| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: Computing for Sustainable Water Forum
Thread: Computing for Sustainable Water Problems Thread |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 254
|
|
| Author |
|
|
Ingleside
Veteran Cruncher Norway Joined: Nov 19, 2005 Post Count: 974 Status: Offline Project Badges: 
|
I'm seeing something unexpected with the quota system for my CFSW WUs. Neither has returned a CFSW WU yet, so I was expecting a quota of 2 WUs per core to start with. My 4-core i5 should have a quota of 8 WUs total, but has 11 downloaded. My 1-core AMD should have only 2 WUs, but has 3 downloaded (and its Messages is saying "...daily quota of 2 tasks"). Is someone playing about with the quota system, or does something need fixing? Edit: The 4-core i5 (no HT) now has 17 WUs downloaded and subsequently is saying "This computer has finished a daily quota of 2 tasks"  The scheduling-server only checks the quota at the beginning of a request so it's possible to get assigned multiple tasks past the quota in a work-request. Example, on a quad that's already downloaded 7, on next work-request it can download N more, bringing total to N + 7. Only on the next work-requests will (N + 7) >= 8 so you'll see the "daily quota of 2 tasks". (Not sure, but it's possible it's only checking strictly larger, so a quad-core with 8 tasks downloaded can still manage to get more work).  "I make so many mistakes. But then just think of all the mistakes I don't make, although I might." |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
The scheduling-server only checks the quota at the beginning of a request so it's possible to get assigned multiple tasks past the quota in a work-request. That's reasonable, thanks for clarifying it. However, the last 9 of the 17 WUs on the 4-core were downloaded in 7 separate requests. All were single tasks except for the final request which gained 3 new ones. That might indicate that a machine limit of 16 was in place, i.e. 4 per core. Still can't explain it! |
||
|
|
deltavee
Ace Cruncher Texas Hill Country Joined: Nov 17, 2004 Post Count: 4842 Status: Offline Project Badges: 
|
I've started getting replication 1, quorum 1 WUs. That seemed to happen quicker than with sn2s.
---------------------------------------- |
||
|
|
sk..
Master Cruncher http://s17.rimg.info/ccb5d62bd3e856cc0d1df9b0ee2f7f6a.gif Joined: Mar 22, 2007 Post Count: 2324 Status: Offline Project Badges: 
|
!
----------------------------------------[Edit 1 times, last edit by skgiven at Jul 18, 2012 9:15:06 PM] |
||
|
|
nanoprobe
Master Cruncher Classified Joined: Aug 29, 2008 Post Count: 2998 Status: Offline Project Badges: 
|
I had to restart this machine again after changing out some memory and I had 1 task error again after being 92.5% completed. Same error as before. I guess I won't be doing any more restarts on this box with tasks that are in progress.
----------------------------------------Result Log Result Name: cfsw_ 0020_ 00020131_ 0-- <core_client_version>7.0.15</core_client_version> <![CDATA[ <message> The system cannot find the path specified. (0x3) - exit code 3 (0x3) </message> <stderr_txt> [11:01:26] INFO:Beginning simulation: 1990:240:2016285312 [11:11:36] INFO: Finished tick number 4 [11:19:59] INFO: Finished tick number 9 [11:27:19] INFO: Finished tick number 14 [11:36:52] INFO: Finished tick number 19 [11:43:47] INFO: Finished tick number 24 [11:53:03] INFO: Finished tick number 29 [12:00:58] INFO: Finished tick number 34 [12:09:00] INFO: Finished tick number 39 [12:18:00] INFO: Finished tick number 44 [12:24:37] INFO: Finished tick number 49 [12:34:09] INFO: Finished tick number 54 [12:41:32] INFO: Finished tick number 59 [12:50:25] INFO: Finished tick number 64 [12:58:48] INFO: Finished tick number 69 [13:05:59] INFO: Finished tick number 74 [13:15:23] INFO: Finished tick number 79 [13:22:16] INFO: Finished tick number 84 [13:31:33] INFO: Finished tick number 89 [13:39:27] INFO: Finished tick number 94 [13:47:26] INFO: Finished tick number 99 [13:56:26] INFO: Finished tick number 104 [14:03:03] INFO: Finished tick number 109 [14:12:33] INFO: Finished tick number 114 [14:19:57] INFO: Finished tick number 119 [14:28:48] INFO: Finished tick number 124 [14:37:11] INFO: Finished tick number 129 [14:44:20] INFO: Finished tick number 134 [14:53:46] INFO: Finished tick number 139 [15:00:39] INFO: Finished tick number 144 [15:09:53] INFO: Finished tick number 149 [15:17:46] INFO: Finished tick number 154 [15:25:45] INFO: Finished tick number 159 [15:34:44] INFO: Finished tick number 164 [15:41:21] INFO: Finished tick number 169 [15:50:51] INFO: Finished tick number 174 [15:58:12] INFO: Finished tick number 179 [16:07:03] INFO: Finished tick number 184 [16:15:25] INFO: Finished tick number 189 [16:22:34] INFO: Finished tick number 194 [16:31:57] INFO: Finished tick number 199 [16:38:50] INFO: Finished tick number 204 [16:48:03] INFO: Finished tick number 209 [16:55:55] INFO: Finished tick number 214 [17:03:56] INFO: Finished tick number 219 [17:24:50] DEBUG: Restarting from checkpoint. [17:24:50]PctComplete = 0.925 [17:24:50]ticks:currentTick:modules:currentModule:restart:seed240:222:6:0:0:16391 </stderr_txt>
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.
  |
||
|
|
Jason1478963
Senior Cruncher United States Joined: Sep 18, 2005 Post Count: 295 Status: Offline Project Badges:
|
Is anybody seeing unusually long run times? For instance, 26 hours/wu and not getting nearly the estimated credit. Other boxes identical except for cpu are finishing in about 4.5 - 6 hrs. The boxes in question are showing 99 to 100 % efficiency, but still are taking over a day to complete. A 6 core running at 1.8 Ghz in the same setup is finishing in around 6 hours, another 6 core running at 2.4 Ghz is finishing in about 4.5 hours, but these quad cores running at 2.3 Ghz are taking 24 to 26 hours per wu.
----------------------------------------Any ideas? Ubuntu 10.10 X64 4 GB Ram Dual Opteron 2356's at 2.3 Ghz Boinc Client 6.10.56  |
|||
|
||||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Is anybody seeing unusually long run times? For instance, 26 hours/wu and not getting nearly the estimated credit. Other boxes identical except for cpu are finishing in about 4.5 - 6 hrs. The boxes in question are showing 99 to 100 % efficiency, but still are taking over a day to complete. A 6 core running at 1.8 Ghz in the same setup is finishing in around 6 hours, another 6 core running at 2.4 Ghz is finishing in about 4.5 hours, but these quad cores running at 2.3 Ghz are taking 24 to 26 hours per wu. Any ideas? Ubuntu 10.10 X64 4 GB Ram Dual Opteron 2356's at 2.3 Ghz Boinc Client 6.10.56 Seems to like cpu's with larger cache sizes on the systems I looked at. I do see some of my older systems (Paxville/Potomac) clocking up 12 to 13 hour tasks on average. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Wow David. Ruby..... Very impressed.
|
||
|
|
Ingleside
Veteran Cruncher Norway Joined: Nov 19, 2005 Post Count: 974 Status: Offline Project Badges:
|
Not a problem, just an observation: On Linux these tasks are >50% faster; 3.1h vs 4.8h on Ubuntu_x64 vs Win_x64 systems. Hardly any variation in run time's either, so this should be fairly accurate. This is nice because the C4CW tasks are about the same on Win and Linux - so C4CW on Win and CFSW on Linux... Well, in my opinion a speed-difference of >50% in advantage of either OS to me indicates a problem. Assuming this is benchmarked on the same system, or two identical systems except for different OS, the speed should be roughly equal between OS, as you're seeing with C4CW. Such a big difference can indicate a problem with the application, either due to wrong compiler-choices, linking against sub-optimal libraries or using a mediocre compiler. Also, while Linux being slower than Windows wouldn't matter much for the project, Windows being slower on the other hand has a much bigger impact. This is because Windows is 81.45% of WCG's production while Linux is 14.10% and Mac is 4.44% (based on BoincStats RAC). Making it simple, if speed-difference is 50% and the project would take 12 months with the "fast" application, if Linux is slow the project will take an additional 26 days for a total project-time close to 13 months. If windows is slow on the other hand the project will take an additional 149 days, roughly 5 months, for a total project-time of nearly 17 months. Oh, and just to include OSX also, if Mac is the slow one this would add an additional 8 days to the total project-time. "I make so many mistakes. But then just think of all the mistakes I don't make, although I might." |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
on my asus laptop i7
+w7 64bit running in 3 hours on my asus desktop i5 2500 +w7 64 bit runnining in 2 hours on my dell i7 +w7 64 bit running in 3.5 hours no linux here to show running time |
||
|
|
|