| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: Android Support Forum
Thread: Is boinc for android lollipop ready? |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 34
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
An extensive article at techradar http://www.techradar.com/news/phone-and-commu...-and-rumours-1091500#null discusses the coming new os release. My tab is expected to be push-upgraded in the coming weeks which then lolli-pops the question: Is bfa ready?
In the meantime, the official agent has not been upgraded for quite some time begging the question if there is going to come a 2 level battery check for computing as in 1) stop computing when charge drops below x percent 2) resume computing when charge reaches y percent. Why? Yet again my tab was alternating in stop/start/stop cycle as the charger was not able to keep up. When this happens, eventually the tasks start collapsing and you get a series like this: FAHV_x3EKX-B-AS_0882903_2074_1 3 fahv Signal Error 193/SIGSEGV Code 11 7,19 07-11-14 05:13Setting computing to only when idle is pointless as when messages/updates/notifications arrive, these elevate os activity and the battery charge starts going down till the critical point of the pause/resume cycle and serial crashing starts. Tired of manually suspending or snoozing the agent and checking up if the charge is back up high enough to allow uninterrupted computing. The platform continues to lose out, appears to not make any inroads whatsoever at wcg. A sorry state, considering that anyone who sees charge draining while on wallsocket even when idle will eventually toss boinc out, sick of micromanaging. Please have this fix of 2 battery charge levels computing implemented. To quote one of the project developers, eric j. kopela from a june 3 alpha mail list post, 'Having a high and low water mark (both on charge and temperature) seems like a good idea to me.' We're now 5 months on. No matter what david anderson thinks, continuous stop start stop is detrimental to the task stability, and the well being of android devices. |
||
|
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7659 Status: Offline Project Badges: 
|
Maybe the android developers are waiting for a more robust battery, or a more robust charging system or a hardware solution to a smaller draw of electricity to bail them out of this dilemma.
----------------------------------------Somehow it begs the question of why the engineering folks designed a product which is incapable of running to its full potential even while on a charger. I am thinking it was a push by the marketing people to overly the hype the capabilities of the product in order to push sales. Many of these products are designed to only last until the next newest upgrade hits the market and then be thrown away. Cheers
Sgt. Joe
*Minnesota Crunchers* |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
By the time the wait is up, there's no one left to compute mobile but those you find on the alpha mail list, the diehards.
Think a major part of the weakness is the micro-usb getting flubby. The charger is more than ample. So, either a software solution to stabilize an engineering shortcoming or waiting till kingdom come. Maybe the nativeboinc guy reads this and implements this feature since berkeley is unresponsive and or unwilling. It'd make the mobile boinc experience, a bunch more valid results, on the go a so much better experience as any -unnecessary- fail is a fail too many. The fact that my tab got over 47 repair tasks dumped thursday night for reasons beyond me is an ominous sign on the general stability of boinc for android. Yeah, our esteemed technicians repeated suggestion was to just let the device compute on one core, when there are 4. Who's going with that in the short/medium/long term? |
||
|
|
enels
Senior Cruncher Joined: Apr 25, 2008 Post Count: 286 Status: Offline Project Badges: 
|
How about a setting like: Suspend while charging.
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
That wont work where the reverse setting is already there, do not compute while on battery. When you add 'suspend when on charger' there's no computing at all where -some- will manage to select both. No matter what, even when battery is at 100 percent, charging continues, so we need that 'resume' top setting when on charger, the target period why we have boinc for android. Then we'll have uninterrupted computing for longer when on juice, suspend at the low setting and no discharged surprises in the morning plus a pile of results that went kaboom.
|
||
|
|
noderaser
Senior Cruncher United States Joined: Jun 6, 2006 Post Count: 297 Status: Offline Project Badges: 
|
I think part of the problem, is that BOINC lost their Android developer... At least, I haven't heard that they found a new one as of yet.
---------------------------------------- |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
A qualification to this, joachim was lost to development at start of october, where he was said to have been the gui man. Rom is the under the hood guy.
Lollipop ready: No! it appears from this google group test post series: Hello, My Nexus 7 (2012 Edition) got the Lollipop update. But the BOINC client does not appear to work properly. I can only see the splash-screen when I launch the app and stops there. Regards, Kam-Yung This has been happening ever since the Lollipop Preview back in June. Tap and hold on the BOINC logo and then release. Tap on the GUI Log/Messages tab and you'll see the log. I sent the log in awhile ago with no response. Thanks for the steps. I can see the GUI messages and they seem to show that the client is having problems connecting to the background service. My guess is the service is having problems running under Lollipop. Unfortunately, I believe the Android BOINC project doesn't have a full time Android developer at the moment. Regards, Kam-Yung No problem at all. Let me know how I can help. Last I heard they weren't doing any testing on 5.0 due to a lack of a Nexus device, but this does not make sense to me as they could just use an emulator. On Tuesday, November 18, 2014 11:02:10 PM UTC-5, Kam-Yung Soh wrote: Running nativeboinc since longer, with great satisfaction as being a ´more buttons´ person. Not expecting lollipop to arrive to the 2013-lte for a number of months, same as 4.4.4 arrive much later than the lte-less version. Then the dilemma is to allow the upgrade. The battery management is claimed to be substantially improved. Maybe then the charging can keep up even when in use, get more hours computing done, less or no signal 11/193 |
||
|
|
Coleslaw
Veteran Cruncher USA Joined: Mar 29, 2007 Post Count: 1343 Status: Offline Project Badges: 
|
Has anyone tried nativeBOINC on an Android 5 device yet though?
----------------------------------------    |
||
|
|
David Autumns
Ace Cruncher UK Joined: Nov 16, 2004 Post Count: 11062 Status: Offline Project Badges: 
|
Oh hum
----------------------------------------Got the OTA upgrade for the Nexus 5 It had got up to about 50% valid workunits on Kitkat 4.4.4 Now no more crunching on lollipop With Lollipop it's just sucks  Might have to wait a little longer for the 20 year badge Dave  |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Something is happening as of last night, indicating that both boinc and the science app for fahv have to be recompiled.
On Wednesday, November 19, 2014 7:04:57 PM UTC-5, davea wrote: We've identified the problem: starting with 5.0, Android will run only position-independent executables (PIE), and the current BOINC client (and all existing science apps) are not position-independent. Support for PIE was added in Android 4.1. So if all the native-mode stuff (BOINC client and apps) are recompiled as PIE, everything will work on 4.1 and later devices, but will fail on pre-4.1. Currently around 10% of the active Android devices (2500 out of 27000) are pre-4.1. What we'll do is: - make a PIE version of the BOINC client - encourage projects to build PIE versions of their apps - tell projects how to ensure that their PIE apps get sent only to 4.1+ clients, and (if they want to maintain their non-PIE apps) that these get sent only to pre-5.0 clients. All this will take a week or two, maybe more. -- David |
||
|
|
|