I got a 0 credit result too, good luck with finding the problem, if it's even yours. ;)
I will do a comparison between clients to find more points to increase the similarity.
I did this comparison, so I see that a two times bigger sin/cos look-up table would be able to reduce the differences between Sxx/Dxx and the original application results. So, I will switch back to the S39 size look-up table. It will decrease the speed with about 5%.
I got a 0 credit result too, good luck with finding the problem, if it's even yours. ;)
I will do a comparison between clients to find more points to increase the similarity.
I did this comparison, so I see that a two times bigger sin/cos look-up table would be able to reduce the differences between Sxx/Dxx and the original application results. So, I will switch back to the S39 size look-up table. It will decrease the speed with about 5%.
I get about 5% invalid results, so it does not make a difference to me, but to the project.
But as far as I remember, I did not have any errors with S40. Just luck?
The interesting point is, that another host got the same error too with this WU.
The three other results got calculated by classic alberts, but I don't think that means anything.
cu,
Michael
If you haven't done so already, you should read and respond to the If you have an Albert crashing on 0xC0000005 ... survey thread. I think the cause of the -1073741819 (0xc0000005) problem is still unknown but you may be the first to report having it with one of Akos' apps.
If you haven't done so already, you should read and respond to the If you have an Albert crashing on 0xC0000005 ... survey thread. I think the cause of the -1073741819 (0xc0000005) problem is still unknown but you may be the first to report having it with one of Akos' apps.
Done. :)
But I don't believe the reason is found in akos's optimizations.
If you haven't done so already, you should read and respond to the If you have an Albert crashing on 0xC0000005 ... survey thread. I think the cause of the -1073741819 (0xc0000005) problem is still unknown but you may be the first to report having it with one of Akos' apps.
Done. :)
But I don't believe the reason is found in akos's optimizations.
cu,
Michael
I don't think the problem is related to Akos' apps either. The first reported occurences happened with regular Albert 4.37. If it is a coding issue, it's in the base code (and Akos has "optimized" the bug along with everything else). Since it is still an "unknown", I just thought I'd help publicize the effort to gather data on it again.
RE: I got a 0 credit result
)
I will do a comparison between clients to find more points to increase the similarity.
Access Violation in
)
Access Violation in here
5.3.12.tx37
- exit code -1073741819 (0xc0000005)
2006-04-19 14:30:10.9843 [normal]: Optimised by akosf S40.04 --> 'projects/einstein.phys.uwm.edu/albert_4.37_windows_intelx86.exe'
2006-04-19 14:30:10.9843 [normal]: Started search at lalDebugLevel = 0
2006-04-19 14:30:12.8906 [normal]: Checkpoint-file 'Fstat.out.ckp' not found.
2006-04-19 14:30:12.8906 [normal]: No usable checkpoint found, starting from beginning.
2006-04-19 14:32:55.1875 [normal]: Fstat file reached MaxFileSizeKB ==> compactifying ... done.
***UNHANDLED EXCEPTION****
Reason: Access Violation (0xc0000005) at address 0x0040AA5F read attempt to address 0x00951090
1: 04/19/06 15:42:38
1: e:\\einsteinathome\\cfs\\windows_build\\albert4.37\\cfslaldemod.c(928) +80 bytes (TestLALDemod)
1: SymGetLineFromAddr(): GetLastError = 126
RE: RE: I got a 0 credit
)
I did this comparison, so I see that a two times bigger sin/cos look-up table would be able to reduce the differences between Sxx/Dxx and the original application results. So, I will switch back to the S39 size look-up table. It will decrease the speed with about 5%.
RE: RE: RE: I got a 0
)
I get about 5% invalid results, so it does not make a difference to me, but to the project.
But as far as I remember, I did not have any errors with S40. Just luck?
http://einsteinathome.org/goto/comment/27337
cu,
Michael
RE: But as far as I
)
Same here... no errors with S40, S40.03, or S40.04. How lucky can I be? Here are some examples... hope they don't purge these too soon...
AMD Athlon XP 1800+ Palomino
Intel P3-800 Coppermine
I got an error -1073741819
)
I got an error -1073741819 (0xc0000005) with S40.04:
***UNHANDLED EXCEPTION****
Reason: Access Violation (0xc0000005) at address 0x0040AABD read attempt to address 0xFF7455C8
1: 04/20/06 14:32:20
1: e:\\einsteinathome\\cfs\\windows_build\\albert4.37\\cfslaldemod.c(941) +7 bytes (TestLALDemod)
result details
The interesting point is, that another host got the same error too with this WU.
The three other results got calculated by classic alberts, but I don't think that means anything.
cu,
Michael
RE: I got an error
)
If you haven't done so already, you should read and respond to the If you have an Albert crashing on 0xC0000005 ... survey thread. I think the cause of the -1073741819 (0xc0000005) problem is still unknown but you may be the first to report having it with one of Akos' apps.
RE: I got an error
)
Welcome aboard, I got the same error:
2006-04-20 20:06:00.2968 [normal]: Optimised by akosf S40.04 --> 'projects/einstein.phys.uwm.edu/albert_4.37_windows_intelx86.exe'
2006-04-20 20:06:00.2968 [normal]: Started search at lalDebugLevel = 0
2006-04-20 20:06:01.1406 [normal]: Checkpoint-file 'Fstat.out.ckp' not found.
2006-04-20 20:06:01.1406 [normal]: No usable checkpoint found, starting from beginning.
2006-04-20 20:08:02.8437 [normal]: Fstat file reached MaxFileSizeKB ==> compactifying ... done.
***UNHANDLED EXCEPTION****
Reason: Access Violation (0xc0000005) at address 0x0040AE2D read attempt to address 0x0114FFF8
1: 04/20/06 20:22:43
1: e:\\einsteinathome\\cfs\\windows_build\\albert4.37\\cfslaldemod.c(1163) +162 bytes (TestLALDemod)
1: SymGetLineFromAddr(): GetLastError = 126
RE: If you haven't done so
)
Done. :)
But I don't believe the reason is found in akos's optimizations.
cu,
Michael
RE: RE: If you haven't
)
I don't think the problem is related to Akos' apps either. The first reported occurences happened with regular Albert 4.37. If it is a coding issue, it's in the base code (and Akos has "optimized" the bug along with everything else). Since it is still an "unknown", I just thought I'd help publicize the effort to gather data on it again.