SETI@home v8 beta to begin on Tuesday
Message boards :
News :
SETI@home v8 beta to begin on Tuesday
Message board moderation
Previous · 1 . . . 6 · 7 · 8 · 9 · 10 · 11 · 12 . . . 99 · Next
| Author | Message |
|---|---|
 Jeff BuckSend message Joined: 11 Dec 14 Posts: 96 Credit: 1,240,941 RAC: 0 
|
In the Stderr for every v8.01 i686-pc-linux-gnu task, I'm seeing a couple lines that don't appear in v8.01 x86_64-pc-linux-gnu or in any of the v8.00 or prior versions: Low memory machine... Disabling graphics. 0.000000 <= 50331648.000000 With 6GB of memory on that box, it seems a little strange. Any explanation for the sudden appearance of that message? ID: 55378 ·  |
 Send message Joined: 12 Jul 06 Posts: 19 Credit: 705,040 RAC: 0 
|
Just an observation: The bunch of Windows 8x tasks I've completed so far, none have shown any Gaussians. Is this a characteristic of the test data, the program or wot? ID: 55379 · |
|
Send message Joined: 3 Jan 07 Posts: 1451 Credit: 3,272,268 RAC: 0
|
In the Stderr for every v8.01 i686-pc-linux-gnu task, I'm seeing a couple lines that don't appear in v8.01 x86_64-pc-linux-gnu or in any of the v8.00 or prior versions: The 32-bit application is probably struggling to decode > 4 GB of memory, and perhaps going 'round the clock' and interpreting it as negative RAM? ID: 55380 · |
|
Send message Joined: 29 May 06 Posts: 1037 Credit: 8,440,339 RAC: 0
|
In the Stderr for every v8.01 i686-pc-linux-gnu task, I'm seeing a couple lines that don't appear in v8.01 x86_64-pc-linux-gnu or in any of the v8.00 or prior versions: ldd the graphic apps on your host, do you have all the libaries? The i386 graphics app can't run on mine, while the x64 can, and does: (Perhaps it's just misreporting that it can't see how much memory is available) stephen@stephen-MM061:~/BOINC/projects/setiweb.ssl.berkeley.edu_beta$ ldd setigraphics_8.01_x86_64-pc-linux-gnu Claggy ID: 55381 · |
|
Jeff Buck Send message Joined: 11 Dec 14 Posts: 96 Credit: 1,240,941 RAC: 0
|
In the Stderr for every v8.01 i686-pc-linux-gnu task, I'm seeing a couple lines that don't appear in v8.01 x86_64-pc-linux-gnu or in any of the v8.00 or prior versions: Okay, took me a bit to figure that out since I'm not Linux literate. :^) On my machine the ldd returns "not found" only for the libglut.so.3, but does so for both x86_64 and i686. The libGLU.so.1 is successfully found for both. So, that doesn't seem to account for this new message only suddenly appearing for the i686 tasks. In any event, FWIW, I never display the graphics anyhow. Whatever it is, something in the memory-detection seems to have changed in the i686 app between v8.00 and v8.01. I'm wondering if that might factor into the increased run times I've seen, perhaps with the app running with less memory than it could. ID: 55383 · |
|
Send message Joined: 18 Jun 08 Posts: 76 Credit: 113,089 RAC: 0 
|
In the Stderr for every v8.01 i686-pc-linux-gnu task, I'm seeing a couple lines that don't appear in v8.01 x86_64-pc-linux-gnu or in any of the v8.00 or prior versions: I got that too but I have only 2.5 GB of RAM, although if I count swap too then it's about 5 GB of VM. ID: 55384 · |
|
Send message Joined: 18 Jun 08 Posts: 76 Credit: 113,089 RAC: 0
|
For anyone wanting to see the graphics, in Debian the packages are: libglut.so.3 : freeglut3 libGLU.so.1 : libglu1-mesa And you install the 32-bit version like so: sudo apt-get install freeglut3:i386 ID: 55385 · |
Eric J Korpela Send message Joined: 15 Mar 05 Posts: 1547 Credit: 27,183,456 RAC: 0
|
Which version of BOINC are you using? The available memory is provided by BOINC to the app. 
ID: 55387 · |
|
Jeff Buck Send message Joined: 11 Dec 14 Posts: 96 Credit: 1,240,941 RAC: 0
|
Which version of BOINC are you using? The available memory is provided by BOINC to the app. It looks like Juha's host is reporting BOINC 7.7.0, but mine is running 7.2.42. I've also seen it come from Claggy running 7.2.47. I'd expect that you'll likely see it with 7.2.33 when your own v8.01 i686-pc-linux-gnu tasks start reporting. It just started appearing with v8.01 but doesn't appear on any of the x86_64-pc-linux-gnu tasks. ID: 55388 · |
|
Send message Joined: 29 May 06 Posts: 1037 Credit: 8,440,339 RAC: 0
|
Which version of BOINC are you using? The available memory is provided by BOINC to the app. stderrgfx.txt in the slot directory of an i686 wu says:
A successfully opening x86_64 graphics app just produces: Successfully loaded '../../projects/setiweb.ssl.berkeley.edu_beta/seti_801.jpg'. Claggy ID: 55389 · |
Raistmer Send message Joined: 18 Aug 05 Posts: 2423 Credit: 15,878,738 RAC: 0 
|
All my main Windows hosts have 20+ completed tasks for 8.00 already, no invalids. Errors only with 7.99 - bad mem allocations. Time to check GPU part and switch to Anonymous platform. News about SETI opt app releases: https://twitter.com/Raistmer ID: 55391 · |
|
Send message Joined: 3 Jan 07 Posts: 1451 Credit: 3,272,268 RAC: 0
|
I've just noticed that I'm getting the rather ugly, and unnecessary, "Message from task: 0" in the Event Log on completion of each task. 21/12/2015 13:53:02 | | log flags: file_xfer, sched_ops, task 21/12/2015 13:57:35 | SETI@home Beta Test | Message from task: 0 21/12/2015 13:57:35 | SETI@home Beta Test | Computation for task 06ap11ag.12706.16836.4.38.26_0 finished This only happens with Beta v8 tasks - the line doesn't appear with the v7 app. As shown, this is with the default (non-debug) Event Log options in effect. Platform is Windows. Message display is conditional on BOINC client in use: it doesn't appear with v6.12.34. I suspect this is some interaction between the new API and newish clients, designed to display an informative message in exceptional circumstances. But constant repetition of null information doesn't add anything. ID: 55392 · |
|
Send message Joined: 29 May 06 Posts: 1037 Credit: 8,440,339 RAC: 0
|
I've just noticed that I'm getting the rather ugly, and unnecessary, "Message from task: 0" in the Event Log on completion of each task. I'm seeing this on my Self built Boinc 7.6.22 on Linux x64 too: Mon 21 Dec 2015 08:58:20 GMT | | log flags: file_xfer, sched_ops, task, coproc_debug, sched_op_debug Claggy ID: 55393 · |
|
Raistmer Send message Joined: 18 Aug 05 Posts: 2423 Credit: 15,878,738 RAC: 0
|
BOINC 7.6.9 (x64): 21/12/2015 18:41:50 | SETI@home | Message from task: 0 That is, from SETI main, not beta.... Hence, v8 not involved in my case. News about SETI opt app releases: https://twitter.com/Raistmer ID: 55395 · |
|
Eric J Korpela Send message Joined: 15 Mar 05 Posts: 1547 Credit: 27,183,456 RAC: 0
|
I'll try a rebuild of the i686-linux BOINC libraries to see if that helps.
ID: 55396 · |
|
Send message Joined: 2 Jul 13 Posts: 505 Credit: 5,019,318 RAC: 0
|
Seems my machine 76256 is not very fond of the new x86_64 Linux App. It didn't have any trouble with the previous version or the tasks on main. SIGSEGV: segmentation violation Seems to happen when running two CPU VLARS, the one run by itself finished. ID: 55397 · |
|
Eric J Korpela Send message Joined: 15 Mar 05 Posts: 1547 Credit: 27,183,456 RAC: 0
|
Believe it or not, that minimal stack trace helps. I'll let you know what I find.
ID: 55399 · |
|
Send message Joined: 3 Jan 07 Posts: 1451 Credit: 3,272,268 RAC: 0
|
BOINC 7.6.9 (x64): Was that with an 'official build', or with a third party app? If it was third party, I suspect it came in with one of the API tweaks - which would explain why it appears at other projects, like NumberFields@Home, too. ID: 55400 · |
|
Raistmer Send message Joined: 18 Aug 05 Posts: 2423 Credit: 15,878,738 RAC: 0
|
BOINC 7.6.9 (x64): I run SETI main in Anonymous platform mode. Message appears just at task completion. News about SETI opt app releases: https://twitter.com/Raistmer ID: 55401 · |
|
Send message Joined: 3 Jan 07 Posts: 1451 Credit: 3,272,268 RAC: 0
|
Was that with an 'official build', or with a third party app? "Message from task", of course, literally means "message from application which ran the task" - so it matters which application ran the task, rather than how the application was deployed. I take it that 'anonymous platform mode' means 'an application different from the Berkeley official CPU baseline build' - which is the comparison I was making in my original report. And hence, crucially, probably an API newer than 13 Mar 2013 (deployment date of the app I was comparing). Anyway, never mind. I've just written to David to say I believe the display of the message is a Client thing, triggered by line 622 of the function which commences at https://github.com/BOINC/boinc/blob/master/client/app_control.cpp#L606 ID: 55402 · |
©2025 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.