Hmmmm, there seems to be a little problem with this application: I have now a second WU which was aborted with a computation error:
ERROR: HoughMap.c 388: map index out of bounds: 2973 [0..2970] j:52 xp[j]:113
The workunit can be found under http://einsteinathome.org/task/90382626
Is this a problem of my computer or is there something wrong with the program?
A Happy New Year to all of you!
Jürgen
I've seen a few of these lately, Bernd will have to look into this. We've seen reports where massive overclocking led to exactly this kind of errors, but that doesn't mean there's no bug that can also trigger this ;-).
Bikeman, I have a question for you since you run lots of older machines. How can I tell if a system has the SSE/2 instruction sets?
I have about 15 different cpu types in varying stages of "put together", so I can't just toss this on all of them and hope for the best. I get the impression it won't run on my P3's, but I have a large spread of those as well as a large spread of P4's...
Bikeman, I have a question for you since you run lots of older machines. How can I tell if a system has the SSE/2 instruction sets?
I have about 15 different cpu types in varying stages of "put together", so I can't just toss this on all of them and hope for the best. I get the impression it won't run on my P3's, but I have a large spread of those as well as a large spread of P4's...
Hi!
You are right, it won't run on your PIIIs , and not on your "plain" Athlon (Athlon 64 is fine).
When you look at the list of your active computers:
Bikeman, I have a question for you since you run lots of older machines. How can I tell if a system has the SSE/2 instruction sets?
I have about 15 different cpu types in varying stages of "put together", so I can't just toss this on all of them and hope for the best. I get the impression it won't run on my P3's, but I have a large spread of those as well as a large spread of P4's...
Hi!
You are right, it won't run on your PIIIs , and not on your "plain" Athlon (Athlon 64 is fine).
When you look at the list of your active computers:
I'll report if I see something. There should be an improvement over on at least one of them - I found a system I forgot about and haven't logged into since I installed it in October.
God I love linux, I don't have to babysit anything.
Past an energy-problem (fuse) and fschk.ext3 without faults, i get "Client errors" on that suse10.2-x86_64-host
2 WU's has been cancelled by myself to clear all einstein-related files.
There are some data files that are kept on the disc for several results, so cancelling workunits might not clear all files. If the start of the problem coincided with the power loss and fschk run, I guess it's really a good reason to do the a "Reset" of the project via BOINC manager, which should clear ALL project related files.
RE: Hmmmm, there seems to
)
I've seen a few of these lately, Bernd will have to look into this. We've seen reports where massive overclocking led to exactly this kind of errors, but that doesn't mean there's no bug that can also trigger this ;-).
Happy New Year
Bikeman
Bikeman, I have a question
)
Bikeman, I have a question for you since you run lots of older machines. How can I tell if a system has the SSE/2 instruction sets?
I have about 15 different cpu types in varying stages of "put together", so I can't just toss this on all of them and hope for the best. I get the impression it won't run on my P3's, but I have a large spread of those as well as a large spread of P4's...
RE: Bikeman, I have a
)
Hi!
You are right, it won't run on your PIIIs , and not on your "plain" Athlon (Athlon 64 is fine).
When you look at the list of your active computers:
http://einsteinathome.org/account/83506/computers
You will actually see "sse2" as one of the abreviations listed after the processor type for those PCs that can handle SSE2.
NO GO: Athlon, Athlon XP , Pentium, Pentium PRO, PII, PIII and earlier
GO : Athlon 64, Opteron, Phenom, Pentium 4, Pentium M, Core, Core 2 and newer
CU
Bikeman
RE: RE: Bikeman, I have a
)
Oh, so I can use the pre-existing identification features of the regular client to figure things out. That saves time.
It looks like six of my systems can use this. I'll have to sort out the handful of new systems I want to bring online as well.
OK, I threw it on all the
)
OK, I threw it on all the compatible systems:
4x P4 2.8 GHz
1x P4 2.0 GHz
1x P4 1.3 GHz
I'll report if I see something. There should be an improvement over on at least one of them - I found a system I forgot about and haven't logged into since I installed it in October.
God I love linux, I don't have to babysit anything.
As a FWIW, I have completed a
)
As a FWIW, I have completed a result running on my virtual machine with this version. It validated successfully...
result 90185860
My X2 5000+ has successfully
)
My X2 5000+ has successfully finished 28 WUs so far.
Past an energy-problem (fuse)
)
Past an energy-problem (fuse) and fschk.ext3 without faults, i get "Client errors" on that suse10.2-x86_64-host
2 WU's has been cancelled by myself to clear all einstein-related files.
[edit]
link
[/edit]
There are some data files
)
There are some data files that are kept on the disc for several results, so cancelling workunits might not clear all files. If the start of the problem coincided with the power loss and fschk run, I guess it's really a good reason to do the a "Reset" of the project via BOINC manager, which should clear ALL project related files.
Good Luck.
BRM
I'm seeing an average of 7-8%
)
I'm seeing an average of 7-8% decrease in time to complete units compared with the stock app, w/ a Celery 2.4.