einstein_S5R1_4.17_i686-pc-linux-gnu buggy on amd64
25 Aug 2006 16:22:46 UTC
Topic 191739
(moderation:
I just checked the resultlist of my computer because the daily average was so low and found that nearly half of the results either results in a compute error or are marked as invalid even when succesful done.
There are 2 types of errors:
1. http://einsteinathome.org/task/40447415
error:
xTemp is not finite
DEBUG: loop=2583, xTemp=nan, f=1100.277395, alpha=6, tempInt1[alpha]=25
DEBUG: skyConst[ tempInt1[ alpha ] ] = 1800.082042, xSum[ alpha ]=-0.007467
2.
http://einsteinathome.org/task/40260365
error:
ERROR! sftIndex = -2147483648 < 0 in TestLALDemod run 0
alpha=4, xTemp=1980034.29570906353183091, Dterms=16, ifmin=1979878
Level 0: $Id: ComputeFStatistic.c,v 1.371 2006/06/09 12:48:58 reinhard Exp $
Function call `TestLALDemod(status, &Fstat, SFTData, DemodParams)' failed.
file /home/bema/einsteinathome/CFS/EaH_build_release_einstein_S5R1_4.17/extra_sources/lalapps-CVS/src/pulsar/FDS_isolated/ComputeFStatistic.c, line 958
2006-08-23 21:07:39.0244 [normal]:
Level 1: $Id: CFSLALDemod_SSEgas.c,v 1.6 2006/07/28 17:01:40 bema Exp $
2006-08-23 21:07:39.0245 [normal]: Status code 3: Invalid input
2006-08-23 21:07:39.0245 [normal]: function TestLALDemod, file /home/bema/einsteinathome/CFS/EaH_build_release_einstein_S5R1_4.17/extra_sources/lalapps-CVS/src/pulsar/FDS_isolated/CFSLALDemod_SSEgas.c, line 190
2006-08-23 21:07:39.0245 [CRITICAL]: BOINC_ERR_EXIT(): now calling boinc_finish()
3. successfull but marked invalid
http://einsteinathome.org/task/39821292
I never had such problems with the 4.01 version
einstein_S5R1_4.17_i686-pc-linux-gnu buggy on amd64
)
Success Valid.
Success = Result downloaded, crunched, returned, reported. Nothing more.
Validation is a different step.
As for why the errors, they all happened in a row, and over a 2 day period. What there something significant during that time? The errors look at little out of my ballpark. It looks like you are getting back on track, though.
The Client Errors are from
)
The Client Errors are from sanity checks in the program, meaning that something is happening that should not be possible (the calculations we do could not yield infinite numbers, and an index into an array could not be negative). Also a result that is sent back from a run where the App didn't crash (i.e. that was "successful") may contain wrong numbers (or values that are not numbers at all).
Errors of this type most likely point to an overheated CPU, possibly during the hotter days of the year. Other hardware problems are also possible (heat related or not), but they are by far less common.
BM
BM