| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: Android Support Forum
Thread: Android Phone Appears to Produce "Invalid" Results 2 Out of 3 Times |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 37
|
|
| Author |
|
|
adamgreb
Cruncher Joined: Apr 27, 2007 Post Count: 11 Status: Offline Project Badges: 
|
So far the first WU is valid with only one core running
|
||
|
|
adamgreb
Cruncher Joined: Apr 27, 2007 Post Count: 11 Status: Offline Project Badges:
|
Two valids and an invalid. I will have to continue to investigate.
|
||
|
|
ish451
Cruncher Joined: Aug 7, 2013 Post Count: 3 Status: Offline Project Badges: 
|
Running a Samsung Galaxy S4 and am also producing a lot of invalid results. Out of 50 results 34 of them were invalid (68% failure rate). I was thinking that maybe the suspend when not plugged in was causing the issue so I let it crunch even when on battery (I have 2 spare battery's so no big deal) but still even those results are invalid!
Even when the phone is plugged in at night, under a fan so it never reaches above 85F, and not under any use for 8 hours still ~65% of the results are invalid. This is when running on all 4 cores haven't tried using only 3, 2 or 1 core but if 65% of my time is wasted, there is no point on crunching numbers on my phone. All are 706 version. If needed I can post any of my invalid 38 status logs. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hi ish451,
Computing on phones is new and we are still learning what the designers have done. Please try running on 1 core only while plugged in. If that works, it means your phone designers have added some tricks to their design that we have not learned about yet. Lawrence |
||
|
|
adamgreb
Cruncher Joined: Apr 27, 2007 Post Count: 11 Status: Offline Project Badges:
|
Two valids and an invalid. I will have to continue to investigate. Even with the one core running I'm seeing invalid results. Of the 4 units yesterday. One is valid, 2 are invalid, and 1 is pending verification. Is there anything of value I can provide to the community? Looking at the other people processing the workunit, others errored out or produced invalid results so at least I'm not the only one. This is one of the invalid units: <core_client_version>7.2.8</core_client_version> <![CDATA[ <stderr_txt> INFO: No state to restore. Start from the beginning. [05:16:03] Number of tasks = 7 [05:16:03] Starting task 0,CPU time is 0.000000. [05:16:03] ./ZINC08424714.pdbqt size = 28 8 ../../projects/www.worldcommunitygrid.org/fahv.x3ZT4_B_IN_LEDGF.pdbqt size = 2853 0 [05:43:07] Finished task #0 cpu time used 1617.230000 [05:43:07] Starting task 1,CPU time is 1617.230000. [05:43:07] ./ZINC08424715.pdbqt size = 29 7 ../../projects/www.worldcommunitygrid.org/fahv.x3ZT4_B_IN_LEDGF.pdbqt size = 2853 0 Quit requested: Exiting [05:52:02] Number of tasks = 7 [05:52:02] Starting task 1,CPU time is 1617.230000. [05:52:02] ./ZINC08424715.pdbqt size = 29 7 ../../projects/www.worldcommunitygrid.org/fahv.x3ZT4_B_IN_LEDGF.pdbqt size = 2853 0 Quit requested: Exiting [05:53:53] Number of tasks = 7 [05:53:53] Starting task 1,CPU time is 1617.230000. [05:53:53] ./ZINC08424715.pdbqt size = 29 7 ../../projects/www.worldcommunitygrid.org/fahv.x3ZT4_B_IN_LEDGF.pdbqt size = 2853 0 [06:17:47] Finished task #1 cpu time used 1427.040000 [06:17:47] Starting task 2,CPU time is 3044.270000. [06:17:47] ./ZINC08424718.pdbqt size = 33 10 ../../projects/www.worldcommunitygrid.org/fahv.x3ZT4_B_IN_LEDGF.pdbqt size = 2853 0 [06:57:52] Finished task #2 cpu time used 2357.350000 [06:57:52] Starting task 3,CPU time is 5401.620000. [06:57:52] ./ZINC08424723.pdbqt size = 24 6 ../../projects/www.worldcommunitygrid.org/fahv.x3ZT4_B_IN_LEDGF.pdbqt size = 2853 0 [07:10:14] Finished task #3 cpu time used 731.290000 [07:10:14] Starting task 4,CPU time is 6132.910000. [07:10:14] ./ZINC08424730.pdbqt size = 24 7 ../../projects/www.worldcommunitygrid.org/fahv.x3ZT4_B_IN_LEDGF.pdbqt size = 2853 0 [07:29:59] Finished task #4 cpu time used 1179.500000 [07:29:59] Starting task 5,CPU time is 7312.410000. [07:29:59] ./ZINC08424733.pdbqt size = 24 5 ../../projects/www.worldcommunitygrid.org/fahv.x3ZT4_B_IN_LEDGF.pdbqt size = 2853 0 [07:44:31] Finished task #5 cpu time used 867.375000 [07:44:31] Starting task 6,CPU time is 8179.785000. [07:44:31] ./ZINC08424736.pdbqt size = 26 8 ../../projects/www.worldcommunitygrid.org/fahv.x3ZT4_B_IN_LEDGF.pdbqt size = 2853 0 [08:04:47] Finished task #6 cpu time used 1206.580000 08:04:47 (9285): called boinc_finish </stderr_txt> ]]> |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hi adamgreb,
You have reached what would be my personal threshold for quitting in disgust. Hardware or software? If any results are valid, I would suspect hardware, but I would wonder if my result was being validated against hardware producing a slightly different result. There is some research into that possibility. ARM computing for BOINC is very new. Lawrence |
||
|
|
captainjack
Advanced Cruncher Joined: Apr 14, 2008 Post Count: 147 Status: Offline Project Badges: 
|
During the beta testing for Android devices and during the early parts of regular processing, my Nexus 7 produced almost 100% valid results. Now days, it is producing mostly invalid results.
And yesterday, it processed work unit FAHV_ x3NF6_ B_ IN_ Y3_ 0056582_ 0747_ 4-- along with 11 of my closest friends. My task was sent to me on 8/15 at 10:22:48 and was returned on 8/15 at 19:33:46; 9 hours 11 minutes later. The status of mine and two others was "Too late". The other 9 people who processed tasks for this work unit were given a status of "Error". Something is definitely amiss. The Nexus 7 will be on vacation for a while. Might as well save some kilowatts. Please let us know when improvements are ready for testing. Thanks, CaptainJack |
||
|
|
ish451
Cruncher Joined: Aug 7, 2013 Post Count: 3 Status: Offline Project Badges:
|
As suggested above by lawrence, for the past 5 days I have been crunching on a single core of my 4-core galaxy S4. Although the results are trickling in 4x slower... the amount of valids vs invalids is still about the same.
----------------------------------------Approximately 50-60% of my results are invalid. And if you look at any of the global results (the valids or invalids) they all have a large amount of errors and invalids associated with them. For example out of 6 replicates 2 will be valid, 2-3 will be invalid and the rest will be errors! If all this data (invalids) is helping WCG/IBM and FAAH run diagnostics/trouble shoot then I am happy to keep crunching (attempting to crunch) in the spirit of "beta testing" Otherwise it might be time to uninstall the android client while they work on a fix. [Edit 1 times, last edit by ish451 at Aug 17, 2013 2:35:45 PM] |
||
|
|
ish451
Cruncher Joined: Aug 7, 2013 Post Count: 3 Status: Offline Project Badges:
|
bump.....
any info? |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hi ish451,
No info so far, but if this is bothering you then stop. I expect problems to eventually be tracked down and fixed, but I am not hearing about any immediate fixes in the pipeline. Don't break your heart on new software. Lawrence |
||
|
|
|