I just installed Boinc on my Verizon Galaxy S3 last night and it has completed one and a half work units so far (it hasn't uploaded the completed one yet so I'm not sure if it was successful.) I'm just running it on one CPU for now because I was wary of overheating. I've hit a weird problem I haven't seen in this thread though, when I look at the event log it seems to be suspending and then resuming computation extremely often, due to the GUI not being active. This is a small segment of the event log:
Mon Jul 29 09:57:18 EDT 2013||Resuming computation
Mon Jul 29 09:56:40 EDT 2013||Suspending network activity - GUI not active
Mon Jul 29 09:56:40 EDT 2013|Einstein@Home|[task] task_state=SUSPENDED for p2030.20130404.G40.87-01.07.N.b1s0g0.00000_3182_2 from suspend
Mon Jul 29 09:56:40 EDT 2013||Suspending computation - GUI not active
Mon Jul 29 09:55:43 EDT 2013||Resuming network activity
Mon Jul 29 09:55:43 EDT 2013|Einstein@Home|[task] task_state=EXECUTING for p2030.20130404.G40.87-01.07.N.b1s0g0.00000_3182_2 from unsuspend
Mon Jul 29 09:55:43 EDT 2013||Resuming computation
Mon Jul 29 09:54:29 EDT 2013||Suspending network activity - GUI not active
Mon Jul 29 09:54:29 EDT 2013|Einstein@Home|[task] task_state=SUSPENDED for p2030.20130404.G40.87-01.07.N.b1s0g0.00000_3182_2 from suspend
Mon Jul 29 09:54:29 EDT 2013||Suspending computation - GUI not active
Mon Jul 29 09:54:06 EDT 2013||Resuming network activity
Mon Jul 29 09:54:06 EDT 2013|Einstein@Home|[task] task_state=EXECUTING for p2030.20130404.G40.87-01.07.N.b1s0g0.00000_3182_2 from unsuspend
Mon Jul 29 09:54:06 EDT 2013||Resuming computation
Mon Jul 29 09:53:38 EDT 2013||Suspending network activity - GUI not active
Mon Jul 29 09:53:38 EDT 2013|Einstein@Home|[task] task_state=SUSPENDED for p2030.20130404.G40.87-01.07.N.b1s0g0.00000_3182_2 from suspend
Mon Jul 29 09:53:38 EDT 2013||Suspending computation - GUI not active
Mon Jul 29 09:52:48 EDT 2013||Resuming network activity
Mon Jul 29 09:52:48 EDT 2013|Einstein@Home|[task] task_state=EXECUTING for p2030.20130404.G40.87-01.07.N.b1s0g0.00000_3182_2 from unsuspend
Mon Jul 29 09:52:48 EDT 2013||Resuming computation
Mon Jul 29 09:52:13 EDT 2013||Suspending network activity - GUI not active
Mon Jul 29 09:52:13 EDT 2013|Einstein@Home|[task] task_state=SUSPENDED for p2030.20130404.G40.87-01.07.N.b1s0g0.00000_3182_2 from suspend
Mon Jul 29 09:52:13 EDT 2013||Suspending computation - GUI not active
Mon Jul 29 09:51:42 EDT 2013||Resuming network activity
Mon Jul 29 09:51:42 EDT 2013|Einstein@Home|[task] task_state=EXECUTING for p2030.20130404.G40.87-01.07.N.b1s0g0.00000_3182_2 from unsuspend
Mon Jul 29 09:51:42 EDT 2013||Resuming computation
Mon Jul 29 09:51:38 EDT 2013||Suspending network activity - GUI not active
Mon Jul 29 09:51:38 EDT 2013|Einstein@Home|[task] task_state=SUSPENDED for p2030.20130404.G40.87-01.07.N.b1s0g0.00000_3182_2 from suspend
Mon Jul 29 09:51:38 EDT 2013||Suspending computation - GUI not active
Mon Jul 29 09:51:07 EDT 2013||Resuming network activity
Mon Jul 29 09:51:07 EDT 2013|Einstein@Home|[task] task_state=EXECUTING for p2030.20130404.G40.87-01.07.N.b1s0g0.00000_3182_2 from unsuspend
Mon Jul 29 09:51:07 EDT 2013||Resuming computation
Mon Jul 29 09:50:38 EDT 2013||Suspending network activity - GUI not active
Mon Jul 29 09:50:38 EDT 2013|Einstein@Home|[task] task_state=SUSPENDED for p2030.20130404.G40.87-01.07.N.b1s0g0.00000_3182_2 from suspend
Mon Jul 29 09:50:38 EDT 2013||Suspending computation - GUI not active
As I said, it seems to be computing OK, but this can't be helping efficiency very much. Anyone have any idea if there's something causing it that I could fix, or is it just the Android OS being stupid?
Now I've run a mix of Pogs and Milkyway, no problem so far.
It really looks like Einstein crashes the system when running a mix with other projects and using all cores.
I have no idea why however I can run pogs and E@H just fine both at the same time. As you said it might be the client
Hi Kiska,
yes, shure, I run the same mix on 2 more devices without error, but I do not use all cores there.
Might well be possible that we have reached a limit within distributed computing / distributed developement; who is responsible for that type of error? Maybe we'll never know.
But as long as there is a workaround I can live with it. And maybe, when HBE or BM, whoever has written this app, reads these cases of malfunction he gets a 'flash of inspiration' and knows where to fix the lever ...
Now I've run a mix of Pogs and Milkyway, no problem so far.
It really looks like Einstein crashes the system when running a mix with other projects and using all cores.
For the moment I will run a mix for some days and then Einstein only, but I hope that he problem will be fixed some time.
Alex
Looking closer into my account (the device has a different name when using official BOINC) I found another one: http://einsteinathome.org/task/394532706
also segemtation violation. This means it is not limited to native Boinc.
My previously reported problem is gone now. I'm not sure how exactly. I needed to reboot my device. I connected the Nexus 4 to its charger and started BOINC right after that. It said that it needed to have 70% battery (even though it was attached to the charger) and after it reached it, it started crunching.
Looks well now!
OK, so it's not only POGS tasks getting these SIGSEGV errors. There's one user constantly getting these with the official BOINC client running NativeBOINC BugCatcher version for me so I can see what the stack traces contain if it happens on NativeBOINC as well. It seems like such a pain to pull these out at the moment. Glad NativeBOINC has this feature. Hopefully BOINC guys can add something similar as well.
Well I have committed my only Raspberry Pi to the Einstein effort. Just a few weeks ago E&H did not support the Pi. Now it does. I am waiting on a new over clock kit to arrive (case with fan and 3 heat sinks). I am currently overclocked to a medium setting and the initial temp was about 53C. I added an external USB fan without removing the cases lid and the temp dropped to about 42C. With the arrival of the new kit I will push to the max setting. The Pi will never be a major contributor but every little bit helps. And if the Pi should find something interesting all of us with bigger and better boxes might be just a little bit jealous. :>)
Well I have committed my only Raspberry Pi to the Einstein effort.
Excellent! I have two Pis doing E@H , one is otherwise used as a Media Center and the other one is used for experiments in home automation and as a webcam platform. I love them :-).
I just installed Boinc on my
)
I just installed Boinc on my Verizon Galaxy S3 last night and it has completed one and a half work units so far (it hasn't uploaded the completed one yet so I'm not sure if it was successful.) I'm just running it on one CPU for now because I was wary of overheating. I've hit a weird problem I haven't seen in this thread though, when I look at the event log it seems to be suspending and then resuming computation extremely often, due to the GUI not being active. This is a small segment of the event log:
As I said, it seems to be computing OK, but this can't be helping efficiency very much. Anyone have any idea if there's something causing it that I could fix, or is it just the Android OS being stupid?
Now I've run a mix of Pogs
)
Now I've run a mix of Pogs and Milkyway, no problem so far.
It really looks like Einstein crashes the system when running a mix with other projects and using all cores.
http://einsteinathome.org/task/394624348 SIGSEGV: segmentation violation
http://einsteinathome.org/task/394583346 SIGSEGV: segmentation violation
http://einsteinathome.org/task/394454362 Exit Status 11
http://einsteinathome.org/task/394382314 SIGSEGV: segmentation violation
For the moment I will run a mix for some days and then Einstein only, but I hope that he problem will be fixed some time.
Alex
I have no idea why however I
)
I have no idea why however I can run pogs and E@H just fine both at the same time. As you said it might be the client
RE: I have no idea why
)
There are other users reporting segementation errors as well:
http://einsteinathome.org/node/197076&nowrap=true#126034
http://einsteinathome.org/node/197076&nowrap=true#126050
RE: I have no idea why
)
Hi Kiska,
yes, shure, I run the same mix on 2 more devices without error, but I do not use all cores there.
Might well be possible that we have reached a limit within distributed computing / distributed developement; who is responsible for that type of error? Maybe we'll never know.
But as long as there is a workaround I can live with it. And maybe, when HBE or BM, whoever has written this app, reads these cases of malfunction he gets a 'flash of inspiration' and knows where to fix the lever ...
Cheers
Alexander
RE: Now I've run a mix of
)
Looking closer into my account (the device has a different name when using official BOINC) I found another one:
http://einsteinathome.org/task/394532706
also segemtation violation. This means it is not limited to native Boinc.
My previously reported
)
My previously reported problem is gone now. I'm not sure how exactly. I needed to reboot my device. I connected the Nexus 4 to its charger and started BOINC right after that. It said that it needed to have 70% battery (even though it was attached to the charger) and after it reached it, it started crunching.
Looks well now!
Daniel here - volunteer
)
Daniel here - volunteer dev/tester over at POGS.
OK, so it's not only POGS tasks getting these SIGSEGV errors. There's one user constantly getting these with the official BOINC client running NativeBOINC BugCatcher version for me so I can see what the stack traces contain if it happens on NativeBOINC as well. It seems like such a pain to pull these out at the moment. Glad NativeBOINC has this feature. Hopefully BOINC guys can add something similar as well.
Cheers
Daniel
Well I have committed my only
)
Well I have committed my only Raspberry Pi to the Einstein effort. Just a few weeks ago E&H did not support the Pi. Now it does. I am waiting on a new over clock kit to arrive (case with fan and 3 heat sinks). I am currently overclocked to a medium setting and the initial temp was about 53C. I added an external USB fan without removing the cases lid and the temp dropped to about 42C. With the arrival of the new kit I will push to the max setting. The Pi will never be a major contributor but every little bit helps. And if the Pi should find something interesting all of us with bigger and better boxes might be just a little bit jealous. :>)
RE: Well I have committed
)
Excellent! I have two Pis doing E@H , one is otherwise used as a Media Center and the other one is used for experiments in home automation and as a webcam platform. I love them :-).
Cheers
HB