know which is right, so it may request that more be sent out.

Not to mention, Folding/Rosetta/SETI is no where near as stressful as Prime95 or LinPack.

Point taken. It is very different for this guy to stress test rack upon rack of blades than for me, at home, to fold a bit to make sure an OC is stable (though that's always phase 2 for me. Prime gets run first.)

Prime95 would be much more appropriate for this guy to find bad hardware.
 
LinPack was made for testing servers with hundreds or thousands of cores and would have been perfect for this.
 
Well, as of the 6am update this guy hasn't dropped off on his production.
Maybe at the 9 or 12 update, after he gets to work, he'll shut down?

Here's hoping it's soon.
 
Well, as of the 6am update this guy hasn't dropped off on his production.
Maybe at the 9 or 12 update, after he gets to work, he'll shut down?

Here's hoping it's soon.
Actually for the science if they can avoid losing so many WU's we hope they keep going...we can wish they were on our Team, but we can't wish that someone stops folding.;)
 
That is very true and I agree that folding shouldn't be used as a stress test on its own. It looks as though he has stopped some clients. The number of WU's returned in the last couple of updates has halved compared to the average so hopefully he'll be done soon.
 
Back