Message boards : Number crunching : Minirosetta android 3.83
Author | Message |
---|---|
![]() Volunteer moderator Project administrator Project developer Project scientist Send message Joined: 1 Jul 05 Posts: 1018 Credit: 4,334,829 RAC: 0 |
Please post issues/bugs regarding the minirosetta android version 3.83 application in this thread. |
Sid Celery Send message Joined: 11 Feb 08 Posts: 2059 Credit: 40,463,815 RAC: 2,908 ![]() |
It seems like there are lots of v3.73 tasks in the buffer at your end. I've seen no v3.83 even download yet. ![]() ![]() |
![]() Send message Joined: 1 Dec 05 Posts: 1958 Credit: 9,099,368 RAC: 978 ![]() |
Please post issues/bugs regarding the minirosetta android version 3.83 application in this thread. And Android surpasses Windows |
![]() Volunteer moderator Project administrator Project developer Project scientist Send message Joined: 1 Jul 05 Posts: 1018 Credit: 4,334,829 RAC: 0 |
It seems like there are lots of v3.73 tasks in the buffer at your end. I've seen no v3.83 even download yet. I'll take a look. So far the success rate is high and 3.83 results are coming in. |
Sid Celery Send message Joined: 11 Feb 08 Posts: 2059 Credit: 40,463,815 RAC: 2,908 ![]() |
It seems like there are lots of v3.73 tasks in the buffer at your end. I've seen no v3.83 even download yet. Sorry, they were in progress. 2 tasks now completed, both with validation errors but not computation errors. Is there a decoy limit of 600? Strange looking output: Device Samsung Android 6 res7_binoculars_binoculars_THR7_SAVE_ALL_OUT_477348_12_0 <core_client_version>7.4.53</core_client_version> res7_binoculars_binoculars_ALA7_SAVE_ALL_OUT_477332_13_0 <core_client_version>7.4.53</core_client_version> ![]() ![]() |
![]() Volunteer moderator Project administrator Project developer Project scientist Send message Joined: 1 Jul 05 Posts: 1018 Credit: 4,334,829 RAC: 0 |
Interesting. I'll check with the researcher who owns these jobs. Looks like the job created TOO MANY models so it failed validation. |
![]() Volunteer moderator Project administrator Project developer Project scientist Send message Joined: 1 Jul 05 Posts: 1018 Credit: 4,334,829 RAC: 0 |
I updated the validator to hopefully address this issue. |
Sid Celery Send message Joined: 11 Feb 08 Posts: 2059 Credit: 40,463,815 RAC: 2,908 ![]() |
I updated the validator to hopefully address this issue. 4 res6 tasks validated ok, including one that hit 600 decoys, after running the full time. One res7 task pulled up short again. Though I probably need to let a few more complete to give it a proper chance. Different tasks going through now. ![]() ![]() |
![]() Volunteer moderator Project administrator Project developer Project scientist Send message Joined: 1 Jul 05 Posts: 1018 Credit: 4,334,829 RAC: 0 |
I updated the validator to hopefully address this issue. I'll look into that one further. It should have been ok with the limits I set in the validator. thanks! |
Sid Celery Send message Joined: 11 Feb 08 Posts: 2059 Credit: 40,463,815 RAC: 2,908 ![]() |
I updated the validator to hopefully address this issue. This was the specific res6 task res6_binoculars_binoculars_DALA6_SAVE_ALL_OUT_477313_61_0 but it may've crossed with the time you made your update The rest of my android tasks - res 6, res5 and 20res - all came in under 600 decoys and validated fine, so all looking good now. ![]() ![]() |
Usuario1_S Send message Joined: 24 Mar 14 Posts: 92 Credit: 3,059,705 RAC: 0 |
I think I can make my Huawei Y330 to work R@H, running 1 WU on 1 core, it's dual core, so that way will occupy 400MBx2 (base + 1 core WU) that's leaving like 400MB free for my use, but now I can't connect to R@H from it, tried like 10 times, even change my password, it just fails, can anyone help please ^ is from another thread I can login from my phone to the website just fine. I did have BOINC client connected to R@H a few hours before with 2 threads running, so I think it's server side problem |
Usuario1_S Send message Joined: 24 Mar 14 Posts: 92 Credit: 3,059,705 RAC: 0 |
OK, another outage passed, WUs getting normally to my PC client and I can't even connect my Y330 to R@H, I guess is my fault since it's not possible since I reporteed it, well uninstalling from it and forgetting about it then, cheers |
Usuario1_S Send message Joined: 24 Mar 14 Posts: 92 Credit: 3,059,705 RAC: 0 |
Another thing: my Tablet: Samsung Galaxy Tab4 10.1" screen, 4 Cores ARM v7 CPU, 1.5 GB RAM. R@H only work when the screen is on, the configuration doesn't have 'Protected Apps' section where you can specify to not close apps on standby mode, the onscreen manual is useless, some apps and games just run on the background as long as I don't close them properly, and I never close BOINC, I can see it it's on RAM but it won't run unless the screen is on, it says "Connect to a power source to continue processing", max on display timer is 30 mins, so it's not practical to check my tablet all the time. I have lots of time free for my tablet always connected to the charger, and I'd like it put to a good use, imagine this replicated nx100,000 worldwide? you're losing lots of CPU power, also its GPU isn't give any WUs, another wasted computing potential |
Usuario1_S Send message Joined: 24 Mar 14 Posts: 92 Credit: 3,059,705 RAC: 0 |
Also please add the option to move the BOINC to and SD Card, or the WUs to be stored there, it's key for mobile devices, the PC version can direct the storage to a different partition/drive on installation time, BOINC for Android should on Preferences |
Usuario1_S Send message Joined: 24 Mar 14 Posts: 92 Credit: 3,059,705 RAC: 0 |
I have discovered why sometimes my device will stop processing even when connected to a power source, it's the way the Power Administrator works': when the power cable is connected, it doesn't take the power from it and give the rest to recharge the battery, it recharges the battery all the time and stops when it has reached 100%, so the device takes power from the battery all the time, so BOINC works in the slim range between 100% an what the Power Admin can detect as sub-100%, discharged battery that needs to be charged, so BOINC is stopping all the time at random times, losing valuable hours of processing time a day. Better Power Admin hardware should be developed for mobile devices. Solution: Set as BOINC default preference as: 1) to use all 4 energy sources 2) Set minimum battery level to 90% This will give the 10% room that BOINC needs and probably will never pass. But anyway please tell BOINC developers to add an option to prevent the Android/OS from stopping BOINC processing, 'keep alive' option, BOINC stops eventually, even when I have disabled the 'Power Saving' option and on Android versions that support it, added it to 'Protected Apps', that prevents apps from closing after the screen is off. |
Usuario1_S Send message Joined: 24 Mar 14 Posts: 92 Credit: 3,059,705 RAC: 0 |
I just found out that BOINC for Android won't process on battery even when configured so, please fix this. I e-malied to BOINC but I think it'd have more influence if you tell them to correct this "feature". BOINC says: Computing will resume when battery charge increases to 90% (Currently 100%) |
Usuario1_S Send message Joined: 24 Mar 14 Posts: 92 Credit: 3,059,705 RAC: 0 |
This can be done and it's not hard, it's old technology, I have 2 apps that use this, both are 2 years old and never fails, 1 is CPU Stats and Netowrk Mini, for now I'm using AndSY Apps |
Sid Celery Send message Joined: 11 Feb 08 Posts: 2059 Credit: 40,463,815 RAC: 2,908 ![]() |
I just found out that BOINC for Android won't process on battery even when configured so, please fix this. I e-malied to BOINC but I think it'd have more influence if you tell them to correct this "feature". That's very odd. My Samsung S6 runs on battery above my designated charge level ![]() ![]() |
supdood Send message Joined: 3 Aug 15 Posts: 6 Credit: 190,389 RAC: 0 |
Every once in a while I've gotten a long runner, but recently I've gotten a higher percentage of them--12-24+ hours of CPU time even though preferences are set to 4 hours. Running on two android devices, they tend to get up to 98% or so and then creep along for many hours until completion. Credit is the same as if the task ran 4-5 hours of CPU time. |
Mod.Sense Volunteer moderator Send message Joined: 22 Aug 06 Posts: 4018 Credit: 0 RAC: 0 |
The post from supdood was originally posted to WUs estimated time way off to elapsed time and was written from that context. But I felt it may be more related to Android than the specific WUs, so I moved it here. Rosetta Moderator: Mod.Sense |
Message boards :
Number crunching :
Minirosetta android 3.83
©2024 University of Washington
https://www.bakerlab.org