New PC with new CPU and Mother Board. 8 tries and
8 client errors/compute errors. Latest version of Boinc. I run all 4 cores at max when I'm not useing the PC. Anyone know what the problem is? I hate to waste the projects time if all I'm getting is garbage and sending the same.
Copyright © 2024 Einstein@Home. All rights reserved.
Client Error
)
The four failures I could spot in the web stats have 3 different error messages, one is an exceptionally strange "file not found" error after an app restart (it's highly unlikely that the file was really missing because the app had started earlier for the same WU). At least one of the WUs was now successfully completed by another host.
Because the error messages are so diverse, I'd suspect a hardware problem. Is this CPU overclocked? Have you checked the HDD and RAM already? A RAM problem would be my first guess.
CU
Bikeman
RE: RE: New PC with new
)
The HDD is also new and so is the RAM. The CPU is not overclocked. I also run 7 other programs on Boinc and so far all the other ones have not had the same problem. Today I detached and re-attached in hopes that would work. It reinsalled the master file and I'm hoping the problem will go away.
The Master File is just a
)
The Master File is just a little file that tells BOINC how to find the scheduler of the project, it won't work magic.
The errors you had were hardware or software (driver) errors.
Even new RAM can have problems, CPU problems may also cause the above errors, as well as an intermittent error on the motherboard (although those would usually also give you a blue screen of death). Overheating could cause it, in such case try to tell BOINC to run with one less core and see if that fixes it. Clean out any dust, even if you think it's not enough to cause problems.
Driver errors are more difficult to figure out. That's really a long term trial and error. Just saying that you run the tasks of 7 other projects doesn't matter much, they may use different parts of the CPU, or use the CPU in a different way as Einstein does.
Well, the two most frequent
)
Well, the two most frequent answers you get when suggesting something is caused by a hardware failure are :
* No, it can't be, it's brand new
and
* No, it can't be, it worked OK before for a long time and I haven't changed anything.
which implies that hardware never fails :-). Well, it does.
CU
Bikeman
Yesterday I took the
)
Yesterday I took the suggestion and had both my CPU and memory checked out
both in my PC and also removed and checked by them selves. They both check out.
Thought it might be the temp because at one time I had a heating issue but the
CPU runs between 49C to 51C max when all 4 cores are at 100%. I also run Prime95 and discontinued that thinking it was the problem. Still had the failures.
Yesterday I took the suggestion of only using 3 cores and I finally had
3 tasks that made it thru even running Prime95 at the same time. Seems
as though it will run on 3 but not 4 cores.
That would indeed point to a
)
That would indeed point to a heat problem, doesn't it? The S5R5 SSE2 optimized E@H app is really good as a stress test for the cooling system. The Pulsar Search runs quite a bit cooler.
CU
Bikeman