| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: Drug Search for Leishmaniasis Forum
Thread: WU with a problem? |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 21
|
|
| Author |
|
|
cjslman
Master Cruncher Mexico Joined: Nov 23, 2004 Post Count: 2082 Status: Offline Project Badges: 
|
Just an update on the WUs that I reported as inconclusive... I was lucky and my WU was declared Valid
---------------------------------------- , but two of the wingmen were not so lucky  DSFL_ 00000093_ 0000015_ 0332_ 3-- 624 Valid DSFL_ 00000093_ 0000015_ 0332_ 2-- 624 Valid DSFL_ 00000093_ 0000015_ 0332_ 0-- 619 Invalid DSFL_ 00000093_ 0000015_ 0332_ 1-- 619 Invalid CJSL |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
After reading the previous replies I suddenly remembered that I actually download these WU:s using one computer, and then perform the calculations on another unit. So, if the CPU type do not match, then the calculations differ, as the CPU type may be be different.
----------------------------------------I was wondering if the CPU type could be excluded from the comparison software, so that The Validator not simply assign Invalid based upon irrelevant data? [Edit 1 times, last edit by Former Member at Jan 7, 2012 3:51:56 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Huh? In the first instance the validator compares the data if the results in the quorum sufficiently match. That's the only things of interest, giving a hoots if OS B or OS A did the calculations. The OS/CPU identifiers and a few more features of the host are used during assignment, during any connection, but particularly during problem analysis.
----------------------------------------If you sneakernetted the tasks, then it's your responsibility to ensure that the same OS and or CPU group does the calculation, noting that the OS/CPU is identified for the device that reports the result. Generally though if a second computer reports the tasks or connects , there's a great chance that the server will respond by telling that the original assignments were to a different host and bin them all. tell the client to abort them and send new work. In your case it reads like you uploaded also with the 2nd computer. That's asking for trouble. Somehow the servers did though not get alerted, so you may have cloned the client to device sufficiently equal but not entirely. It's for you to tell what you exactly did. --//-- [Edit 1 times, last edit by Former Member at Jan 7, 2012 4:23:45 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
If the scientific results actually do differ simply depending on the CPU type, also for working and correct machines, then this seems to be serious, and personally I would not find that acceptable without a proper investigation. My thought is that the Beta Runs are also used in this respect?
The machine that I use for loading WU:s have a double-boot, and is very carefully configured not to upset the project server. After loading the WU:s the machine take part in the WCG as usual, using a different computer name on the normal boot. The operating system on this machine, and several other of my machines, has been installed and configured only to take part in the WCG, and this is how I normally participate in the WCG effort. The double-boot is a fix to "work-around" the "Minimum memory requirement", that was a problem for me already in 2007. This was also found to actually be a bug in this thread, even though I have not taken the effort of responding and actually describing the bug/problem in detail, due to possibly lack of interest on the system-part that will be needed to fix it. A network error found and reported in 2007 stayed with me at least for all 5.** versions of the BOINC client. The proper fix was never installed, though newer operating systems (Windows 7) may not need it any more. I also have a slow machine, that no longer can meet any kind of deadline on any WCG project. The problem with this is that the WCG is a volunteer-project, and running the WCG on an old and no longer used machine was my start to participating in 2007. Possibly it is also so for many others; in my case 25,000 faah-WU:s followed. I have so forth failed to properly describe the importance of accepting these older and slower machines; it do not at all relate to the overall performance seen from the project server's point of view. If I wish to cut-down on the computing, this is unfortunately best done on the fastest machines. Since the CPU-type issue will likely not go away, I have changed the project for the target machine (so the problem go away). |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
"If the scientific results actually do differ simply depending on the CPU type, also for working and correct machines, then this seems to be serious..." The knreed [Jan 5, 2012 2:22:45 PM] post touches on the top-level state of things with the AMD FX chip as a take-off point but generally applies to machines, OS, CPU types vis-a-vis the applications.; ------ edit1:2012.01.09Mo.0019.utc: spell-check and readability improvements. [Edit 1 times, last edit by Former Member at Jan 9, 2012 12:19:22 AM] |
||
|
|
pvh513
Senior Cruncher Joined: Feb 26, 2011 Post Count: 260 Status: Offline Project Badges: 
|
I also had quite a number of WUs declared invalid because I ran them with version 6.19 and the wingmen with 6.24. I never had any invalid WUs before the version change, so it seems to me that validating WUs with different versions of the code automatically fails, or at least has a significantly higher chance of failing. It has been mentioned in this thread that from now on AMD results will be matched to AMD, and Intel to Intel. But is there also code in place to assure that the same version of the code will be used in the validation?
|
||
|
|
knreed
Former World Community Grid Tech Joined: Nov 8, 2004 Post Count: 4504 Status: Offline Project Badges: 
|
The BOINC server has recently added 'homogenous app version' that we are still testing in conjunction with BOINC. When we update our server code, that will be included as part of the update. In the meantime, we do not need to have the 6.19 version active anymore so we have deprecated it.
|
||
|
|
nanoprobe
Master Cruncher Classified Joined: Aug 29, 2008 Post Count: 2998 Status: Offline Project Badges: 
|
The BOINC server has recently added 'homogenous app version' that we are still testing in conjunction with BOINC. When we update our server code, that will be included as part of the update. In the meantime, we do not need to have the 6.19 version active anymore so we have deprecated it. Thanks for the update. Out of curiosity I looked and noticed I had 3 DSFL vs. 6.19 WUs declared invalid and all had been compared to vs. 6.24 WUs which were all valid.
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.
  |
||
|
|
pvh513
Senior Cruncher Joined: Feb 26, 2011 Post Count: 260 Status: Offline Project Badges:
|
Indeed, thanks for the update. Incidentally, I also lost another three WUs today because of this issue. It's good to know that this will be fixed in the future.
|
||
|
|
gb009761
Master Cruncher Scotland Joined: Apr 6, 2005 Post Count: 3010 Status: Offline Project Badges: 
|
Well, unless two "No Reply's" show up PDQ, it looks as though I'll also be loosing out with this issue...
----------------------------------------DSFL_ 00000091_ 0000056_ 0591_ 1-- Inconclusive 30/12/11 16:04:30 08/01/12 21:46:31 9.46 196.1 / 0.0 Workunit Status Project Name: Drug Search for Leishmaniasis Created: 12/29/2011 04:05:54 Name: DSFL_00000091_0000056_0591 Minimum Quorum: 2 Replication: 3 Result Name App Version Number Status Sent Time Time Due / Return Time CPU Time (hours) Claimed/ Granted BOINC Credit DSFL_ 00000091_ 0000056_ 0591_ 2-- 624 Inconclusive 09/01/12 16:06:14 11/01/12 05:00:32 7.03 149.1 / 0.0 DSFL_ 00000091_ 0000056_ 0591_ 0-- - No Reply 30/12/11 16:05:27 09/01/12 16:05:27 0.00 0.0 / 0.0 DSFL_ 00000091_ 0000056_ 0591_ 1-- 619 Inconclusive 30/12/11 16:04:30 08/01/12 21:46:31 9.46 196.1 / 0.0 DSFL_ 00000091_ 0000056_ 0591_ 3-- - Waiting to be sent — — 0.00 0.0 / 0.0 DSFL_ 00000091_ 0000056_ 0602_ 0-- Inconclusive 30/12/11 16:03:47 09/01/12 15:37:01 8.46 173.4 / 0.0 Workunit Status Project Name: Drug Search for Leishmaniasis Created: 12/29/2011 04:05:51 Name: DSFL_00000091_0000056_0602 Minimum Quorum: 2 Replication: 3 Result Name App Version Number Status Sent Time Time Due / Return Time CPU Time (hours) Claimed/ Granted BOINC Credit DSFL_ 00000091_ 0000056_ 0602_ 3-- - In Progress 10/01/12 07:14:21 14/01/12 07:14:21 0.00 0.0 / 0.0 DSFL_ 00000091_ 0000056_ 0602_ 2-- 624 Inconclusive 09/01/12 16:00:51 10/01/12 07:10:51 7.07 109.5 / 0.0 DSFL_ 00000091_ 0000056_ 0602_ 0-- 619 Inconclusive 30/12/11 16:03:47 09/01/12 15:37:01 8.46 173.4 / 0.0 DSFL_ 00000091_ 0000056_ 0602_ 1-- - No Reply 30/12/11 15:59:49 09/01/12 15:59:49 0.00 0.0 / 0.0 Hopefully not, although if either of them go from Inconclusive to Invalid, it'll be the very first time it's happened to me   |
||
|
|
|