Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
World Community Grid Forums
Category: Completed Research Forum: The Clean Energy Project - Phase 2 Forum Thread: Why a 18 hour limit ??? |
No member browsing this thread |
Thread Status: Active Total posts in this thread: 9
|
Author |
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I'm starting to notice that I'm receiving errors on some work units because of a 18 hour CPU run time limit .... While FAAH2 and others are able to run until they are done, some up to 35 hr CPU times.
----------------------------------------Can someone that matters tell us why this is when we have 240 hours to return the UW. I'm sure by now any coding problems that allowed for endless loops, are gone right ?? So what's the real current reason for keeping that restriction in place?? Be sure it's a good one.... [Edit 2 times, last edit by Former Member at Dec 12, 2015 3:48:14 AM] |
||
|
deltavee
Ace Cruncher Texas Hill Country Joined: Nov 17, 2004 Post Count: 4842 Status: Offline Project Badges: |
Be sure it's a good one.... Is that a threat? This issue has been discussed ad nauseam. The duration of CEP2 workunits has been a topic for the more than five years this project has been running. And yes, the answer is a good one. You could have found this yourself in a simple forum search. Here it is straight from "someone that matters", back when the limit was 12 hours. http://www.worldcommunitygrid.org/forums/wcg/...ead,30547_offset,0#307466 edit - fixed link edit - because of the current forum paging problem you can not see the link unless you are logged in and have posts per page set to 50. [Edit 2 times, last edit by deltavee at Dec 11, 2015 7:49:19 PM] |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Gee Bro
----------------------------------------Don't be so up tight, I wouldn't know if it's old news or not many wouldn't. No matter how much it's been talked about I feel I have a right to ask. Still that link is 5 years old when it had 12 steps. Not a current valid reason, so I'd still like to hear from someone that matters.... Next.... [Edit 4 times, last edit by Former Member at Dec 12, 2015 3:34:23 AM] |
||
|
petehardy
Senior Cruncher USA Joined: May 4, 2007 Post Count: 318 Status: Offline Project Badges: |
Not a current valid reason, so I'd still like to hear from someone that matters.... I don't matter... But if you read the link above, you'd see that the slow computers are just used to validate the results, and that there's no point in wasting time completing WUs that have already been validated. "Patience is a virtue", I can't wait to learn it! |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I did see all that, even before I asked the question thanks very much.
----------------------------------------However there is no reason for it when most times letting it run just a few more hours gives the full result with out having to send it back out. I asked for a current valid reason, what else do you have that's not 5 years old structured for different WU's and has been changed once already anyway?? Can you offer something related to my question in terms of current information?? This after all is a Forum, not a Q&A page I'm not searching it out. I have enough to read every day and a good number of things to deal with. Maybe it's my nature in that if my team can't work with facts they don't work for me and I've done all the real research within my life I care to do, I'm well and long ago retired. If I didn't have to banter back and forth with you and/or we had a honest/current Q&A page this could be done already. Is it because they don't want the WU's to get done to fast so they send them out over and over to balance that with validations?? Is it because they feel they take to long fewer people will run them?? Is it because they only want to deal with compounds that fit within that range?? To me that makes the least sense. I'd also like to know why there are no metals, but I bet that's past most cruncher's range. Is it all some or none of that?? I mean really it's just a number in a program, whats the big deal in changing it ?? After all it places a limit on the number of electron shells and complexity any one compound can have .... If you have nothing else to offer your information has been noted and I'm disregarding it. I feel you did not add anything new. That is after all what a forum is best for. Next......PLZ [Edit 6 times, last edit by Former Member at Dec 12, 2015 10:24:14 AM] |
||
|
Falconet
Master Cruncher Portugal Joined: Mar 9, 2009 Post Count: 3294 Status: Offline Project Badges: |
---------------------------------------- AMD Ryzen 5 1600AF 6C/12T 3.2 GHz - 85W AMD Ryzen 5 2500U 4C/8T 2.0 GHz - 28W AMD Ryzen 7 7730U 8C/16T 3.0 GHz |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Yes I seen that....a discussion on what happens with a 18 hour limit.
----------------------------------------Lets get into the why? One of my WU's got to 18 hours and didn't even get past job one ("0") and it's a older computer I made from peaces and parts that so far has done two years running two at a time. With a newer computer odds are it's twice as fast as mine and one got all the way to saving job 4 in 18 hours he/she got credit I got zero for the 3rd time in two weeks. The 18 hour limit isn't based on a computers ability to preform it's based on the clock within it so it would seem. There are FAAH2 WU's that can take 24-26++ hours why is CEP2 based on 18 ??? A 18 hour limit may be a fast way to screen them based on a filter program that kind of works so they can pick what they like on the 1 out of what a few 100 that fit the bill of what they are looking for. So whats the matter with 24 or 30 or 36 hours why are we forced to eat ANY dead CPU time because a WU never got passed the first job. Why not base the limit on a different value like getting past the first job and then a 18 hour limit?? Because it seems the code is good enough to not get into endless loops let us at least complete the first job......whats the big deal????????????? To me it really displaces the value they tell us we are worth to them, like it'as all so much plant food. [Edit 3 times, last edit by Former Member at Dec 12, 2015 6:50:16 PM] |
||
|
ca05065
Senior Cruncher Joined: Dec 4, 2007 Post Count: 325 Status: Recently Active Project Badges: |
From my memory of reading this forum, there are two reasons why the 18 hour limit is in place:
1) for a specific molecule the work unit is only converging slowly to an answer - this is worse if it is a slower CPU. Such work units are re-run by the scientists on a superfast machines. 2) for a specific molecule the work unit will never converge to an answer even on the fastest CPU. |
||
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7545 Status: Offline Project Badges: |
This project started with a 12 hour work limit. It was increased to an 18 limit on December 18, 2013. here In reading the forums it appears the reason for the new limit had several reasons. The 12 hour limit was imposed as a limit so most of the machines could finish most of the time. It was also a preventive measure against a divergent condition versus a convergent condition (different from an infinitely looping problem). It appears the limit was changed in response to a change in project to accommodate some larger molecule sizes. There was also a change in the work units from 16 jobs per WU to 8 jobs per WU. These larger molecules started generating upload files approaching 100 mbs. I presume if the length of the jobs was allowed to go even longer, the size of the upload files would have been proportionally bigger. For those on slow ISP's the upload times would have become untenable. In conclusion it seems the 18 hour limit was an arbitrary compromise to satisfy most of the crunchers most of the time and yet still get the needed information to the scientists.
----------------------------------------Cheers
Sgt. Joe
*Minnesota Crunchers* |
||
|
|