draining seti-queue under BOINC?


log in

Advanced search

Message boards : Number crunching : draining seti-queue under BOINC?

Author Message
Profile Mr.Pernod
Volunteer tester
Avatar
Send message
Joined: 8 Feb 04
Posts: 350
Credit: 1,015,988
RAC: 0
Netherlands
Message 309 - Posted: 12 Jun 2004, 22:49:20 UTC

even though I have changed my preferences to:
Keep enough work on disk to last between 0 and 0 days
On multiprocessors, use at most 1 processors

my dual Xeon (HT enabled) keeps downloading WU's as soon as the number of WU in the queue drops under 4 with the following entries in the stdout.txt:
---: Fewer active results than CPUs; requesting more work
SETI@home: Requesting 1.000000 seconds of work

is there a way to drain the queue without destroying the assigned WU's??

Ingleside
Volunteer developer
Send message
Joined: 4 Feb 03
Posts: 1546
Credit: 3,575,760
RAC: 37
Norway
Message 312 - Posted: 12 Jun 2004, 23:51:45 UTC

Till a better solution is implemented, one way that should work is increasing the disk-space-limits. If "Leave at least" is larger than your free space you will continue asking, but will not download more work.

John McLeod VII
Volunteer developer
Volunteer tester
Avatar
Send message
Joined: 15 Jul 99
Posts: 23702
Credit: 493,107
RAC: 137
United States
Message 317 - Posted: 13 Jun 2004, 2:32:36 UTC

The reason that 0 to 0 does not work is to fix a multi proocessor bug. If a reasonable amount of time is set for the queue, and you have a slow multi processor computer (I have 2 dual 200Mhz machines in my mix) one of the CPUs is constantly starved unless a notice of an idle CPU is sent to the server. This notice is sent by incrementing the work time request to 1 second from 0 seconds. As you can see if a single processor requests 0 seconds with an idle CPU, the time is incremented to 1 second, and a WU is delivered.
jm7

Profile Mr.Pernod
Volunteer tester
Avatar
Send message
Joined: 8 Feb 04
Posts: 350
Credit: 1,015,988
RAC: 0
Netherlands
Message 325 - Posted: 13 Jun 2004, 9:21:07 UTC - in response to Message 312.
Last modified: 13 Jun 2004, 9:21:53 UTC

> Ingleside:
> Till a better solution is implemented, one way that should work is increasing
> the disk-space-limits. If "Leave at least" is larger than your free space you
> will continue asking, but will not download more work.
>

"SETI@home: Message from server: No work available (there was work but you don't have enough disk space allocated)"
"SETI@home: No work from project"

the disk-space trick works
thanks :D

Message boards : Number crunching : draining seti-queue under BOINC?

Copyright © 2014 University of California