| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: Computing for Clean Water Forum
Thread: Repeating errors on 64-bit server |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 31
|
|
| Author |
|
|
armstrdj
Former World Community Grid Tech Joined: Oct 21, 2004 Post Count: 695 Status: Offline Project Badges: 
|
minaev,
If you are able to test this I would be curious if you see the issue if you limit the number of cores being used by BOINC to say 50% or 25%. Thanks, armstrdj |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
armstrdj,
I have created a new profile for this server, limited it to this project only and set 'On multiprocessors, use' to 50% of CPUs. Thanks. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
50% restriction doesn't help. Decreased to 25%.
|
||
|
|
sk..
Master Cruncher http://s17.rimg.info/ccb5d62bd3e856cc0d1df9b0ee2f7f6a.gif Joined: Mar 22, 2007 Post Count: 2324 Status: Offline Project Badges: 
|
Any other observations; continuous increase in RAM usage for example?
Xeon E5540's are HT capable, so I expect you have HT enabled. You might want to try it with HT off. I would be inclined to use the report tasks immediately cc_config switch to see if that helps free something up after the tasks finish. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
No, I see no signs of leaks of any kind. Memory, file descriptors, sockets, all resources seem to be released correctly.
Yes, hyperthreading is enabled, but, frankly speaking, I'm rather reluctant about turning it off. With 25% of CPUs dedicated to the calculations, results still often contain errors. I think that periods of errors are more or less regularly interspersed with periods of successful results now. I will let it run for a couple of days longer, but I'm getting more and more inclined to leaving C4CW and moving on to other projects... After all, the server can do more than running on a quarter of its resources :) |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Given that there are a select assortment of very hi core count devices crunching at WCG to include a 127 core Linux monster http://boincstats.com/stats/host_stats.php?pr=wcg&st=0, it could be something in the OS/Library front. I wonder how we could force your client to receive the 32 bit version of Clean Water **. The performance differential is kind of marginal depending on device, even negative for some.
** One of the older 32 clients would only fetch 32 bit sciences could be one way. The client has no speed impact itself... it's just a science app thread manager. Of course, the question is how far does one want to push finding the solution / cause ?!? --//-- |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Results are still unstable. There may be twenty errors followed by eight successful results, then four segfaults, then sixteen successful ones and so on.
Thank you all for your help, but I thought I'll better switch to other projects. As of yet, not a single task from other projects has failed. |
||
|
|
retsof
Former Community Advisor USA Joined: Jul 31, 2005 Post Count: 6824 Status: Offline Project Badges:
|
I am running 64 bit C4CW (shows correctly on task list) on a 3.3 GHz non overclocked AMD Phenom II X6 and all are ending correctly with no errors. The projects on that computer are also mixed with HCC and CEP, a few betas and an occasional high priority -2 item.
----------------------------------------The results bonus on these seems always to be about 5% more than I have claimed since the wingman is probably running 32 bit. This is NOT hyperthreaded.
SUPPORT ADVISOR
----------------------------------------Work+GPU i7 8700 12threads School i7 4770 8threads Default+GPU Ryzen 7 3700X 16threads Ryzen 7 3800X 16 threads Ryzen 9 3900X 24threads Home i7 3540M 4threads50% [Edit 1 times, last edit by retsof at Jul 6, 2011 8:47:32 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I would just like to add that yesterday I switched from HCC to C4CW.
I run an overclocked Athlon II X4 630, @3.5ghz and have done so for 18 months with very few BSOD's. Yesterday I had multiple crashes within a 2 hour period.  In the end I have to reduce the FSB to reduce the memory overclock...........it then ran stable for 5-6 hours.....I will test further today. I run Win 7 64 bit. Whilst I am no expert on Overclocking my commonsense conclusion is that C4CW is in someway exceptionally hard on RAM. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
One or the other science are known to cause higher CPU temps, so would not be surprised this applies to RAM too. OCers often have these copper heatsinks. Used to have DDR2 ram with these pre-mounted for better cooling, but saw them as ''after market'' or whatever that term is in English for clipon.
--//-- |
||
|
|
|