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: 156
|
![]() |
Author |
|
sk..
Master Cruncher http://s17.rimg.info/ccb5d62bd3e856cc0d1df9b0ee2f7f6a.gif Joined: Mar 22, 2007 Post Count: 2324 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Bearcat, the crunching is Good for this project. It is just that 2 of the credit systems are poor!
About 12% of tasks report low run time. This effects the same people and it’s not transient. The same people also get less credit, but as credit is balanced between 2 crunchers, this usually only means a credit loss of 25%. Wingmen without run time credit issues will only lose about 3% credit overall. So if we reduce the number of crunchers with poor runtime it will help the credit of everyone doing CEP2. |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
well, apart from being unable to run cep2 right now, this project is #1 of my all-time credits/cpu-hour chart:
cep2 (linux): 72.40 (after 8 units back in june) hcmd2 (win): 70.30 cep1 (win): 68.69 dddt2 (win): 67.74 c4cw (win): 67.14 beta (win): 64.38 hcc (win): 60.69 nrfw (win): 55.68 hpf2 (win): 54.11 hfcc (win): 52.07 iads (win): 51.88 faah (win): 51.78 dddt1 (win): 41.13 (was only 1 wu) |
||
|
RaymondFO
Veteran Cruncher USA Joined: Nov 30, 2004 Post Count: 561 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
I use Kubuntu 10.04 x64. Run times varied between 8.70h and 4.57h So far: 61 Valid, 8 PV, 3 IP 8 Errors - On two occasions 3 tasks failed at the same time; probably due to me not closing Boinc correctly after an update and doing a restart. CPU usage to runtime ratio always over 90% Q6600, 4GB DDR2, 1xGTX260-216, using 3 CPU cores for WCG (1 free for the GPU). CPU usage constantly at 100%. With no GPU running, the ratio is over 95%. I would suggest you try the x64 version; as it works for me and is faster anyway. I doubt the 64bit version will close the gap to where it should be. I know I will gain efficiency, but not where it should be. When I tried Ubuntu 64bit from 32bit version with a clean install, the efficiency did improve but a huge gap still existed. I will be installing Debian tomorrow as I want to finish crunching the remaining WU's on the normally used hard drive (Ubuntu 10.10) for that computer. I am having difficulty downloading Dotsck/UX. Anyone have a good site location for Intel quad chip? [Edit 1 times, last edit by RaymondFO at Oct 16, 2010 7:20:09 PM] |
||
|
X-Files 27
Senior Cruncher Canada Joined: May 21, 2007 Post Count: 391 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Update on my previous post:
----------------------------------------Recompiling the kernel didn't proive to have gain in performance. But what i noticed was that, running linux without x-window helps gpu cuda apps. ubuntu lucid desktop (wcg & seti) - wcg cpu time diffence: >30mins ubuntu maverick server (wcg & seti) - wcg cpu time diffence: <30mins ubuntu maverick server (wcg) - wcg cpu time diffence: <15mins all of my test have this on my sysctl.conf: kernel.shmmax = 2147483648 vm.swappiness = 0 vm.vfs_cache_pressure = 50 ![]() ![]() |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
herna, Boinc Manager, Advanced Preferences, Processor Usage, "While Processor Usage is less than", set this to Zero (0). slade52, that is about right; a loss of up to 30min is not too bad. It is partially due to system overhead and partially your usage. Other projects are similar (not all, but some both here and outside WCG). It is actually a good bit more than the difference on other projects. Most of them are a minute or less per clock time hour. None the less, it does not seem all that important. It could be that I have my cpu usage higher than most folks, I do not know. Right now cpu usage is at 100% for all 4. If I get to doing something that actually uses serious cpu then cpu usage actually drops as boinc slows down. Seems like a pretty robust system of control that boinc uses. Real good way to test your system for robustness too. |
||
|
nanoprobe
Master Cruncher Classified Joined: Aug 29, 2008 Post Count: 2998 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
I am having difficulty downloading Dotsck/UX. Anyone have a good site location for Intel quad chip? Have you tried here?
In 1969 I took an oath to defend and protect the U S Constitution against all enemies, both foreign and Domestic. There was no expiration date.
![]() ![]() |
||
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Based on some comments, I've increased the WTS (Write to Disk) to 36000 seconds as an experiment. Expecting not to see any checkpoint writing on any WCG task except for CEP2 which will store away the 16 jobs segments, these by BOINC perceived as EOTs. The quad is on a UPS so not worried too much of power outs plus the Linux is currently operating very stable.
----------------------------------------NB: Changing WTS does not take effect until a new task or the client is restarted i.e. wont know until probably tomorrow night if this has any performance improvement at all.
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
The product of running this 10 hour WTS interval is, that with 2 CEP2 and anything else on the other 2 cores of the quad indeed only CEP2 logs checkpoints. The efficiency improved only marginally... less than 1 percent.
----------------------------------------3048 WCG 18-10-2010 05:48 [checkpoint_debug] result E200461_808_A.24.C19H14N2S2Si.317.0.set1d06_1 checkpointed 3049 WCG 18-10-2010 05:53 [checkpoint_debug] result E200461_808_A.24.C19H14N2S2Si.317.0.set1d06_1 checkpointed 3050 WCG 18-10-2010 06:01 [checkpoint_debug] result E200460_620_A.24.C19H12N2OSSe.133.4.set1d06_1 checkpointed 3051 WCG 18-10-2010 06:07 [checkpoint_debug] result E200460_620_A.24.C19H12N2OSSe.133.4.set1d06_1 checkpointed 3052 WCG 18-10-2010 06:12 [checkpoint_debug] result E200460_620_A.24.C19H12N2OSSe.133.4.set1d06_1 checkpointed 3053 WCG 18-10-2010 06:18 [checkpoint_debug] result E200460_620_A.24.C19H12N2OSSe.133.4.set1d06_1 checkpointed 3054 WCG 18-10-2010 06:23 [checkpoint_debug] result E200460_620_A.24.C19H12N2OSSe.133.4.set1d06_1 checkpointed 3055 WCG 18-10-2010 06:29 [checkpoint_debug] result E200460_620_A.24.C19H12N2OSSe.133.4.set1d06_1 checkpointed 3056 WCG 18-10-2010 06:35 [checkpoint_debug] result E200460_620_A.24.C19H12N2OSSe.133.4.set1d06_1 checkpointed 3057 WCG 18-10-2010 06:41 [checkpoint_debug] result E200460_620_A.24.C19H12N2OSSe.133.4.set1d06_1 checkpointed 3058 WCG 18-10-2010 06:53 [checkpoint_debug] result E200460_620_A.24.C19H12N2OSSe.133.4.set1d06_1 checkpointed 3059 WCG 18-10-2010 07:01 [checkpoint_debug] result E200460_620_A.24.C19H12N2OSSe.133.4.set1d06_1 checkpointed 3060 WCG 18-10-2010 07:21 [checkpoint_debug] result E200461_808_A.24.C19H14N2S2Si.317.0.set1d06_1 checkpointed 3061 WCG 18-10-2010 07:27 [checkpoint_debug] result E200461_808_A.24.C19H14N2S2Si.317.0.set1d06_1 checkpointed 3062 WCG 18-10-2010 07:32 [checkpoint_debug] result E200461_808_A.24.C19H14N2S2Si.317.0.set1d06_1 checkpointed 3063 WCG 18-10-2010 07:34 [checkpoint_debug] result E200460_620_A.24.C19H12N2OSSe.133.4.set1d06_1 checkpointed 3064 WCG 18-10-2010 07:37 [checkpoint_debug] result E200461_808_A.24.C19H14N2S2Si.317.0.set1d06_1 checkpointed Word is, that the techs are developing code and instructions how to set and automate the limitation of the number of concurrent CEP2 jobs running, so for instance only 2 on a quad, 1 on a duo, 4 on an octo and so on. We'll see how that pans out after the next Beta, which as was reported is planned to be for Linux too.
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
RaymondFO
Veteran Cruncher USA Joined: Nov 30, 2004 Post Count: 561 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
I will be installing Debian tomorrow as I want to finish crunching the remaining WU's on the normally used hard drive (Ubuntu 10.10) for that computer. I am having difficulty downloading Dotsck/UX. Anyone have a good site location for Intel quad chip? I attempted to install Debian, however that was not successful. The image on the screen was very badly distorted and when I tried to switch back to Ubuntu on the other hard drive, I was almost not able to stabilize the image. Every time I attempt to download Dotsck/UX, I get a "failure to mount" error message. I have one final idea, and that is to re-flash the motherboard. Any ideas if anyone believe that could work, or am I just wasting my time? |
||
|
Somervillejudson@netscape.net
Veteran Cruncher USA Joined: May 16, 2008 Post Count: 1065 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
What a pain!
|
||
|
|
![]() |