| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: Android Support Forum
Thread: Work Unit resets maybe every hour |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 11
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Added boinc to my new phone and after running other projects for a while I start working on ebola, which I have my old phone doing on both cores at the moment. After it was working for a few hours I decide to check it and see it had started a new work unit, decide to check the website to see how long it took and find that nothing happened. Go back an hour later and find it had started over again. This time the website said it had errored. Disappointing but whatever. Check later and come to the conclusion that it is just resetting over and over again. Sorry this is a bit of a story but what can I do to fix this?
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
It depends on the settings to help minimize fails and restarts... seeing the same thing at times
 . Suggest the web device profile 'Default' is set to ensure that 'Leave application in memory when suspended' [For customization, Android would only listen to the Default device profile, if it does at all]. Also allow computing up to 50% for 'Pause at CPU usage above' [tick the 'Show advanced preferences and controls' if you don't see it]. Ram limit set to 70% and min. spare storage to 0.5GB. Only running max cores minus 1 i.e. my 4 core only runs 3 Ebola at this time, mostly only 2 if the fail rate jumps. Also, it seems the Ebola app is more stable than the FightAIDS, so not running latter at this time on the mobile device. Untick the 'Pause computation when screen is on', else nothing much will happen if on charger and using the device, which is how I work the tablet most of the times. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I am running 1 core on a dual core processor
----------------------------------------pause cpu usage is set at 50% Ram limit was at 50%, I'll up it to 70% like you are suggesting Pause with screen on is turned off. I have a Droid Maxx if it matters. I'll see if it helps tonight when I plug it back in, thanks. [Edit 1 times, last edit by Former Member at Mar 24, 2015 7:30:33 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
It reset again. Maybe around the two and a half to three hour mark. After seeing that it did reset I went to bed and it is now showing seven and a half hours and 100% but hasn't uploaded. My phone worked on an Einstein project for a while so it is possible it didn't reset again.
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Had one reset myself, but on 3 cores concurrent, the 2 others just fell back to last checkpoint [WCG only].
Can you check if the event log shows as if the client was just restarted...scroll to bottom, which should be short. Mine says it was today at 07:34, though the device has not been booted for weeks. Suspect it's one of 2 storage/process/ram cleaner apps I've got doing this, though both are set to exclude BOINC [no way to exclude the science apps themselves... they don't show in the process manager either] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Only goes back to about 20 minutes ago in the event and gui log
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
whenever I get a phone call. This tends to blow away ~2 hours of CPU time. It is taking two days or more to get a WU done.
|
||
|
|
numbermaniac
Cruncher Australia Joined: Mar 28, 2014 Post Count: 46 Status: Offline Project Badges: 
|
The problem is Ebola barely checkpoints. FAAH checkpoints every few minutes, but Ebola will usually go 2+ hours without checkpoint. IMO FAAH is way more stable and more'friendly' than Ebola tasks.
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
It does at the same rate as FAHV, every completed dock run. Since calculations are non-deterministic, a docking attempt for optimal energy can take longer, it can be shorter.
Here's a piece of log from one running on a device that is going at 1300Mhz [normally 1500]... about every 42 minutes this one: World Community Grid 6/18/2015 2:37:09 PM [cpu_sched] Starting task OET1_0000966_xZAGP-S_rig_75304_0 using oet1 version 719 in slot 0 World Community Grid 6/18/2015 3:18:28 PM [checkpoint] result OET1_0000966_xZAGP-S_rig_75304_0 checkpointed World Community Grid 6/18/2015 4:00:23 PM [checkpoint] result OET1_0000966_xZAGP-S_rig_75304_0 checkpointed World Community Grid 6/18/2015 4:42:58 PM [checkpoint] result OET1_0000966_xZAGP-S_rig_75304_0 checkpointed |
||
|
|
numbermaniac
Cruncher Australia Joined: Mar 28, 2014 Post Count: 46 Status: Offline Project Badges:
|
Well on my 1.8GHz phone, Ebola tasks only checkpoint ever 2 - 3 hours or so, while FAAH checkpoints every 2-5 minutes. Ebola is also misleading because after it reaches 100% it stays there for several hours. FAAH finishes as soon as it reaches 100%.
|
||
|
|
|