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.

1 · 2 · 3 · 4 . . . 6 · Next

AuthorMessage
Profile Eric J Korpela
Volunteer moderator
Project administrator
Project developer
Project scientist
Avatar

Send message
Joined: 15 Mar 05
Posts: 1547
Credit: 26,981,856
RAC: 717
United States
Message 48104 - Posted: 19 Nov 2013, 23:14:32 UTC


I don't know if a BOINC version capable of running it exists yet.

It's identical to the v7.01 i686-linux version.
ID: 48104 · Report as offensive
Alex Storey
Volunteer tester
Avatar

Send message
Joined: 10 Feb 12
Posts: 107
Credit: 305,151
RAC: 0
Greece
Message 48107 - Posted: 20 Nov 2013, 13:40:09 UTC - in response to Message 48104.  

I don't know if a BOINC version capable of running it exists yet.


Yeah, no!:) Not on Play Store anyway. My proc is still being 'detected' as an ARM by 7.2.15

Also slightly noteworthy (but unrelated to Seti) I tried uninstalling/re-installing BOINC just in case it made any difference. But I used the Play Store uninstall option. That freaked BOINC out, probably because the uninstall had kept the app's data? Locally uninstalling while clearing data first seamed to do the trick.
ID: 48107 · Report as offensive
Ken Sharp
Volunteer tester
Avatar

Send message
Joined: 10 Nov 13
Posts: 7
Credit: 6,080
RAC: 0
United Kingdom
Message 48194 - Posted: 28 Nov 2013, 9:05:00 UTC

Nothing built, but someone could always give building it a go. :-)
ID: 48194 · 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: 26,981,856
RAC: 717
United States
Message 48242 - Posted: 3 Dec 2013, 2:23:22 UTC - in response to Message 48194.  

Rom said he was planning to build it and put it in the Beta section of the play store.

I have no idea what the URL for that is.
ID: 48242 · Report as offensive
Alex Storey
Volunteer tester
Avatar

Send message
Joined: 10 Feb 12
Posts: 107
Credit: 305,151
RAC: 0
Greece
Message 48269 - Posted: 4 Dec 2013, 15:15:20 UTC

Cassandra is back (again)

Eric, I (think) I understand the need for handing out Credit to errored-out tasks on ARM (saves you from a lot of moaning/groaning from the credit hounds?) but the fact that my phone suddenly has a higher RAC than my laptop is just too funny:)

Especially since the phone has an Intel proc and has yet to complete a single workunit successfully!

Is there no easy way to fix this, like increasing the max amount of errors per task or sending replicated WUs to at least one reliable host (or even CPU)?

Just throwing out ideas and info, I don't pretend to know what's best for the project.
ID: 48269 · 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: 26,981,856
RAC: 717
United States
Message 48272 - Posted: 4 Dec 2013, 16:32:11 UTC - in response to Message 48269.  

The hazards of giving out a few months worth of credit in an hour is that the RACs go through the roof. Credit has been so devalued by some other projects that its hard for me to justify putting a whole lot of time into making sure that beta isn't over crediting or getting spikes like this. I can barely find the time to get new versions out.

Rom had put out a beta BOINC client for x86 android, but it was failing, so he unpublished it. I'll let you know when he releases a new version.
ID: 48272 · Report as offensive
Profile Raistmer
Volunteer tester
Avatar

Send message
Joined: 18 Aug 05
Posts: 2423
Credit: 15,878,738
RAC: 0
Russia
Message 48276 - Posted: 4 Dec 2013, 19:54:29 UTC - in response to Message 48272.  

The hazards of giving out a few months worth of credit in an hour is that the RACs go through the roof. Credit has been so devalued by some other projects that its hard for me to justify putting a whole lot of time into making sure that beta isn't over crediting or getting spikes like this. I can barely find the time to get new versions out.

Rom had put out a beta BOINC client for x86 android, but it was failing, so he unpublished it. I'll let you know when he releases a new version.


Save time on credit manipulations. Credits not right with or w/o your manual actions. Better to fix issues with 7.23 that prevent it from running on old android devices.
ID: 48276 · Report as offensive
jason_gee
Volunteer tester

Send message
Joined: 11 Dec 08
Posts: 198
Credit: 658,573
RAC: 0
Australia
Message 48277 - Posted: 4 Dec 2013, 20:27:06 UTC - in response to Message 48276.  
Last modified: 4 Dec 2013, 20:56:29 UTC

The hazards of giving out a few months worth of credit in an hour is that the RACs go through the roof. Credit has been so devalued by some other projects that its hard for me to justify putting a whole lot of time into making sure that beta isn't over crediting or getting spikes like this. I can barely find the time to get new versions out.

Rom had put out a beta BOINC client for x86 android, but it was failing, so he unpublished it. I'll let you know when he releases a new version.


Save time on credit manipulations. Credits not right with or w/o your manual actions. Better to fix issues with 7.23 that prevent it from running on old android devices.


Yeah Eric, somewhere floating in your email is an analysis of 4 identified, so far, CreditNew design issues. It's undercrediting here anyway, against the cobblestone scale. For multibeam underclaim is by a factor of >3.3 times, and >1.65 times for AP. Respective 'Best apps' claim lower raw operation counts than the minimum theoretically possible to do the tasks. In addition to the [deflationary] scaling error, self similar oscillation instabilities are also apparent (making it non-deterministic at best).
ID: 48277 · Report as offensive
William
Volunteer tester
Avatar

Send message
Joined: 14 Feb 13
Posts: 606
Credit: 588,843
RAC: 0
Message 48288 - Posted: 5 Dec 2013, 10:21:41 UTC - in response to Message 48277.  

The hazards of giving out a few months worth of credit in an hour is that the RACs go through the roof. Credit has been so devalued by some other projects that its hard for me to justify putting a whole lot of time into making sure that beta isn't over crediting or getting spikes like this. I can barely find the time to get new versions out.

Rom had put out a beta BOINC client for x86 android, but it was failing, so he unpublished it. I'll let you know when he releases a new version.


Save time on credit manipulations. Credits not right with or w/o your manual actions. Better to fix issues with 7.23 that prevent it from running on old android devices.


Yeah Eric, somewhere floating in your email is an analysis of 4 identified, so far, CreditNew design issues. It's undercrediting here anyway, against the cobblestone scale. For multibeam underclaim is by a factor of >3.3 times, and >1.65 times for AP. Respective 'Best apps' claim lower raw operation counts than the minimum theoretically possible to do the tasks. In addition to the [deflationary] scaling error, self similar oscillation instabilities are also apparent (making it non-deterministic at best).


It might help to know, that I sent that email on the 25th of November. Bernd Machenschalk from Einstein@home replied two days later.
A person who won't read has no advantage over one who can't read. (Mark Twain)
ID: 48288 · 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: 26,981,856
RAC: 717
United States
Message 48305 - Posted: 5 Dec 2013, 21:15:40 UTC - in response to Message 48288.  

I did get the email, and scanned it, but haven't had time to deal with a subject that complicated since then.

ID: 48305 · Report as offensive
Claggy
Volunteer tester

Send message
Joined: 29 May 06
Posts: 1037
Credit: 8,440,339
RAC: 3
United Kingdom
Message 49098 - Posted: 1 Feb 2014, 22:40:14 UTC
Last modified: 1 Feb 2014, 22:43:47 UTC

The x86 Android Boinc has hit the Play store, anyone with one might want to try asking for work again,

Alex Storey's host has a Wu, and hasn't errored it yet:

http://setiweb.ssl.berkeley.edu/beta/results.php?hostid=68839

Claggy
ID: 49098 · Report as offensive
Alex Storey
Volunteer tester
Avatar

Send message
Joined: 10 Feb 12
Posts: 107
Credit: 305,151
RAC: 0
Greece
Message 49108 - Posted: 3 Feb 2014, 14:05:32 UTC

Apologies for not being able to make it to the boards sooner.

Well the good news is the Intel proc is now detected. More good news is that I haven't managed to error out the task by unplugging the charger, making a phone call, switching off the device, pausing the task, and changing the ASUS Energy Saving options. All good signs :)

Here's my initial Event Log (in reverse order remember) from a couple of days ago:

Sat Feb 01 11:33:15 EET 2014||Platform changed from arm-android-linux-gnu to x86-android-linux-gnu
Sat Feb 01 11:33:15 EET 2014||Version change (7.2.15 -> 7.2.39)
Sat Feb 01 11:33:15 EET 2014||Config: report completed tasks immediately
Sat Feb 01 11:33:15 EET 2014||Local time is UTC +2 hours
Sat Feb 01 11:33:15 EET 2014||Disk: 11.72 GB total, 7.78 GB free
Sat Feb 01 11:33:15 EET 2014||Memory: 967.44 MB physical, 0 bytes virtual
Sat Feb 01 11:33:15 EET 2014||OS: Android: 3.0.34
Sat Feb 01 11:33:15 EET 2014||Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe nx constant_tsc arch_perfmon pebs bts nonstop_tsc aperfmperf s3_nonstop_tsc pni dtes64 monitor ds_cpl vmx est tm2 ssse3 xtpr pdcm movbe lahf_lm arat dts tpr_shadow vnmi flexpriority
Sat Feb 01 11:33:15 EET 2014||Using 1 CPUs
Sat Feb 01 11:33:15 EET 2014||Processor: 2 GenuineIntel Intel(R) Atom(TM) CPU Z2420 @ 1.20GHz [Family 6 Model 39 Stepping 2]
Sat Feb 01 11:33:15 EET 2014||Host name: android_433e0eff
Sat Feb 01 11:33:15 EET 2014||No usable GPUs found
Sat Feb 01 11:33:15 EET 2014||Data directory: /data/data/edu.berkeley.boinc/client
Sat Feb 01 11:33:15 EET 2014||Running as a daemon
Sat Feb 01 11:33:15 EET 2014||Libraries: libcurl/7.28.1 OpenSSL/1.0.1c zlib/1.2.6
Sat Feb 01 11:33:15 EET 2014||log flags: file_xfer, sched_ops, task, task_debug
Sat Feb 01 11:33:15 EET 2014||Starting BOINC client version 7.2.39 for x86-android-linux-gnu

The bad news is something is slowing down crunching and I'm never gonna be able to get this WU (the one Claggy linked to... thanx for keeping an eye out Claggy!) done in time. Rough estimate is 50 days to completion running a theoretical 24/7, or realistically probably closer to 150 Days since I only really leave the phone charging at night (for Seti). Any ideas? On a side note: how do I check my battery's temp? You know, so I can rule it out as a suspect...

Probably a more serious problem for now is that my task does not appear to be Checkpointing. Is there any way I can confirm this? Progress appears to reset to zero/nil... likely really does so, I guess.

That's all for now, I think. I'll be around & available for the next few days. And I do take requests:) So if there's anything you gurus want me to try/test/relay/whatever, just let me know and I'll do my best. Just keep in mind you'll need to walk me through any 'advanced' requests:)
ID: 49108 · Report as offensive
Claggy
Volunteer tester

Send message
Joined: 29 May 06
Posts: 1037
Credit: 8,440,339
RAC: 3
United Kingdom
Message 49115 - Posted: 3 Feb 2014, 18:35:15 UTC - in response to Message 49108.  

On a side note: how do I check my battery's temp? You know, so I can rule it out as a suspect...

Get the CPU-Z app, that will tell you Battery temp, as as loads of other interesting info.

Claggy
ID: 49115 · Report as offensive
Alex Storey
Volunteer tester
Avatar

Send message
Joined: 10 Feb 12
Posts: 107
Credit: 305,151
RAC: 0
Greece
Message 49116 - Posted: 3 Feb 2014, 19:12:55 UTC - in response to Message 49115.  

Get the CPU-Z app...


D'oh! Of course... That's what I used last summer when we started this (boy time flies). Thanx Claggy.
Temp is mid to high twenties so no problems there...

Just errored out btw...
Use your link above to have a look at the task... it's all kinds of wrong.

From Event log:
Couldn't "detach & destroy shared memory" whatever that means...
ID: 49116 · Report as offensive
Claggy
Volunteer tester

Send message
Joined: 29 May 06
Posts: 1037
Credit: 8,440,339
RAC: 3
United Kingdom
Message 49117 - Posted: 3 Feb 2014, 19:27:58 UTC - in response to Message 49116.  

Just errored out btw...
Use your link above to have a look at the task... it's all kinds of wrong.

From Event log:
Couldn't "detach & destroy shared memory" whatever that means...

It didn't error out, seems to have completed and validated, app also seems to have been progressing too, But couldn't report it's progress to the client:

http://setiweb.ssl.berkeley.edu/beta/result.php?resultid=15950039
<core_client_version>7.2.39</core_client_version>
<![CDATA[
<stderr_txt>
shmget in attach_shmem: No such file or directory
15:27:53 (16839): Can't set up shared mem: -1. Will run in standalone mode.
setiathome_v7 7.00 Revision: 1782 g++ (GCC) 4.4.1 20090725 (Red Hat 4.4.1-2)
libboinc: BOINC 7.1.0

Work Unit Info:
...............
WU true angle range is : 0.444902
Optimal function choices:
--------------------------------------------------------
name timing error
--------------------------------------------------------
v_BaseLineSmooth (no other)
v_vGetPowerSpectrumUnrolled2 0.000759 0.00000
sse3_ChirpData_ak8 0.050569 0.00000
v_vTranspose4x16ntw 0.009230 0.00000
BH SSE folding 0.005835 0.00000
shmget in attach_shmem: No such file or directory
19:12:58 (23461): Can't set up shared mem: -1. Will run in standalone mode.
Restarted at 10.18 percent.
shmget in attach_shmem: No such file or directory
20:01:44 (24898): Can't set up shared mem: -1. Will run in standalone mode.
Restarted at 12.56 percent.
shmget in attach_shmem: No such file or directory
03:17:30 (15865): Can't set up shared mem: -1. Will run in standalone mode.
Restarted at 23.81 percent.
shmget in attach_shmem: No such file or directory
10:21:54 (25032): Can't set up shared mem: -1. Will run in standalone mode.
Restarted at 46.79 percent.
shmget in attach_shmem: No such file or directory
10:27:50 (25810): Can't set up shared mem: -1. Will run in standalone mode.
Restarted at 46.88 percent.
shmget in attach_shmem: No such file or directory
10:32:00 (26376): Can't set up shared mem: -1. Will run in standalone mode.
Restarted at 46.88 percent.
shmget in attach_shmem: No such file or directory
16:40:35 (3206): Can't set up shared mem: -1. Will run in standalone mode.
Restarted at 46.88 percent.
shmget in attach_shmem: No such file or directory
17:08:59 (4146): Can't set up shared mem: -1. Will run in standalone mode.
Restarted at 48.39 percent.
shmget in attach_shmem: No such file or directory
17:13:13 (4525): Can't set up shared mem: -1. Will run in standalone mode.
Restarted at 48.41 percent.
shmget in attach_shmem: No such file or directory
17:16:34 (4662): Can't set up shared mem: -1. Will run in standalone mode.
Restarted at 48.47 percent.
shmget in attach_shmem: No such file or directory
17:56:02 (6011): Can't set up shared mem: -1. Will run in standalone mode.
Restarted at 50.90 percent.
shmget in attach_shmem: No such file or directory
01:13:40 (17401): Can't set up shared mem: -1. Will run in standalone mode.
Restarted at 67.95 percent.
shmget in attach_shmem: No such file or directory
18:29:18 (4049): Can't set up shared mem: -1. Will run in standalone mode.
Restarted at 99.73 percent.
shmget in attach_shmem: No such file or directory
18:33:45 (4319): Can't set up shared mem: -1. Will run in standalone mode.
Restarted at 99.78 percent.
shmget in attach_shmem: No such file or directory
18:37:48 (4720): Can't set up shared mem: -1. Will run in standalone mode.
Restarted at 99.84 percent.
shmget in attach_shmem: No such file or directory
18:45:32 (5429): Can't set up shared mem: -1. Will run in standalone mode.
Restarted at 99.98 percent.
shmget in attach_shmem: No such file or directory
18:49:03 (5654): Can't set up shared mem: -1. Will run in standalone mode.
Restarted at 99.98 percent.

Flopcounter: 36435481902336.296875

Spike count: 0
Autocorr count: 1
Pulse count: 0
Triplet count: 0
Gaussian count: 0
18:52:51 (5654): called boinc_finish


Claggy
ID: 49117 · Report as offensive
Alex Storey
Volunteer tester
Avatar

Send message
Joined: 10 Feb 12
Posts: 107
Credit: 305,151
RAC: 0
Greece
Message 49118 - Posted: 3 Feb 2014, 20:17:31 UTC
Last modified: 3 Feb 2014, 20:45:16 UTC

Wow! So the task is legit? That's great news. I just assumed it was a false positive.

And just so you don't think I'm crazy, both times I left my phone crunching overnight I woke up to find the task progress at 0.5% which is why I thought it would take about 50 days to finish:)

Edit:

Even a GTX 780Ti is no match for my Atom!!
Average processing rate 788.85 GFLOPS

LOL, sorry, couldn't resist ;)
ID: 49118 · Report as offensive
Profile Coleslaw
Volunteer tester
Avatar

Send message
Joined: 1 May 09
Posts: 16
Credit: 1,948,815
RAC: 210
United States
Message 49126 - Posted: 4 Feb 2014, 3:06:37 UTC
Last modified: 4 Feb 2014, 3:09:21 UTC

Claggy, thanks for pointing this thread out to me. I will start that laptop up again tomorrow when I'm at work. Here is the VM http://setiweb.ssl.berkeley.edu/beta/show_host_detail.php?hostid=70313

I am more interested in if the app works under VM's. As a proof of concept it should get more people interested in porting apps. Google is working on a 64bit version of Android and Intel supposedly has a 64bit kernel out there. Once the 64bit version officially hits, there could be a pretty good gain at some projects due to it being essentially Linux. The Android BOINC app is much easier to update and Android is the only version with the auto update option. Android is typically pretty RAM light already. I find all of these very appealing.
ID: 49126 · Report as offensive
Alex Storey
Volunteer tester
Avatar

Send message
Joined: 10 Feb 12
Posts: 107
Credit: 305,151
RAC: 0
Greece
Message 49130 - Posted: 4 Feb 2014, 10:26:42 UTC

No news really, just a small update:

After 15hrs non-stop crunching my 2nd WU, progress is showing as 1.1% (but probably closer to 50% in reality).

I'll try and keep my phone plugged in until this WU finishes so you guys can get a clearer picture. Maybe in 6-12hrs from now? I have no way of knowing...

http://setiweb.ssl.berkeley.edu/beta/workunit.php?wuid=6049432

Also, the main reason my first task had so many restarts was probably because of the CPU settings. I had it @80% but changed that yesterday to 100% (so the 2nd task probably won't have any restarts).
ID: 49130 · Report as offensive
Claggy
Volunteer tester

Send message
Joined: 29 May 06
Posts: 1037
Credit: 8,440,339
RAC: 3
United Kingdom
Message 49138 - Posted: 4 Feb 2014, 13:52:06 UTC - in response to Message 49126.  

Claggy, thanks for pointing this thread out to me. I will start that laptop up again tomorrow when I'm at work. Here is the VM http://setiweb.ssl.berkeley.edu/beta/show_host_detail.php?hostid=70313

As I said in thread on the Main project, you need to check out your project preferences, your hosts seem to be asking for work but not getting it,
Looks as if you don't have 'Seti v7' selected, probably just got 'SETI@home v7 large workunits' selected of which there are no Wu's available.

Claggy
ID: 49138 · Report as offensive
Alex Storey
Volunteer tester
Avatar

Send message
Joined: 10 Feb 12
Posts: 107
Credit: 305,151
RAC: 0
Greece
Message 49141 - Posted: 4 Feb 2014, 14:47:26 UTC

Dunno if this is useful info but here's what my EV says whenever I make a call on my phone:

Tue Feb 04 13:26:57 EET 2014|SETI@home Beta Test|[task] task_state=EXECUTING for 16au08aa.12942.1299.438086664204.16.95_2 from start
Tue Feb 04 13:26:57 EET 2014|SETI@home Beta Test|[task] ACTIVE_TASK::start(): forked process: pid 3799
Tue Feb 04 13:26:57 EET 2014|SETI@home Beta Test|[error] Couldn't destroy shared memory: shmctl() failed
Tue Feb 04 13:26:57 EET 2014|SETI@home Beta Test|[error] Couldn't detach shared memory: shmget() failed
Tue Feb 04 13:26:57 EET 2014|SETI@home Beta Test|[task] task_state=UNINITIALIZED for 16au08aa.12942.1299.438086664204.16.95_2 from handle_exited_app
Tue Feb 04 13:26:57 EET 2014|SETI@home Beta Test|[task] process got signal 9
Tue Feb 04 13:26:57 EET 2014|SETI@home Beta Test|[task] Process for 16au08aa.12942.1299.438086664204.16.95_2 exited, status 9, task state 1
Tue Feb 04 13:26:56 EET 2014|SETI@home Beta Test|Restarting 16au08aa.12942.1299.438086664204.16.95_2 - message timeout
Tue Feb 04 13:23:56 EET 2014||Resuming network activity
Tue Feb 04 13:23:56 EET 2014|SETI@home Beta Test|[task] task_state=EXECUTING for 16au08aa.12942.1299.438086664204.16.95_2 from unsuspend
Tue Feb 04 13:23:56 EET 2014||Resuming computation
Tue Feb 04 13:23:03 EET 2014||Suspending network activity - computer is in use
Tue Feb 04 13:23:03 EET 2014|SETI@home Beta Test|[task] task_state=SUSPENDED for 16au08aa.12942.1299.438086664204.16.95_2 from suspend
Tue Feb 04 13:23:03 EET 2014||Suspending computation - computer is in use
Mon Feb 03 20:56:30 EET 2014||Resuming network activity

So I guess my 2nd WU will have a few restarts too...
ID: 49141 · Report as offensive
1 · 2 · 3 · 4 . . . 6 · Next

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


 
©2019 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.