| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
| World Community Grid Forums
|
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 56
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
How about putting some figures into the CPU Speed column? At least some of the projects have CPU time limits enforced, which means that if you run too slow a machine, the WUs will be aborted before they complete.
eg. I tried a Pentium 200 MMX on HPF2 and it got aborted by WCG when 95% through. So, HPF2 should have around 225 MHz as an absolute minimum. If you can get the CPU maximums from the techs for other projects, it might be possible to guesstimate other minimums. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
We used to enforce minimum speed requirements ( http://www.worldcommunitygrid.org/help/viewTopic.do?shortName=minimumreq ) but we have done away with them since varying work unit sizes meant that we were arbitrarily keeping old machines from doing useful work. Now our members get to make that decision.
Lawrence |
||
|
|
MD-Crusher
Advanced Cruncher Joined: Jul 23, 2007 Post Count: 52 Status: Offline |
How about putting some figures into the CPU Speed column? At least some of the projects have CPU time limits enforced, which means that if you run too slow a machine, the WUs will be aborted before they complete. eg. I tried a Pentium 200 MMX on HPF2 and it got aborted by WCG when 95% through. So, HPF2 should have around 225 MHz as an absolute minimum. If you can get the CPU maximums from the techs for other projects, it might be possible to guesstimate other minimums. As i understand it CPU speed is irrelevant in relation to the time-out. The HPF2 task header has current average OPS in the header and is allowed to run up to 6x those estimates, none related to real time. On receipt BOINC computes how much time those estimated OPS would take on the machine with 200mhz. Thus if the estimate was 5 hours on a 2ghz machine, it would be 50 hours on a 200mhz machine (not exactly linear). At six times it would be allowed to run 300 hours. More probable is that you ran into the hang-up bug of HPF2 and it cycling in a loop when it reached the 95%, which is when it hit the OPS times 6 limit. Think the programmers chose to use the "Exceeded Time" message over "Exceeded OPS" for the non-technical members. [Edit 1 times, last edit by MD-Crusher at Dec 31, 2007 10:03:00 AM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Okay. I'm only unconvinced because I'm pretty sure it wasn't in a loop ... everything was progressing in the usual fashion. (I'm not saying "95%" as in what BOINC reported, since the job was killed, so I have no idea what the reported % was. I'm referring to 95% of what earlier "current CPU time"/"fraction done" came out to.)
----------------------------------------Change of hardware may have confused BOINC but, if so, it means that BOINC ignored current benchmark data and ignored current "current CPU time"/"fraction done" for the limit and somehow used old, irrelevant data instead. Seeing as there's no decent documentation on almost anything about BOINC, it's hard to tell what it might do though. (There should be a nice, big, standard "man" entry for BOINC describing all of its data fields, how they are created and what they are used for. Instead, there's a huge pile of incomplete wikis scattered about the web.) [Edit 1 times, last edit by Former Member at Jan 1, 2008 1:28:11 AM] |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Long Last finally got the checkpoints added to the matrix (opening post of this thread). Matrix can now also can be found in the Start Here forum FAQ's.
----------------------------------------The waiting was on AC@H jobs to get the last missing bits **. Monsters taking 20 hours on my P4. Also applied other tweaks like CPU time indications. They can widely vary so plz dont go shooting from the hip. (My gatling is loaded and magazine to the brim). Suggestion, corrections, tweaks always welcome.... plz volunteer the info. (note to Jal2. Still not figured what it is about what the fpu has of impact on the different sciences. Think the L2 cache size and gigabytes of ram and fsb, all the bigger the better, is of far greater effect on crunching throughput and tell you what, replacing an old IDE drive and replacing with a 16mb cached 500mb 7200 rpm monster alone had material impact for larger jobs.... my 2.1 cents opinion)
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Added the Quorum and Initial Distribution of the projects to the table.
----------------------------------------
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Table updated. Per post by Uplinger, the Official Checkpoint count for AfricanClimate is 56 per job. We're on cycle 19 meaning the project phase is about 67% complete (18/27th)
----------------------------------------
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
The DDD-T row needs to be updated for the new, bigger WUs.
|
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Actually there are a few more bits which were waiting for the web update.... find the difference
----------------------------------------![]() Edit: F5 ![]()
WCG
----------------------------------------Please help to make the Forums an enjoyable experience for All! [Edit 1 times, last edit by Sekerob at Mar 7, 2008 4:15:29 AM] |
||
|
|
JmBoullier
Former Community Advisor Normandy - France Joined: Jan 26, 2007 Post Count: 3716 Status: Offline Project Badges:
|
Sek,
----------------------------------------The new DDDT WUs are not yet on a 1/1 quorum scheme. And they are already on a 12-day deadline. Thank you for your efforts to keep this chart up-to-date (tough job!). Jean. |
||
|
|
|