| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: speeding up BOINC calcs... |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 14
|
|
| Author |
|
|
KLiK
Master Cruncher Croatia Joined: Nov 13, 2006 Post Count: 3108 Status: Offline Project Badges: 
|
I noticed that when dual-core or dual-CPU computers are crunching different projects, that the projects gets crunched more quickly, than to crunch two (or more) of the same type of project...
----------------------------------------but I haven't found any option in BOINC to crunch only different projects data! is there a way to do that? how? |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Sorry, KLiK, there is no such control (at any level).
The effect you see is real, but cannot be generalised. It only occurs when a research application causes a bottleneck effect on some resource other than the processor pipeline. The most common example is L2 cache, and the page fault problem. As a result, we don't advise that multicore crunchers work exclusively on HCC, which has the biggest page fault problem. However, on a dual core machine, the problem isn't severe and generally isn't worth worrying about. |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
What platform is this HCC page faulting, Didactylos? None of the sorts here, very slick on Vista/Quad or XP/Duo in the last incarnation. The baddies in my book are CEP and at distance HPF2. Think Jean mentioned HCC on a particular Linux(?) showing large numbers in last few days.
----------------------------------------And KLiK, it's a long known fact, also some projects generate because of their specificity, a number of degrees hotter CPU running. Selecting an all project mix so the probability of getting different task to run concurrently is great. It's been requested to have a control to allow only 1 of some and someone actually wrote a script to facilitate this for a few none WCG project, but think it was limited to duo.
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
KLiK
Master Cruncher Croatia Joined: Nov 13, 2006 Post Count: 3108 Status: Offline Project Badges:
|
@Didactylos
----------------------------------------OK, so HCC is not recommended for the simultaneous work on multi-core machines! does that involve the new HCCC? what other projects are not recommended for the simultaneous work on multi-core machines? and which one are not recommended for the simultaneous work on dual-core/core2duo machines? @Sekerob like you said: it should be looked at, 'cause the WCG only yesterday delivered 251.3years of CPU work...if there is 1% chance of improvement (and it's higher than that, I'm pretty sure of it), that means WCG will be more efficient in 2.5years of CPU work per day of crunching! efficiency is something that will not waste our time for finding a cure (more than 90% of projects are for help finding cures for diseases) & will not waste the current that out PCs spend for calculations! I'm not worried about the CPU heating, though it does heat up my room and the surrounding areas (and atmosphere) - 'cause that is causality of crunching DATA with our CPUs...though I'm thinking about installing a liquid cooling for the 3 computers that I use... |
||
|
|
JmBoullier
Former Community Advisor Normandy - France Joined: Jan 26, 2007 Post Count: 3716 Status: Offline Project Badges: 
|
Think Jean mentioned HCC on a particular Linux(?) showing large numbers in last few days. Yes, I did mention it several times here or there since January. It is for HCC under Ubuntu 8.10 64-bit with Boinc 6.2.15. I still have no idea of who is the culprit, Ubuntu, Boinc, or some particular library. The soft page fault rate is 125,000 pf/s per task on a Q6600 running at 2.88 GHz. It does not matter at all if there is one or more HCC running at the same time, i.e. you get a total 125,000 rate for one task alone and 500,000 for four concurrent tasks. Despite that the disaster is limited: the average duration of HCC tasks is about 20-25 minutes more than before (with Ubuntu 8.04 64-bit and Boinc 5.10.45), but still much less than when running that same machine under XP SP3 32-bit. Cheers. Jean. |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
@Didactylos OK, so HCC is not recommended for the simultaneous work on multi-core machines! does that involve the new HCCC? what other projects are not recommended for the simultaneous work on multi-core machines? and which one are not recommended for the simultaneous work on dual-core/core2duo machines? @Sekerob like you said: it should be looked at, 'cause the WCG only yesterday delivered 251.3years of CPU work...if there is 1% chance of improvement (and it's higher than that, I'm pretty sure of it), that means WCG will be more efficient in 2.5years of CPU work per day of crunching! efficiency is something that will not waste our time for finding a cure (more than 90% of projects are for help finding cures for diseases) & will not waste the current that out PCs spend for calculations! I'm not worried about the CPU heating, though it does heat up my room and the surrounding areas (and atmosphere) - 'cause that is causality of crunching DATA with our CPUs...though I'm thinking about installing a liquid cooling for the 3 computers that I use... KLiK, to emphasize the PF ing in large numbers for HCC is platform and even OS version dependent. My devices show negligible numbers e.g. For HFCC on Vista Quad and XP Duo is negligible, certainly less and not more than for instance FAAH & DDDT that run on AutoDock science Engine. As for 1%, you can throw an army of programmers at it to find the cause and optimize the code. There's a major cost to that to WCG and it's not limitless so the concentration is on the low hanging fruit, and sometimes a Eureka. Set a high project mix and the chance of like meeting gets smaller.
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
KLiK
Master Cruncher Croatia Joined: Nov 13, 2006 Post Count: 3108 Status: Offline Project Badges:
|
@Didactylos OK, so HCC is not recommended for the simultaneous work on multi-core machines! does that involve the new HCCC? what other projects are not recommended for the simultaneous work on multi-core machines? and which one are not recommended for the simultaneous work on dual-core/core2duo machines? @Sekerob like you said: it should be looked at, 'cause the WCG only yesterday delivered 251.3years of CPU work...if there is 1% chance of improvement (and it's higher than that, I'm pretty sure of it), that means WCG will be more efficient in 2.5years of CPU work per day of crunching! efficiency is something that will not waste our time for finding a cure (more than 90% of projects are for help finding cures for diseases) & will not waste the current that out PCs spend for calculations! I'm not worried about the CPU heating, though it does heat up my room and the surrounding areas (and atmosphere) - 'cause that is causality of crunching DATA with our CPUs...though I'm thinking about installing a liquid cooling for the 3 computers that I use... KLiK, to emphasize the PF ing in large numbers for HCC is platform and even OS version dependent. My devices show negligible numbers e.g. For HFCC on Vista Quad and XP Duo is negligible, certainly less and not more than for instance FAAH & DDDT that run on AutoDock science Engine. As for 1%, you can throw an army of programmers at it to find the cause and optimize the code. There's a major cost to that to WCG and it's not limitless so the concentration is on the low hanging fruit, and sometimes a Eureka. Set a high project mix and the chance of like meeting gets smaller. where can I suggest to put another command fo cc_config.xml, for example: <avoid_run_same_app>0|1</avoid_run_same_app> ? |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
At http://boinc.berkeley.edu/dev/index.php
----------------------------------------Formulate your thoughts carefully. If someone runs a multi project client on a quad, WCG and say CPDN and only elects 1 project of WCG, you'll find cores idling. It will further kill the project weighting since if attached to 4 unique projects and cache filled, you again will find cores idling. So, you'll need exception rules too. Not exactly an option for beginners.
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
gb009761
Master Cruncher Scotland Joined: Apr 6, 2005 Post Count: 3010 Status: Offline Project Badges:
|
There is an alternative way to do this - one which, depending upon your set-up, you may/may not wish to pursue... and that's to manage your WU's manually
---------------------------------------- Me, I've just got the 1 dual-core machine that I can get access to very easily, and therefore, I can closely monitor what's running at any particular time. Thus, as I'm "working" on 3 badges at the moment (CEP, HCC & HFCC), I endeavour to keep 1 core running on CEP whilst the other is on either of the other 2 cancer projects. In fact, I find it quite interesting to see if I can manually juggle the WU's so that none of them would end up being returned 'too late' whilst still keeping enough in my 'awaiting processing' queue so as to allow me to do this juggling act.  |
||
|
|
KLiK
Master Cruncher Croatia Joined: Nov 13, 2006 Post Count: 3108 Status: Offline Project Badges:
|
At http://boinc.berkeley.edu/dev/index.php Formulate your thoughts carefully. If someone runs a multi project client on a quad, WCG and say CPDN and only elects 1 project of WCG, you'll find cores idling. It will further kill the project weighting since if attached to 4 unique projects and cache filled, you again will find cores idling. So, you'll need exception rules too. Not exactly an option for beginners. I just did that here: http://boinc.berkeley.edu/dev/forum_thread.php?id=3769 so someone should adress that! |
|||
|
||||
|
|