| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: OpenZika
Thread: super low credit for some super long Zika WU |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 44
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
OMG, I see what you mean!
|
||
|
|
SekeRob
Master Cruncher Joined: Jan 7, 2013 Post Count: 2741 Status: Offline |
The fiction... and the facts: https://bit.ly/WCGCPH1
----------------------------------------Try petitioning uplinger to port the sniftiness he recently applied to HST one, to address the undercrediting of runtime 'outliers', but it's just not as bad as it is made out to be, on average, in contra to the fiction, the facts show there's the last week(s) been a strong convergence of all sciences towards the 26-27 credit per hour average mark. Possibly a factor in this issue is/was the speed disparity between the *nux based platforms (10% or so of contribution and twice as fast as 90% done Windows] Given that a large portion of the office crunching resource went missing in the last weeks and months [an uncovered certificate issue], the averages might go down again.  [Edit 1 times, last edit by SekeRob* at May 26, 2016 6:02:07 PM] |
||
|
|
keithhenry
Ace Cruncher Senile old farts of the world ....uh.....uh..... nevermind Joined: Nov 18, 2004 Post Count: 18667 Status: Offline Project Badges: 
|
That's great that the sciences are converging on the same average credit per hour. Seems like that it should be that way. As I recall though, vina makes getting a new project up and running much faster and easier. Sure can't blame the scientists or admins for using it. However, over the years, we've seen that it is best suited for projects with highly homogeneous WUs. When the WUs start having increasingly varying run times, we start seeing the credits given drop off for the outliers. If what Keith did for the TB outlier issue could be ported to Zika, that would be great but I didn't think TB was vina-based.
---------------------------------------- |
||
|
|
jmps
Cruncher Joined: Nov 21, 2004 Post Count: 13 Status: Offline Project Badges: 
|
I though Zika was VINA based? Vina. See what the error says: Error: could not open ".\ZINC12591756.pdbqt" for reading. VINA failed. rc = 1. Exiting |
||
|
|
pvh513
Senior Cruncher Joined: Feb 26, 2011 Post Count: 260 Status: Offline Project Badges:
|
but at least look on the bright side - at least the long units are not erroring any more (so far). I have one of those very long units, and it is still erroring out. They all get error -131 "file size too big". Is this from before the fix, and if so why hasn't it been removed from the database? ZIKA_ 000001186_ x1nb7_ HCVJ4_ RNAPol_ wRNAand2Mn_ chnA_ 0309_ 4-- Linux 4.1.15-8-default 705 Error 5/28/16 10:33:05 5/29/16 19:33:04 23.52 396.2 / 0.0 ZIKA_ 000001186_ x1nb7_ HCVJ4_ RNAPol_ wRNAand2Mn_ chnA_ 0309_ 3-- Linux 2.6.32-573.26.1.el6.x86_64 705 Error 5/27/16 04:05:50 5/28/16 10:32:50 15.29 148.8 / 0.0 ZIKA_ 000001186_ x1nb7_ HCVJ4_ RNAPol_ wRNAand2Mn_ chnA_ 0309_ 2-- Linux 4.3.0-1-amd64 - In Progress 5/27/16 04:05:31 5/30/16 16:05:31 0.00 0.0 / 0.0 ZIKA_ 000001186_ x1nb7_ HCVJ4_ RNAPol_ wRNAand2Mn_ chnA_ 0309_ 1-- Linux 2.6.32-504.1.3.el6.x86_64 705 Error 5/26/16 03:30:34 5/27/16 04:05:18 23.95 29.8 / 0.0 ZIKA_ 000001186_ x1nb7_ HCVJ4_ RNAPol_ wRNAand2Mn_ chnA_ 0309_ 0-- Microsoft Windows 10 Professional x64 Edition, (10.00.10586.00) 705 Error 5/21/16 22:47:03 5/26/16 03:30:26 19.40 29.8 / 0.0 |
||
|
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7847 Status: Offline Project Badges:
|
I think these ultra long units are getting fewer and farther between. I initially saw a couple of these, but have not seen one now for several days. Most of these units should have been caught when Uplinger increased the acceptable file size, but perhaps these last units even exceed that. If it is not too much trouble, could you post the number of tasks in this unit ? My guess is it will exceed 300 tasks.
----------------------------------------Cheers
Sgt. Joe
*Minnesota Crunchers* |
||
|
|
Sandvika
Advanced Cruncher United Kingdom Joined: Apr 27, 2007 Post Count: 112 Status: Offline Project Badges: 
|
I'm only doing Zika on Android at the moment as I'm chasing my 10 year UGM badge before the project ends on Intel - and it's going to be close - how exciting! The low credit issue is affecting me too, not just for long WUs, which is why I'm making an appearance here
---------------------------------------- I had assumed that I've been getting very low credit for OET on Android because the devices are slow compared to Intel CPUs and thus both my Android devices had averaged out on this low level of credit having achieved reliability. If I'm wrong, it's probably a good thing that Zika is dragging the credit upwards by a factor of 4! My device was the _3 re-issue on this WU. Project Name: OpenZika Created: 05/19/2016 21:26:18 Name: ZIKA_000000540_x1nb7_HCVJ4_RNAPol_JustProt_chnA_0033 Minimum Quorum: 2 Replication: 2 Result Name OS type OS version App Version Number Status Sent Time Time Due / Return Time CPU Time / Elapsed Time (hours) Claimed/ Granted BOINC Credit ZIKA_ 000000540_ x1nb7_ HCVJ4_ RNAPol_ JustProt_ chnA_ 0033_ 3-- Android 3.4.0-5508620 (Android 5.0) 705 Valid 30/05/16 04:05:27 31/05/16 08:24:21 7.07 17.9 / 77.7 ZIKA_ 000000540_ x1nb7_ HCVJ4_ RNAPol_ JustProt_ chnA_ 0033_ 2-- Android 3.0.21-00002-OMAP-Android 705 Error 30/05/16 03:42:47 30/05/16 04:04:58 0.00 27.7 / 0.0 ZIKA_ 000000540_ x1nb7_ HCVJ4_ RNAPol_ JustProt_ chnA_ 0033_ 1-- Android 3.18.14-7538748 (Android 6.0.1) - No Reply 20/05/16 03:42:16 30/05/16 03:42:16 0.00 0.0 / 0.0 ZIKA_ 000000540_ x1nb7_ HCVJ4_ RNAPol_ JustProt_ chnA_ 0033_ 0-- Android 3.4.0-4445722 (Android 5.0) 705 Valid 20/05/16 03:42:05 22/05/16 12:00:08 9.01 137.5 / 77.7   |
||
|
|
SekeRob
Master Cruncher Joined: Jan 7, 2013 Post Count: 2741 Status: Offline |
That's great that the sciences are converging on the same average credit per hour. Seems like that it should be that way. As I recall though, vina makes getting a new project up and running much faster and easier. Sure can't blame the scientists or admins for using it. However, over the years, we've seen that it is best suited for projects with highly homogeneous WUs. When the WUs start having increasingly varying run times, we start seeing the credits given drop off for the outliers. If what Keith did for the TB outlier issue could be ported to Zika, that would be great but I didn't think TB was vina-based. Sadly, a week on the high variability continues to be a feeder for killing credit... Zika avg runtime at it's lowest since start, credit at lowest, convergence out the window. What though is a riddle is, why is FAHB with a fairly regular runtime the lowest? BEDAM is hard to crunch, but is that a reason to give least of all [ATM]... is it harder to crunch due inefficiency, thus putting through less fpops/mips, the base measure for credit? The crunching time is is still spend, no instrument available to determine what's running most efficient [but by measure of points]. Not the first time the question is asked... how to optimize the use of the pool resource? [Then really there needing to be a pool badge [argh], so total contribution is recognized 'visibly', instead of this constant emphasis on individual sciences. Charts: Credit/Hour https://bit.ly/WCGCPH1 Runtime/Result: https://bit.ly/WCGART1 |
||
|
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7847 Status: Offline Project Badges:
|
The crunching time is is still spend, no instrument available to determine what's running most efficient [but by measure of points]. Not the first time the question is asked... how to optimize the use of the pool resource? Off topic alert. A reasonable question which would really need some inside research by the techs. It would make sense to tailor the units to the hosts which are best suited for those units. Might make a good thesis for a masters level computer science student. Cheers
Sgt. Joe
*Minnesota Crunchers* |
||
|
|
SekeRob
Master Cruncher Joined: Jan 7, 2013 Post Count: 2741 Status: Offline |
\ot Think there's a working paper out there already, optimizing use of a non-homogeneous computing resource. Would have to dig, vaguely remembering knreed put a link out to the document in his previous tenure, possibly in places I can't go anymore /ot
|
||
|
|
|