Message boards :
News :
SETI@home v7 7.06 released for ARMv7 Android
Message board moderation
Author | Message |
---|---|
![]() Send message Joined: 15 Mar 05 Posts: 1547 Credit: 27,183,456 RAC: 0 ![]() |
That includes most everything from DROID 1 to the present. The server shouldn't send this version to android devices without an VFP or NEON FPU. I may build a "no-FPU" version at a later date to see how slow it is. Let me know how it goes. 7.05 had problems on low memory DROIDs, so I've released 7.06 which should have a smaller RAM footprint, at the expense of a larger executable. ![]() |
Send message Joined: 29 May 06 Posts: 1037 Credit: 8,440,339 RAC: 0 ![]() |
I've attached my HTC One S (running the official Boinc 7.2.8) to Seti Beta and grabbed two Wu's. Claggy |
Send message Joined: 29 May 06 Posts: 1037 Credit: 8,440,339 RAC: 0 ![]() |
So far had four computational errors, one got signall 11, while the others had exit code 193, this is the same error I've occasionally had with other projects apps, they mostly just work, Not managed to successfully run the NativeBoinc supplied Seti app on this device eithier, it had the same or similar error. Claggy |
![]() Send message Joined: 15 Mar 05 Posts: 1547 Credit: 27,183,456 RAC: 0 ![]() |
It seems to be having problems on many machines. It doesn't seem to run under BOINC on my Droid RAZR (it hangs out at 0% done and chews CPU), but I haven't rooted that machine, so I can't get to the state file or stderr.txt. It runs fine in stand alone on that machine, though. On by 256MB Droid A855 it dies in stand alone mode because it runs out of memory. I see that I'll need to add an available memory check before deciding to transpose. Would have been nice if android put a swapfile on flash. But it may be running slowly under BOINC. (It's at 0.4%) ![]() |
![]() Send message Joined: 28 Jan 11 Posts: 619 Credit: 2,580,051 RAC: 0 ![]() |
I've added 2 Sony ericsson 1ghz phones. I also increased boinc manager memory aval from 50% to 80%. I haven't got any tasks so far.... And only 1 is visible on "computors". I've resetted both. I'll remove and re attach if needed. |
![]() Send message Joined: 15 Mar 05 Posts: 1547 Credit: 27,183,456 RAC: 0 ![]() |
That's weird. I wonder if they are registering as the same host. They've got the same name (localhost) the same processor and memory and the same OS. I wonder what android reports for the MAC address? ![]() |
![]() Send message Joined: 28 Jan 11 Posts: 619 Credit: 2,580,051 RAC: 0 ![]() |
|
![]() Send message Joined: 16 Jan 07 Posts: 155 Credit: 194,400 RAC: 0 ![]() |
I'm guessing that a HTC One V does not have a FPU as I do not get any work for the phone. |
Send message Joined: 29 May 06 Posts: 1037 Credit: 8,440,339 RAC: 0 ![]() |
I'm guessing that a HTC One V does not have a FPU as I do not get any work for the phone. You might want to check what location that host is on, and what the preferences are for that location (I kept changing my Albert preferences so to only get GPU work, then forgot the mobile was on that location too) Claggy |
![]() Send message Joined: 28 Jan 11 Posts: 619 Credit: 2,580,051 RAC: 0 ![]() |
I'm guessing that a HTC One V does not have a FPU as I do not get any work for the phone. +1 |
![]() Send message Joined: 16 Jan 07 Posts: 155 Credit: 194,400 RAC: 0 ![]() |
I'm guessing that a HTC One V does not have a FPU as I do not get any work for the phone. Yep, was on default that did not have the CPU option checked. Moved to Home and added the CPU check to that one. This will probably take a long time. 1.4% in one hour, so around 80 hours or so. |
![]() Send message Joined: 28 Jan 11 Posts: 619 Credit: 2,580,051 RAC: 0 ![]() |
I also had the cpu option unchecked. It is now marked. I also downloaded 1 task for each phone. So I guess that it is www that only displayes one of them. |
Send message Joined: 29 May 06 Posts: 1037 Credit: 8,440,339 RAC: 0 ![]() |
My HTC One S has finally managed to complete a Wu without erroring it, it did some/all of it in parallel with an Asteroids task, and not a Seti v7 task. http://setiweb.ssl.berkeley.edu/beta/workunit.php?wuid=5550621 Claggy |
![]() Send message Joined: 15 Mar 05 Posts: 1547 Credit: 27,183,456 RAC: 0 ![]() |
We're probably going to have to chose a large minimum memory on android, since many devices don't support swapping. On my 256MB Droid A855 it continually restarts because it gets killed to free memory. 36 hours has gotten me to 2.9%. While on my RAZR, which in theory is only twice as fast, 18 hours has gotten me to 90%. But first I'll add physical memory checking. ![]() |
![]() Send message Joined: 28 Jan 11 Posts: 619 Credit: 2,580,051 RAC: 0 ![]() |
We're probably going to have to chose a large minimum memory on android, since many devices don't support swapping. On my 256MB Droid A855 it continually restarts because it gets killed to free memory. 36 hours has gotten me to 2.9%. While on my RAZR, which in theory is only twice as fast, 18 hours has gotten me to 90%. I have like 340MB of memory and default BM uses only 50% of it. 170Mb left for app. BM should ask when installing how much of %mem it should use. But I guess that is BM(android)-dev team issue. I set mine to 80% usage for BM. 272MB left for app. |
![]() Send message Joined: 15 Mar 05 Posts: 1547 Credit: 27,183,456 RAC: 0 ![]() |
Unfortunately it doesn't really work that way. My RAZR typically has 20MB free. Currently BOINC itself is using 38MB. The SETI app is using 33MB. The way android gets that memory is by killing processes to free up memory. That's the unwritten rule, you can be killed at any time to free up memory. Google apps and location services are taking up most of the memory. Killed background apps usually restart themselves within a few seconds. So when SETI@home needs memory, other apps are killed. When they start back up, SETI@home might get killed. On my 256MB Droid it does. Often it takes the form of the BOINC client itself getting killed, which kills SETI@home. When it comes back there might be enough memory free to start doing some computation. For a while at least. It's restarted more 60 times on my 256MB droid and it's not yet reached 3% ![]() |
![]() Send message Joined: 28 Jan 11 Posts: 619 Credit: 2,580,051 RAC: 0 ![]() |
Unfortunately it doesn't really work that way. My RAZR typically has 20MB free. Currently BOINC itself is using 38MB. The SETI app is using 33MB. The way android gets that memory is by killing processes to free up memory. That's the unwritten rule, you can be killed at any time to free up memory. Google apps and location services are taking up most of the memory. Wow. Einstein, Ateroids and WCG apps works so I guess there's some memory wizard skills that is needed. My first task is as yours only at 0.9% after aprox 8 hours. |
![]() ![]() Send message Joined: 18 Aug 05 Posts: 2423 Credit: 15,878,738 RAC: 0 ![]() |
attached to beta on 3 Android devices (all successfully done V7 tasks under anonymous platform on main already). Using NativeBOINC as BOINC client. All hosts recived tasks (VLARs). |
![]() Send message Joined: 15 Mar 05 Posts: 1547 Credit: 27,183,456 RAC: 0 ![]() |
That 0.9% is a sign that it's restarting all the time. I plan to have a version that can run in less memory out next week. I'll talk to the Einstein guys to see what they are doing. One way around it would be to do memory allocations through a memory mapped file on the internal flash drive, which would allow memory to be swapped out. I'm not doing that because it's a good way to reduce the lifetime of a phone, flash just can't handle that many write cycles for long. ![]() |
![]() Send message Joined: 28 Jan 11 Posts: 619 Credit: 2,580,051 RAC: 0 ![]() |
I readded and upgraded SETI BETA and BM to 7.2.8 on both phones. Old tasks where abandoned and now I run new tasks. Also increased memory for boinc to 100% on 1 phone the other is 80%. Task is at 15% and other is on 6% after running aprox 9hours. Still I can only see one of the phones on SETI Beta www. Eric, I PM you the Macadresses of the phones. |
©2023 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.