computation error on completed tasks

Finomenn
Finomenn
Joined: 30 May 08
Posts: 1
Credit: 238
RAC: 0
Topic 193956

When I finish a task my computer keeps saying computation error instead of completed, and I don't get any credit for doing the tasks. This is very annoing, and if it continues I will detach from the Progect.

Jord
Joined: 26 Jan 05
Posts: 2952
Credit: 5779100
RAC: 0

computation error on completed tasks

Uhm, OK... a computing error means something bad is happening on your system. How would you detaching from this project help in that?

You are getting:

too many exit(0)s

So the big question is, do you use BOINC's CPU throttling on that system?
On multiple CPUs or cores or hyperthreading the CPU throttling causes problems. This is going to be fixed in a future version of BOINC (even BOINC 6.2 has problems with it, as will 6.4, as far as I know).

Set the throttling option to 100% and use an external program such as BES or Threadmaster to throttle the CPUs.

metalius
metalius
Joined: 29 Dec 05
Posts: 44
Credit: 172607243
RAC: 24804

I think, I have sometimes a

I think, I have sometimes a similar problem.
A fragment of stderr out:
...
done.
2008-10-26 07:45:51.4714 [normal]: done. calling boinc_finish(0).
called boinc_finish

h1_1063.40_S5R4__860_S5R4a_1_0
-197
user requested transfer abort

h1_1063.40_S5R4__860_S5R4a_1_0
-197

]]>
Is it possible to avoid such upload errors?

Alan
Alan
Joined: 7 Jan 08
Posts: 5
Credit: 1189195
RAC: 0

The last several runs with

The last several runs with S5R4 have come up with "Computation error" after completing 100%. I am running BOINC 6.2.18 on an E6750 with 4GB ram and a minimum of 1 TB HD free. I am also running:

Malariacontrol
Rosetta
World Community Grid
SETI
Cosmology &
LHCAhome

with NO problems. Although I am running both processors, I have % CPU time set to 50%.

What is the problem?

Regards,

Alan

Byron S Goodgame
Byron S Goodgame
Joined: 16 Jan 06
Posts: 187
Credit: 56581
RAC: 0

RE: The last several runs

Message 85786 in response to message 85785

Quote:

The last several runs with S5R4 have come up with "Computation error" after completing 100%. I am running BOINC 6.2.18 on an E6750 with 4GB ram and a minimum of 1 TB HD free. I am also running:

Malariacontrol
Rosetta
World Community Grid
SETI
Cosmology &
LHCAhome

with NO problems. Although I am running both processors, I have % CPU time set to 50%.

What is the problem?

Regards,

Alan

There is a known problem with BOINC and multiprocessor system when using anything other than 100% cpu time. You should set that back to 100% and if needed for heat issues, set the preferences to only use one core.

michbret
michbret
Joined: 14 Oct 07
Posts: 1
Credit: 5169
RAC: 0

Hi, Same problems! All the

Hi,

Same problems! All the works end with "computation error" after completion

OK! I'm using hyperthreading and running 2 Boinc tasks at the same time and also using Boinc CPU capping feature ... But, why only Einstein fails ???

Boinc is last official version (6.2.19). System is winXP SP3 on P4 (32b) CPU

Never got any problem with SETI, POEM (e.g.). However, I often ran into trouble with other astrophysics applications (Einstein, Cosmology, ...) so I had to cut off these projects since they were wasting time (energy and some good will ...)

By the way, I hove science will progress even with those minor hick-up.

Regards

Byron S Goodgame
Byron S Goodgame
Joined: 16 Jan 06
Posts: 187
Credit: 56581
RAC: 0

RE: Hi, Same problems! All

Message 85788 in response to message 85787

Quote:

Hi,

Same problems! All the works end with "computation error" after completion

OK! I'm using hyperthreading and running 2 Boinc tasks at the same time and also using Boinc CPU capping feature ... But, why only Einstein fails ???

Boinc is last official version (6.2.19). System is winXP SP3 on P4 (32b) CPU

Never got any problem with SETI, POEM (e.g.). However, I often ran into trouble with other astrophysics applications (Einstein, Cosmology, ...) so I had to cut off these projects since they were wasting time (energy and some good will ...)

By the way, I hove science will progress even with those minor hick-up.

Regards


It seems to be a software bug in BOINC. They are working on it and will have a fix for it in a later version, though from what I've read, it won't be in the 6.2x.

It appears to affect systems differently and on different projects. I had the same problem in Einstein myself, (though only once since I changed it right after and no errors from it since) and never had a problem doing the MB and AP over at SETI before that time. I assume though if I had left it at the 75% cpu time I had it set at, the other project would have eventually had the same effect from it. I've seen a number of folks over at SETI with the same problem also.

Delete me
Delete me
Joined: 1 Jan 09
Posts: 3
Credit: 0
RAC: 0

If this is a problem

If this is a problem specifically with BOINC, why is it only giving errors when crunching Einstein WU's and WU's from other projects get crunched without error?

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.