SETI@home v7.01 released for x86-android

Message boards : News : SETI@home v7.01 released for x86-android
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3 · 4 · 5 · 6 · Next

AuthorMessage
Profile Coleslaw
Volunteer tester
Avatar

Send message
Joined: 1 May 09
Posts: 16
Credit: 2,031,013
RAC: 0
United States
Message 49238 - Posted: 8 Feb 2014, 6:29:00 UTC

I now have a VM running Android 4.3 x86 on my i7 920

http://setiweb.ssl.berkeley.edu/beta/show_host_detail.php?hostid=70432
ID: 49238 · Report as offensive
Profile Coleslaw
Volunteer tester
Avatar

Send message
Joined: 1 May 09
Posts: 16
Credit: 2,031,013
RAC: 0
United States
Message 49245 - Posted: 8 Feb 2014, 20:58:09 UTC
Last modified: 8 Feb 2014, 21:02:05 UTC

First i7 920 http://setiweb.ssl.berkeley.edu/beta/show_host_detail.php?hostid=70432 (stock speeds) work unit finished successfully.

http://setiweb.ssl.berkeley.edu/beta/result.php?resultid=16013380

This system is running multiple VM's and has BOINC running on the core system, so the actual speeds might be better if ran stand alone. I may add a VM of Android 4.2 to see if that makes a difference in speed too.
ID: 49245 · Report as offensive
Alex Storey
Volunteer tester
Avatar

Send message
Joined: 10 Feb 12
Posts: 107
Credit: 305,151
RAC: 0
Greece
Message 49470 - Posted: 20 Feb 2014, 22:02:30 UTC

My first error:
http://setiweb.ssl.berkeley.edu/beta/result.php?resultid=16082318

BTW should I set to NNT until checkpointing gets fixed?
ID: 49470 · Report as offensive
Claggy
Volunteer tester

Send message
Joined: 29 May 06
Posts: 1037
Credit: 8,440,339
RAC: 0
United Kingdom
Message 49473 - Posted: 20 Feb 2014, 22:56:48 UTC - in response to Message 49470.  
Last modified: 20 Feb 2014, 22:59:46 UTC

I think Eric needs to rebuild the app with an Android compatible api, as a first step, and then see how it goes.

Claggy
ID: 49473 · Report as offensive
Josef W. Segur
Volunteer tester

Send message
Joined: 14 Oct 05
Posts: 1137
Credit: 1,848,733
RAC: 0
United States
Message 49479 - Posted: 21 Feb 2014, 3:15:51 UTC - in response to Message 49470.  

My first error:
http://setiweb.ssl.berkeley.edu/beta/result.php?resultid=16082318

BTW should I set to NNT until checkpointing gets fixed?

Checkpointing is working fine, so the application restarts at xx.x percent rather than at the beginning. The communication channel between BOINC and the app which doesn't get established is the problem. That just means BOINC doesn't know the task has checkpointed nor how much time the work took, which won't matter until "Number of tasks completed" grows from its present 5 to 11.

The "finish file present too long" which caused the EXIT_ABORTED_BY_CLIENT means the task had been fully processed, but maybe the OS stopped the app just before it actually exited. I consider it a poor design choice by Dr. Anderson rather than an issue with the application.
                                                                   Joe
ID: 49479 · Report as offensive
Alex Storey
Volunteer tester
Avatar

Send message
Joined: 10 Feb 12
Posts: 107
Credit: 305,151
RAC: 0
Greece
Message 49492 - Posted: 21 Feb 2014, 12:33:24 UTC - in response to Message 49479.  

Checkpointing is working fine...


Yes. Sorry. Thank you. I think that's the second time I've used that word haphazardly. Otherwise I'd still be on my first task!:) What I meant was "elapsed time" and how it could eventually mess up APR.

I do not mind crunching a task that Boinc cannot recognize and I don't mind messing up my APR. It's the hazard of doing business.

But I'm a Seti buffoon and I need a Seti guru to tell me if I should set to NNT or not. What's best for the project?
ID: 49492 · Report as offensive
Profile David S
Volunteer tester
Avatar

Send message
Joined: 10 Sep 13
Posts: 1187
Credit: 2,791,507
RAC: 0
United States
Message 49499 - Posted: 21 Feb 2014, 14:43:01 UTC - in response to Message 49492.  

But I'm a Seti buffoon and I need a Seti guru to tell me if I should set to NNT or not. What's best for the project?

I think most of us are just letting it run. If you get a lot of errors, the server will impose a limit on you.
David
signature sent back to alpha testing
ID: 49499 · Report as offensive
Josef W. Segur
Volunteer tester

Send message
Joined: 14 Oct 05
Posts: 1137
Credit: 1,848,733
RAC: 0
United States
Message 49528 - Posted: 23 Feb 2014, 1:01:31 UTC - in response to Message 49499.  

But I'm a Seti buffoon and I need a Seti guru to tell me if I should set to NNT or not. What's best for the project?

I think most of us are just letting it run. If you get a lot of errors, the server will impose a limit on you.

And the limit could eventually get down to one task a day, effectively no limit at all for a slow enough processor.

Alex, my take is that when you get to 11 completed you'll still be OK so long as the application gets stopped and restarted several times during each task. My judgement is getting a variety of tasks done with the application on your system is valuable enough that setting NNT isn't desirable.

A different way to avoid any difficulty with the bad estimates would be to force the servers to give the system a new hostID if that's possible, the only way I know is to reduce the rpc_seqno for the project in client_state.xml while BOINC is shut down. If that can be done the servers would begin building a new speed estimate for the "new" host. You should run down the cached work on the "old" host prior to trying that.
                                                                  Joe
ID: 49528 · Report as offensive
Alex Storey
Volunteer tester
Avatar

Send message
Joined: 10 Feb 12
Posts: 107
Credit: 305,151
RAC: 0
Greece
Message 49550 - Posted: 24 Feb 2014, 1:33:35 UTC

Thank you Joe (and N9JFE).

I'll get back on it now...
ID: 49550 · Report as offensive
Alex Storey
Volunteer tester
Avatar

Send message
Joined: 10 Feb 12
Posts: 107
Credit: 305,151
RAC: 0
Greece
Message 49560 - Posted: 24 Feb 2014, 14:41:52 UTC

Off topic - Boinc related

Liked the new "no crunching when screen on" option (even though I'm probably not going to be using it, a lot of people will) and especially liked the fact that the message was right there, so I knew why I was paused for no reason and to disable this new option. Clever, bravo.

Not fond of the new menu (enforced in previous version) but I guess that is a Google design. Extra click/tap and just feels overcomplicated, fussy and messy. I liked the tabs... oh well.

Also like the "suspend" option on the Android Home-screen drop-down menu. But maybe that was already there and I only just noticed.

Anywayz, good luck today going mainstream! :)
ID: 49560 · Report as offensive
Alex Storey
Volunteer tester
Avatar

Send message
Joined: 10 Feb 12
Posts: 107
Credit: 305,151
RAC: 0
Greece
Message 50028 - Posted: 18 Mar 2014, 20:56:26 UTC
Last modified: 18 Mar 2014, 21:09:25 UTC

Joe, I was actually worried about getting too many tasks (instead of too few) but then I thought "oh well.. how bad can it be?"

Well I passed the 11-task milestone and now I've got about 3 weeks worth of work even though I'm only asking for half a day:)

OK, I'm just joking around... I know why this is happening and knew it would happen too. So not complaining or anything...

But is there any kind of voodoo I can do to get Boinc to ask for one task at a time?

Edit: 100th post :)
ID: 50028 · Report as offensive
Josef W. Segur
Volunteer tester

Send message
Joined: 14 Oct 05
Posts: 1137
Credit: 1,848,733
RAC: 0
United States
Message 50029 - Posted: 19 Mar 2014, 5:06:14 UTC - in response to Message 50028.  

...
But is there any kind of voodoo I can do to get Boinc to ask for one task at a time?
...

If you set resource share for this project to zero, BOINC will only ask for 1 second of work, and only ask when you have an idle CPU. That's usually used for setting up a backup project to keep the host occupied when the main project isn't delivering work, but looks like it should also work even though this is the only project the host is doing.

A slightly less extreme setting would be to simply set the cache to less than the estimated time of one WU, perhaps 0.02 days for the "maintain enough" setting and 0 for "additional". But that 28 minute 48 second cache setting probably wouldn't be much different than the zero resource share since BOINC isn't getting fraction done nor time from the application.

The other main concern now is avoiding the "Maximum elapsed time exceeded" error. The splitters are setting that bound to 20 times the estimate now, but with your estimate being so much shorter than reality that's not nearly enough. If the elapsed time shown for a task ever reaches that limit, BOINC will kill the task. But IIRC when BOINC restarts your tasks the elapsed time is considered zero since it thinks the task is starting at the beginning, so as long as there is no uninterrupted section that reaches the limit the error should be avoided. I could be wrong about that, I haven't traced the BOINC client logic fully.
                                                                  Joe
ID: 50029 · Report as offensive
Claggy
Volunteer tester

Send message
Joined: 29 May 06
Posts: 1037
Credit: 8,440,339
RAC: 0
United Kingdom
Message 50038 - Posted: 19 Mar 2014, 13:44:36 UTC - in response to Message 50029.  

...
But is there any kind of voodoo I can do to get Boinc to ask for one task at a time?
...

A slightly less extreme setting would be to simply set the cache to less than the estimated time of one WU, perhaps 0.02 days for the "maintain enough" setting and 0 for "additional". But that 28 minute 48 second cache setting probably wouldn't be much different than the zero resource share since BOINC isn't getting fraction done nor time from the application.

The Official Android Boinc doesn't use the Web Computing preferences, so you can't set the cache level any lower.

Claggy
ID: 50038 · Report as offensive
William
Volunteer tester
Avatar

Send message
Joined: 14 Feb 13
Posts: 606
Credit: 588,843
RAC: 0
Message 50045 - Posted: 19 Mar 2014, 17:39:19 UTC - in response to Message 50038.  

...
But is there any kind of voodoo I can do to get Boinc to ask for one task at a time?
...

A slightly less extreme setting would be to simply set the cache to less than the estimated time of one WU, perhaps 0.02 days for the "maintain enough" setting and 0 for "additional". But that 28 minute 48 second cache setting probably wouldn't be much different than the zero resource share since BOINC isn't getting fraction done nor time from the application.

The Official Android Boinc doesn't use the Web Computing preferences, so you can't set the cache level any lower.

Claggy

ROFLMAO

sorry it was either that or hang my head and cry.
A person who won't read has no advantage over one who can't read. (Mark Twain)
ID: 50045 · Report as offensive
Profile Raistmer
Volunteer tester
Avatar

Send message
Joined: 18 Aug 05
Posts: 2423
Credit: 15,878,738
RAC: 0
Russia
Message 50055 - Posted: 19 Mar 2014, 21:29:33 UTC

Solution quite simple - use NativeBOINC.
I used both and NativeBOINC much more convenient.
BTW, anyone knows how to polite exit from stock BOINC for Android? I didn't find a way. Only to hard stop process and service via system properties/app management...
ID: 50055 · Report as offensive
Profile Eric J Korpela
Volunteer moderator
Project administrator
Project developer
Project scientist
Avatar

Send message
Joined: 15 Mar 05
Posts: 1547
Credit: 27,183,456
RAC: 0
United States
Message 50057 - Posted: 19 Mar 2014, 23:43:56 UTC - in response to Message 50055.  

Exiting is against the android design philosophy. ;)
ID: 50057 · Report as offensive
Richard Haselgrove
Volunteer tester

Send message
Joined: 3 Jan 07
Posts: 1451
Credit: 3,272,268
RAC: 0
United Kingdom
Message 50058 - Posted: 19 Mar 2014, 23:47:55 UTC - in response to Message 50057.  

Exiting is against the android design philosophy. ;)

Nice one.
ID: 50058 · Report as offensive
Claggy
Volunteer tester

Send message
Joined: 29 May 06
Posts: 1037
Credit: 8,440,339
RAC: 0
United Kingdom
Message 50061 - Posted: 20 Mar 2014, 4:56:08 UTC - in response to Message 50058.  

Exiting is against the android design philosophy. ;)

Nice one.

Doesn't mean you can't have an app that exits, the Steam app can be exited.

Claggy
ID: 50061 · Report as offensive
Profile David S
Volunteer tester
Avatar

Send message
Joined: 10 Sep 13
Posts: 1187
Credit: 2,791,507
RAC: 0
United States
Message 50073 - Posted: 20 Mar 2014, 15:33:26 UTC - in response to Message 50061.  

Exiting is against the android design philosophy. ;)

Nice one.

Doesn't mean you can't have an app that exits, the Steam app can be exited.

Claggy

On my phone, I touch the left soft key (the one that, from the phone's home screen, opens Settings and such) and it brings up 5 buttons at the bottom of the screen. One of them is "Exit BOINC". Works quite nicely, too.
David
signature sent back to alpha testing
ID: 50073 · Report as offensive
Grumpy Swede
Volunteer tester
Avatar

Send message
Joined: 10 Mar 12
Posts: 1700
Credit: 13,216,373
RAC: 0
Sweden
Message 50076 - Posted: 20 Mar 2014, 15:49:39 UTC - in response to Message 50073.  

Exiting is against the android design philosophy. ;)

Nice one.

Doesn't mean you can't have an app that exits, the Steam app can be exited.

Claggy

On my phone, I touch the left soft key (the one that, from the phone's home screen, opens Settings and such) and it brings up 5 buttons at the bottom of the screen. One of them is "Exit BOINC". Works quite nicely, too.

You run Boinc 7.2.15. In the latest Boinc versions they actully stupidly enough removed the Exit option. I have the last version which still had the Exit option, V 7.3.0.

I caught it in time that the latest ones available from Google Play had the Exit option removed, which means that if it isn't brought back again, I will never upgrade. If there is a requirement to upgrade to run SETI in the future, I will stop running SETI on my android phone.
WARNING!! "THIS IS A SIGNATURE", of the "IT MAY CHANGE AT ANY MOMENT" type. It may, or may not be considered insulting, all depending upon HOW SENSITIVE THE VIEWER IS, to certain inputs to/from the nervous system.
ID: 50076 · Report as offensive
Previous · 1 · 2 · 3 · 4 · 5 · 6 · Next

Message boards : News : SETI@home v7.01 released for x86-android


 
©2020 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.