In article <odysseus1479-at-049D3A.21440328102007@news.telus.net>,
Odysseus  <odysseus1479-at@yahoo-dot.ca> wrote:

In article <13i97m3hemdlb50@corp.supernews.com>,
gheston@hiwaay.net (Gary Heston) wrote:

<snip>

That's pretty much my situation. Without a full time connection and the
equivalent of SetiQueue, I can't do much with BOINC.

That shouldn't be the case: BOINC does its own queuing. You can tell it 
you only connect every week or whatever, and it'll download an 
appropriate amount of work each time -- after a 'training' period while 
it figures out how much time it actually gets to run during a given 
interval, and how your CPU's crunching performance differs from the 
benchmarks.

I was feeding a farm that peaked at about 25 systems with SetiQueue from
the one system used for dialin. The self-queueing of BOINC doesn't do
me any good if only one system connects.

Some S@h tasks have deadlines of only a few days (but they range up to 
several weeks), so having only a week between connections could be a 
problem when a batch of these 'shorties' comes out. Most other BOINC 
projects have more consistent deadlines.

However, if you can connect for a few minutes every couple of days 
(which allows at least a day or so to 'call back' if the servers are 
down) you shouldn't have any trouble making use of your spare CPU cycles 
in the meantime.

On the one system, yes--doesn't help with the farm.


Gary

-- 
Gary Heston  gheston@hiwaay.net   http://www.thebreastcancersite.com/

Yoko Onos' former driver tried to extort $2M from her, threating to
"release embarassing recordings...". What, he has a copy of her album?