Advanced search

Message boards : Number crunching : How to calculate the granted credit of a given WU?

Author Message
Profile Retvari Zoltan
Avatar
Send message
Joined: 20 Jan 09
Posts: 2343
Credit: 16,201,255,749
RAC: 7,520
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 21658 - Posted: 9 Jul 2011 | 21:23:12 UTC
Last modified: 9 Jul 2011 | 21:28:48 UTC

The new TONI_SH2MS tasks are granted much less credit/time than previous WUs.
The project's total granted credit is dropped by 7M (from 56.5M to 49.5M, it's a 12.4% drop)
I tried to figure out the basis of how many credit granted for a WU, but I can't. Maybe I've overlooked something.

____________ Name _____________________ Run __ Granted _ credit _ Steps __ Time ____ credit
______________________________________ Time ___ Credit __ /sec _________ /Step ____ /Mstep
s0r368-TONI_SH2MS3-0-100-RND4752_0 ___ 15830s _ 37485 __ 2.368 __ 2M __ 7.921 ms _ 18742
I67R2-GIANNI_KKFREE4-7-10-RND5385_1 __ 25812s _ 67498 __ 2.615 __ 4M __ 6.459 ms _ 16874
A596-TONI_AGGdense1-18-100-RND3626_1 _ 15893s _ 50647 __ 3.187 __ 3M __ 5.327 ms _ 16882

So I'd like to be enlightened in this matter.

Bedrich Hajek
Send message
Joined: 28 Mar 09
Posts: 467
Credit: 8,185,946,966
RAC: 10,563,100
Level
Tyr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 21659 - Posted: 10 Jul 2011 | 4:11:01 UTC - in response to Message 21658.

So I'd like to be enlightened in this matter.

Me too!

Complain long enough and loud enough, and they'll raise the credits awarded. They are utterly worthless, but there should be consistency. And some people are more competitive than others.

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 21661 - Posted: 10 Jul 2011 | 10:21:03 UTC - in response to Message 21659.
Last modified: 10 Jul 2011 | 10:21:22 UTC

APPLICATIONS
There are two applications which can be selected in your project preferences, acemd2 and acemd_long. By default both are selected. The acemd2 workunits are normal workunits to complete in few hours on the most common cards of the time (up to 3 times slower than top cards). The acemd_long workunits are several times longer and should complete in maximum 12 hours on the top cards.

CREDITS
The credits are returned by the workunit using an estimate of the amount of flops based on number of atoms and number of iterations. Variation of 10-20% between workunits which take the same time to complete are probably normal and can depend on many factors. Some as simple as the FFT uses multiple of 2 for one workunits and not for another one. acemd_long doubles this estimate to compensate for the burden of having your machine computing for as much as 12 hours on a top card and the additional risk of losing the credits to due to a crash.

DEADLINES
The server deadline is 5 days. The server assigns 50% more credits if you return within 24 hours and 25% more credits within 48 hours. After 3 days the server reissue a workunit to make sure it is computed as fast as possible, although if the first one comes in, it cancels the new one immediately if it is not being computed already.

Hope it helps. We should probably put this somewhere more visible.

gdf

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 21667 - Posted: 10 Jul 2011 | 14:17:55 UTC - in response to Message 21661.

done

Profile Retvari Zoltan
Avatar
Send message
Joined: 20 Jan 09
Posts: 2343
Credit: 16,201,255,749
RAC: 7,520
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 21668 - Posted: 10 Jul 2011 | 21:56:21 UTC - in response to Message 21661.

CREDITS
The credits are returned by the workunit using an estimate of the amount of flops based on number of atoms and number of iterations. Variation of 10-20% between workunits which take the same time to complete are probably normal and can depend on many factors.

These general principles were already known to me. A common cruncher doesn't know the number of atoms. My original intention was to understand credit calculation as much as I can do the calculation.
Comparing TONI_SH2MS and TONI_AGGdense workunits, while their runtime is within 0.4%, their granted credit is 26% apart (or 35%, depending on which one I take as base). Based on the variation you've defined above, the crediting of one of these workunits doesn't qualify as "probably normal".

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 21669 - Posted: 10 Jul 2011 | 22:15:33 UTC - in response to Message 21668.

In this case, one of the wu does not use the CPU at all, the other does but we do not compensate for it in the current scheme.

gdf

Post to thread

Message boards : Number crunching : How to calculate the granted credit of a given WU?

//