Excessive size of stderr.txt after graphics displayed on Linux w/ ATI Radeon

Message boards : BOINC client : Excessive size of stderr.txt after graphics displayed on Linux w/ ATI Radeon
Message board moderation

To post messages, you must log in.

AuthorMessage
Profile Alexander Klietz
Avatar

Send message
Joined: 15 Nov 06
Posts: 10
Germany
Message 6467 - Posted: 15 Nov 2006, 5:16:31 UTC
Last modified: 15 Nov 2006, 5:17:32 UTC

Hello,

I would like to point to a potential problem when running BOINC with a graphical output on a Linux system with a ATI Radeon R3xx card (such as Radeon 9800 Pro):

The stderr.txt file starts to grow to an excessive size, in my case up to a maximum of 15 GB during approx. 10 minutes, it only stopped at this size because the hard disk (or file system) was full! Apparently the max. disk use value set in preferences does not have any effect in limiting the file size here!

This was observed both with the QMC@home client and the SETI@home Beta client with graphical output. It did not appear (on my system) with the Einstein@home client, however. Others I haven't tested so far.

Stopping the project in the BOINC manager does not stop the increasing of the file size, closing the BOINC manager apparently does stop the increase.

I would recommend, that anyone who has a similar system should be aware of this potential problem.

I include some technical details and screenshots in separate posts to this thread.

Best regards

Alex
I keep trying to attend a meeting of quantum physicists, but every time they agree on a location the appointed time changes and vice versa.


SETI.Germany: 0.3 billion credits
ID: 6467 · Report as offensive
Profile Alexander Klietz
Avatar

Send message
Joined: 15 Nov 06
Posts: 10
Germany
Message 6468 - Posted: 15 Nov 2006, 5:39:20 UTC
Last modified: 15 Nov 2006, 6:16:30 UTC

Here are some technical details about the system:

OS: Ubuntu 6.10 (Linux 2.6.17-10-generic) for Intel x86
BOINC version: boinc_5.4.9_i686-pc-linux-gnu
SETI@home Beta application: setiathome_enhanced 5.17
QMC@home application: Amolqc-beta 5.07

Graphics card: ATI Radeon 9800 Pro (R350) using default driver from Ubuntu 6.10
Other hardware: please see computer summary.

And here is the beginning of the contents of both stderr.txt files:

1. stderr.txt from QMC@home


------
libGL warning: 3D driver claims to not support visual 0x4b
app_graphics_init() called.
scene_init() called.
Scene initialized.
Try R300_SPAN_DISABLE_LOCKING env var if this hangs.
*********************************WARN_ONCE*********************************
File r300_render.c function r300Fallback line 412
Software fallback:ctx->Line.SmoothFlag
***************************************************************************
freeglut ERROR: Function <glutDestroyWindow> called without first calling 'glutInit'.
freeglut ERROR: Function <glutMainLoop> called without first calling 'glutInit'.
freeglut ERROR: Function <glutMainLoop> called without first calling 'glutInit'.
freeglut ERROR: Function <glutMainLoop> called without first calling 'glutInit'.
freeglut ERROR: Function <glutMainLoop> called without first calling 'glutInit'.
freeglut ERROR: Function <glutMainLoop> called without first calling 'glutInit'.
freeglut ERROR: Function <glutMainLoop> called without first calling 'glutInit'.
freeglut ERROR: Function <glutMainLoop> called without first calling 'glutInit'.
freeglut ERROR: Function <glutMainLoop> called without first calling 'glutInit'.
freeglut ERROR: Function <glutMainLoop> called without first calling 'glutInit'.


2. stderr.txt from SETI@home Beta


Work Unit Info:
...............
WU true angle range is : 0.426138
Optimal function choices:
-----------------------------------------------------
name timing error
-----------------------------------------------------
v_BaseLineSmooth 0.43251 0.00000
v_GetPowerSpectrum 0.00283 0.00000
v_ChirpData 0.10632 0.00000
v_Transpose4 0.05310 0.00000
libGL warning: 3D driver claims to not support visual 0x4b
*********************************WARN_ONCE*********************************
File r300_render.c function r300Fallback line 412
Software fallback:ctx->Line.SmoothFlag
***************************************************************************
Try R300_SPAN_DISABLE_LOCKING env var if this hangs.
setiathome-5.17.i686-pc-linux-gnu: freeglut_structure.c:439: fgEnumWindows: Assertion `fgState.Initialised' failed.
SIGABRT: abort called
Stack trace (18 frames):
setiathome-5.17.i686-pc-linux-gnu(boinc_catch_signal+0x91)[0x81400b1]
[0xffffe420]
[0xffffe410]
/lib/tls/i686/cmov/libc.so.6(gsignal+0x50)[0xb7e14770]
/lib/tls/i686/cmov/libc.so.6(abort+0x103)[0xb7e15ef3]
/lib/tls/i686/cmov/libc.so.6(__assert_fail+0xfb)[0xb7e0ddbb]
../../projects/setiweb.ssl.berkeley.edu_beta/setiathome-5.17.i686-pc-linux-gnu.so(fgEnumWindows+0x73)[0xb7dbf853]
../../projects/setiweb.ssl.berkeley.edu_beta/setiathome-5.17.i686-pc-linux-gnu.so(fgWindowByID+0x26)[0xb7dbfa06]
../../projects/setiweb.ssl.berkeley.edu_beta/setiathome-5.17.i686-pc-linux-gnu.so(glutDestroyWindow+0x14)[0xb7dc0394]
../../projects/setiweb.ssl.berkeley.edu_beta/setiathome-5.17.i686-pc-linux-gnu.so(_Z10KillWindowv+0x33)[0xb7db09d7]
../../projects/setiweb.ssl.berkeley.edu_beta/setiathome-5.17.i686-pc-linux-gnu.so[0xb7db0b8d]
../../projects/setiweb.ssl.berkeley.edu_beta/setiathome-5.17.i686-pc-linux-gnu.so(_Z24xwin_graphics_event_loopv+0x122)[0xb7db1256]
../../projects/setiweb.ssl.berkeley.edu_beta/setiathome-5.17.i686-pc-linux-gnu.so(boinc_init_options_graphics_impl+0x47)[0xb7dab9ff]
setiathome-5.17.i686-pc-linux-gnu(_Z31boinc_init_options_graphics_libR13BOINC_OPTIONSPFvvEPc+0x122)[0x8148726]
setiathome-5.17.i686-pc-linux-gnu(boinc_init_graphics_lib+0x53)[0x814885f]
setiathome-5.17.i686-pc-linux-gnu(main+0x245)[0x808f075]
/lib/tls/i686/cmov/libc.so.6(__libc_start_main+0xdc)[0xb7e008cc]
setiathome-5.17.i686-pc-linux-gnu[0x808ecf1]

Exiting...
setiathome-5.17.i686-pc-linux-gnu: freeglut_main.c:1069: glutMainLoop: Assertion `fgState.Initialised' failed.
SIGABRT: abort called
Stack trace (14 frames):
setiathome-5.17.i686-pc-linux-gnu(boinc_catch_signal+0x91)[0x81400b1]
[0xffffe420]
[0xffffe410]
/lib/tls/i686/cmov/libc.so.6(gsignal+0x50)[0xb7e14770]
/lib/tls/i686/cmov/libc.so.6(abort+0x103)[0xb7e15ef3]
/lib/tls/i686/cmov/libc.so.6(__assert_fail+0xfb)[0xb7e0ddbb]
../../projects/setiweb.ssl.berkeley.edu_beta/setiathome-5.17.i686-pc-linux-gnu.so(glutMainLoop+0xa1)[0xb7dbd721]
../../projects/setiweb.ssl.berkeley.edu_beta/setiathome-5.17.i686-pc-linux-gnu.so(_Z24xwin_graphics_event_loopv+0xe2)[0xb7db1216]
../../projects/setiweb.ssl.berkeley.edu_beta/setiathome-5.17.i686-pc-linux-gnu.so(boinc_init_options_graphics_impl+0x47)[0xb7dab9ff]
setiathome-5.17.i686-pc-linux-gnu(_Z31boinc_init_options_graphics_libR13BOINC_OPTIONSPFvvEPc+0x122)[0x8148726]
setiathome-5.17.i686-pc-linux-gnu(boinc_init_graphics_lib+0x53)[0x814885f]
setiathome-5.17.i686-pc-linux-gnu(main+0x245)[0x808f075]
/lib/tls/i686/cmov/libc.so.6(__libc_start_main+0xdc)[0xb7e008cc]
setiathome-5.17.i686-pc-linux-gnu[0x808ecf1]



Screenshot to follow...
I keep trying to attend a meeting of quantum physicists, but every time they agree on a location the appointed time changes and vice versa.


SETI.Germany: 0.3 billion credits
ID: 6468 · Report as offensive
Profile Alexander Klietz
Avatar

Send message
Joined: 15 Nov 06
Posts: 10
Germany
Message 6469 - Posted: 15 Nov 2006, 5:49:23 UTC

The following screenshot shows the disk space taken by two slots running QMC@home after the graphical output has been started:



ID: 6469 · Report as offensive
Profile Alexander Klietz
Avatar

Send message
Joined: 15 Nov 06
Posts: 10
Germany
Message 6470 - Posted: 15 Nov 2006, 5:51:15 UTC

And this shows the contents of one of the slots. Please note the size of the stderr.txt file here:

ID: 6470 · Report as offensive
Profile Alexander Klietz
Avatar

Send message
Joined: 15 Nov 06
Posts: 10
Germany
Message 6471 - Posted: 15 Nov 2006, 5:59:18 UTC
Last modified: 15 Nov 2006, 6:01:30 UTC

I first mentioned this problem to the QMC Forum (thread is here, but is in German language). Martin Korth from the QMC team kindly pointed me to this bug report on freedesktop.org, which might be related to the problem experienced here.

Hope this helps other users to prevent similar problems. The most simple workaround appears to be not to display any graphical output.

Best regards

Alexander Klietz
ID: 6471 · Report as offensive
Eric J Korpela

Send message
Joined: 5 Oct 06
Posts: 5
United States
Message 6481 - Posted: 16 Nov 2006, 1:03:07 UTC - in response to Message 6471.  

Thanks for pointing this out. Apparently we're not handling glut error returns properly. The proper result for this should probably be to terminate the graphics thread and set flags for no graphics.


ID: 6481 · Report as offensive

Message boards : BOINC client : Excessive size of stderr.txt after graphics displayed on Linux w/ ATI Radeon

Copyright © 2024 University of California.
Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.