Installed OpenSuse 11.3 64 Bit and get errors with this host.
Kernel version: 2.6.34.7-0.5
Example:
APP DEBUG: Application caught signal 8.
FPU status word ffff88e1, flags: ERR_SUMM STACK_FAULT PRECISION INVALID
Obtained 5 stack frames for this thread.
Use gdb command: 'info line *0xADDRESS' to print corresponding line numbers.
../../projects/einstein.phys.uwm.edu/einstein_S5GC1_1.05_i686-pc-linux-gnu__S5GCESSE2[0x805f078]
[0xffffe410]
../../projects/einstein.phys.uwm.edu/einstein_S5GC1_1.05_i686-pc-linux-gnu__S5GCESSE2[0x80534ce]
../../projects/einstein.phys.uwm.edu/einstein_S5GC1_1.05_i686-pc-linux-gnu__S5GCESSE2[0x805e2a3]
/lib/libc.so.6(__libc_start_main+0xfe)[0xf7550c0e]
Stack trace of LAL functions in worker thread:
UpdateSemiCohToplist at line 2052 of file /home/bema/einsteinathome/HierarchicalSearch/EinsteinAtHome/source/lalsuite/lalapps/src/pulsar/GCT/HierarchSearchGCT.c
At lowest level status code = 0, description: NO LAL ERROR REGISTERED
06:25:42 (14840): called boinc_finish
OpenSuse 10.3 and 11.1 never caused such errors. I will boot 10.3 until this is hopefully solved.
Michael
Copyright © 2024 Einstein@Home. All rights reserved.
New OS and suddenly facing errors
)
Signal 8 is a floating point exception that is (usually) caused either by an overheated FPU or a kernel bug, where the FPU registers are not (immediately consistently) restored after a context switch.
What kernel versions are these SuSE distros using?
BM
BM
Kernel version:
)
Kernel version: 2.6.34.7-0.5
I think OpenSuse also offers different flavors of the same kernel version (desktop, default, pae, rt ...) and it would probably be useful to know which one you are running. You could also try a different flavor and see if that helps.
CU
HB
It's the desktop version and
)
It's the desktop version and I think the problems started after a recent kernel update, but I'm not sure.
Michael
RE: It's the desktop
)
That uses preemption, I had similar problems with it (here and with C4CW at WCG). Switching to the default version took care of that.
Oh not again......I thought
)
Oh not again......I thought that bug was fixed a long time ago :-(
Anybody made a bug report already against the kernel?
CU
HBE
RE: RE: It's the desktop
)
Thanks a lot! That was the reason. :-)
The default kernel has the following lines in the config file:
CONFIG_PREEMPT_NONE=y
# CONFIG_PREEMPT is not set
Desktop kernel config:
# CONFIG_PREEMPT_NONE is not set
CONFIG_PREEMPT=y
Michael
RE: Oh not again......I
)
I haven't made a bug report, but, a few days ago, I recounted my experience with the same problem in the "Summer Heat Wave" thread. I found the solution to the problem by entering "Signal 8" in the Advanced Search box above.