Advanced search

Message boards : Number crunching : GIANNI ligand3-0-1 bad batch?

Author Message
Profile Stoneageman
Avatar
Send message
Joined: 25 May 09
Posts: 224
Credit: 34,057,374,498
RAC: 744
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 36376 - Posted: 18 Apr 2014 | 7:14:49 UTC
Last modified: 18 Apr 2014 | 7:19:40 UTC

All errorring out after 2 secs with
ERROR: file pme.cpp line 85: PME NX too small

Profile GDF
Volunteer moderator
Project administrator
Project developer
Project tester
Volunteer developer
Volunteer tester
Project scientist
Send message
Joined: 14 Mar 07
Posts: 1957
Credit: 629,356
RAC: 0
Level
Gly
Scientific publications
watwatwatwatwat
Message 36379 - Posted: 18 Apr 2014 | 9:46:17 UTC - in response to Message 36376.

sorry my fault testing a new submission system. It should be fixed now in lig3.

g

Profile Stoneageman
Avatar
Send message
Joined: 25 May 09
Posts: 224
Credit: 34,057,374,498
RAC: 744
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 36395 - Posted: 18 Apr 2014 | 16:33:29 UTC
Last modified: 18 Apr 2014 | 16:35:29 UTC

They seem to be working OK now except this one I noticed was going at a snails pace, so I aborted it.

Richard Haselgrove
Send message
Joined: 11 Jul 09
Posts: 1618
Credit: 8,604,144,351
RAC: 16,254,926
Level
Tyr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 36459 - Posted: 20 Apr 2014 | 11:44:56 UTC - in response to Message 36395.

They seem to be working OK now except this one I noticed was going at a snails pace, so I aborted it.

I'd say 'long', rather than 'slow'.

My GTX 670 machine got one of these too, and took over 14 hours to complete it. The credit awarded was proportional to the time spent, so I presume the team knew in advance how big these tasks are.

By coincidence, the next two tasks allocated to the host were 'GIANNI_trypben', at the opposite end of the size range: well under 4 hours on this card.

Gianni, if your tasks are going to have such a variable runtime, and if you're working on a new job submission system, could you extend that system to adjust <rsc_fpops_est> to match the job sizes, please? Our BOINC clients have a hard time scheduling everything properly when the estimates turn out not to match reality.

Profile Stoneageman
Avatar
Send message
Joined: 25 May 09
Posts: 224
Credit: 34,057,374,498
RAC: 744
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 36462 - Posted: 20 Apr 2014 | 16:42:02 UTC

When I said 'snails pace' I meant it had stopped making any meaningful progress rather than just slow. Since then I've had two error after 7hrs on different cards (780 & 570). Need to watch these like a hawk.

Matt
Avatar
Send message
Joined: 11 Jan 13
Posts: 216
Credit: 846,538,252
RAC: 0
Level
Glu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 36463 - Posted: 20 Apr 2014 | 16:50:38 UTC

I've received and successfully processed one of these so far. They do run much longer than anything else I've seen.

Profile skgiven
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,995,359,260
RAC: 0
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 36478 - Posted: 21 Apr 2014 | 14:57:51 UTC - in response to Message 36463.
Last modified: 21 Apr 2014 | 15:03:07 UTC

I will be keeping a lookout for tasks that don't progress; I've been caught out by those before!

I've just finished a GIANNI_lig3 on a GTX770@1267MHz (W7). Uses 76% Power, 90% GPU utilization, ~750MB GDDR. It took 12h 39min. This of course means if I have a large cache I will miss the 24h bonus. So a nice low cache will be needed by many, and WRT the power usage, I suggest people check these don't push some cards too hard (check your temps and OC's). I would expect these to sneak in under 12h on a GTX770 on Linux and XP (with a good setup). The upload proportionally larger, 77MB. The time on a GTX670 should be about 13h10min (W7), possibly 15min under 12h on Linux but a bit too close to call.

It's the first proper LONG run I've seen in a while, but does fit the description of 8h to 12h on the fastest card, the Titan Black, which is almost 50% faster than a GTX770.
____________
FAQ's

HOW TO:
- Opt out of Beta Tests
- Ask for Help

Profile Retvari Zoltan
Avatar
Send message
Joined: 20 Jan 09
Posts: 2343
Credit: 16,201,255,749
RAC: 0
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 36487 - Posted: 21 Apr 2014 | 16:35:17 UTC - in response to Message 36478.

I would expect these to sneak in under 12h on a GTX770 on Linux and XP (with a good setup).

You are right.
This one took 12 hours 1 minute and 41 seconds on a factory clocked (1084MHz) GTX 680 under WinXPx64.

Profile Mumak
Avatar
Send message
Joined: 7 Dec 12
Posts: 92
Credit: 225,897,225
RAC: 0
Level
Leu
Scientific publications
watwatwatwatwatwatwatwatwatwatwat
Message 36530 - Posted: 22 Apr 2014 | 10:30:33 UTC

Took 19h:43m on a 750 Ti. GPU Temp ~48 C.

biodoc
Send message
Joined: 26 Aug 08
Posts: 183
Credit: 10,085,929,375
RAC: 6,193,153
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 36533 - Posted: 22 Apr 2014 | 11:03:07 UTC

9 hours, 38 minutes on my GTX780Ti (underclocked; 906 MHz, linux). 2 have finished successfully.

Profile Carlesa25
Avatar
Send message
Joined: 13 Nov 10
Posts: 328
Credit: 72,619,453
RAC: 0
Level
Thr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 36592 - Posted: 23 Apr 2014 | 18:02:01 UTC - in response to Message 36533.

Hello: Finished a task correctly on my GTX 770 Linux 11H 20min.

The CPU load is too low does not pass of 9% over the GPU 98%.

Profile dskagcommunity
Avatar
Send message
Joined: 28 Apr 11
Posts: 456
Credit: 817,865,789
RAC: 0
Level
Glu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 36632 - Posted: 25 Apr 2014 | 9:55:07 UTC
Last modified: 25 Apr 2014 | 9:55:58 UTC

~18:40 on 570gtx with XP 32
____________
DSKAG Austria Research Team: http://www.research.dskag.at



Post to thread

Message boards : Number crunching : GIANNI ligand3-0-1 bad batch?

//