• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

PrimeGrid Autumnal Equinox Challenge September 23 - October 6

Page 7 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Good luck, bad luck or both, I'm not sure. About 3 hours after the race ended, my power supply in my main rig gave up the ghost. Just a click and everything went dark. 😱

Thankfully, I've got a Microcenter close by, new PS in place and up and running again.

I think it simply overheated. It's in a case with the PS intake on the bottom that had a little dust filter. The filter did it's job, but I had totally forgotten about it. Oops.
 
RobertE said:
Good luck, bad luck or both, I'm not sure. About 3 hours after the race ended, my power supply in my main rig gave up the ghost. Just a click and everything went dark. 😱

Thankfully, I've got a Microcenter close by, new PS in place and up and running again.

I think it simply overheated. It's in a case with the PS intake on the bottom that had a little dust filter. The filter did it's job, but I had totally forgotten about it. Oops.
Click to expand...

Hopefully you got a good quality unit, an overheated PSU should simply shut off.
 
theAnimal said:
Hopefully you got a good quality unit, an overheated PSU should simply shut off.
Click to expand...

It did. It smelled hot, not burnt. Computer is back up and running. It did not revive itself (power supply) after the trip to microcenter and back. So now to deal with OCZs RMA process. At least nothing appears to be dead.
 
I'll keep track of the development of the stats and how much the stats change during the clean-up phase.

Teams' stats as of October 7, 17:00:46
06 ______ PrimeSearch Team_______ 36 004 816.19 _____ (+/-0)
07 ______ Czech National Team ___ 31 422 841.19 _____ (-60 884.81)
08 ______ Dutch Power Cows ______ 23 336 031.62 _____ (+/-0)
09 ______ TeAm AnandTech ________ 17 780 900.19 _____ (-84 072.81) :\
10 ______ BOINC@Poland __________ 17 252 481.57 _____ (-58 128.43)
11 ______ BOINCstats ____________ 16 879 494.55 _____ (+/-0)
12 ______ Duke University _______ 16 155 748.93 _____ (-2 481.07)

Not too bad, though ... but I really hope that we will not loose 500K points!
 
198 results validated, 13 results pending, 0 with errors, 4 marked as invalid.

The 4 "invalid results" reached consensus before my result was sent in. Thus I crunched the WU, sent in successfully a correct result and got no points.
<Petrus shrugs>.
Those 4 results would have given some 105K points in the race, which would have given me one step up the ladder.
Thus, if they send out 4 - 5 replications and got two incorrect results and three correct results, then only the two, which arrive first, get the points.
Hmmm, I am not so sure I like that.
 
petrusbroder said:
198 results validated, 13 results pending, 0 with errors, 4 marked as invalid.

The 4 "invalid results" reached consensus before my result was sent in. Thus I crunched the WU, sent in successfully a correct result and got no points.
<Petrus shrugs>.
Those 4 results would have given some 105K points in the race, which would have given me one step up the ladder.
Thus, if they send out 4 - 5 replications and got two incorrect results and three correct results, then only the two, which arrive first, get the points.
Hmmm, I am not so sure I like that.
Click to expand...

I'm sure Ken will correct me if I'm wrong, but as long as you return the result within the challenge window and it validates as "valid" against the other results, then you should get points for it.
 
I wonder if there's some sort of threshold for invalid results. Like if the WU gets two invalids within a certain amount of time, then the WU is "dumped" and no one past the first couple of results gets the points...
 
RobertE said:
I'm sure Ken will correct me if I'm wrong, but as long as you return the result within the challenge window and it validates as "valid" against the other results, then you should get points for it.
Click to expand...
I actually asked about this at the start of the race, as I had one computer that was too slow to complete one WU in 7 days! (It took 9.5 or so.) I verified that the WU was sent in, late, and I got credit for it. I'm sure you did too, Peter.
somethingsketchy said:
I wonder if there's some sort of threshold for invalid results. Like if the WU gets two invalids within a certain amount of time, then the WU is "dumped" and no one past the first couple of results gets the points...
Click to expand...
Note this line on every work unit page:

max # of error/total/success tasks 15, 15, 5

That means if there are 15 total attempts, or 5 successful but different results, no more work is sent out for that task until the admins take a look at it. This is rare enough that they usually send it out for at least one more round of 15/15/5.
 
Last edited:
Peter hasn't posted our progress regress today, so:

6_____36004816___PrimeSearchTeam (+/-0)
7_____31422841___Czech National Team (-60 885, ~2 bad WUs)
8_____23336031___Dutch Power Cows (+/-0)
9_____17665505___TeAm AnandTech (-199 468, ~6 bad WUs :\)
10____17132460___BOINC@Poland (-178 150, ~5 bad WUs)
11____16850512___BOINCstats (-28 983, ~1 bad WU)
12____16155748___Duke University (-2 482)
 
Is there any way to find out if the invalid result came from your username/account? I am curious if my overclock isn't truely stable. It passes one hour of OCCT Linpack 64-bit with 90&#37; RAM (out of 8GB), so I figured it was stable.

Edit: On my primary rig, I didn't have to boost the vcore. Since my Q9300 is essentially two E5200 chips sandwiched together, and E5200 chips often do 3.0 Ghz on stock voltage, I didn't think it was too much of a stretch for my chip to do 3.0 Ghz at stock voltage. My secondary rig, which has identical components (except for the case), needed a couple of minor clicks upward on the vcore to be OCCT one-hour stable.

Edit: I found a way on the PrimeGrid web site to drill down to the active computers, and then to the WUs. None of mine show Invalid, but there is one Woodall LLR that is still Pending credit.
 
Last edited:
Regression testing and reporting. It's what I do for a living, and I'm doing it here now too. (Sort of.) :whiste:

6_____36004816___PrimeSearchTeam (+/-0)
7_____31392432___Czech National Team (-91 294, ~3 bad WUs)
8_____23336031___Dutch Power Cows (+/-0)
9_____17606690___TeAm AnandTech (-258 283, ~8 bad WUs :\)
10____16957322___BOINC@Poland (-353 288, ~12 bad WUs 😱)
11____16820580___BOINCstats (-58 915, ~2 bad WUs)
12____16155748___Duke University (-2 482)

Well, doesn't look like we're going to fall back behind BOINC@Poland. But they might fall behind BOINCstats!
 
PrimeSearchTeam consists of the PrimeGrid admins, plus others very experienced in prime searching. They know better than to overclock their systems too high. 😉 (And/or they're always testing for primes, so if something was producing errors they'd know about it long before the race.)

And, FYI, as of this morning, the Cows have lost one.
 
Faulty WU reports should be slowing down, so I thought I'd wait a bit until my next report. (Or Peter can do some reporting.)

6_____36004816___PrimeSearchTeam (+/-0)
7_____31266849___Czech National Team (-216 877, ~7 bad WUs)
8_____23278384___Dutch Power Cows (-57 647, ~2 bad WUs)
9_____17491217___TeAm AnandTech (-373 756, ~13 bad WUs :\)
10____16806669___BOINC@Poland (-503 941, ~15 bad WUs 😱)
11____16763543___BOINCstats (-115 952, ~4 bad WUs)
12____16123998___Duke University (-34 232 ~1 bad WU)

Nobody's perfect, except PrimeSearchTeam. BOINC@Poland may still lose to BOINCstats - 2 more bad WUs would do it!

Edit: As of 10/12, BOINCstats is now 10th and BOINC@Poland is now 11th!
 
Last edited:
You must log in or register to reply here.

TRENDING THREADS

Back
Top