Subject: Re: Basic Boinc facts.
From: f/fgeorge
Date: 17/09/2005, 14:21
Newsgroups: alt.sci.seti

On Sat, 17 Sep 2005 05:58:19 GMT, Bill Jillians <"\"Bill
Jillians\""@virgin.SPAM_NO_THANKS.net.SPAM_OFF> wrote:


Hello Eric, thanks for the comprehensive reply. 
I have no idea how this scoring system works.  I have now submitted 
about 6 WUs but my total score is 190 and my average score is 10 or so. 
What do these numbers mean?
Every workunit currently only has so much Science in it, each machine
crunching the same unit should ideally ask for the same amount of
credit when crunching the same unit. In reality that does not happen,
has alot to do with the fact that your machine is doing this while
mineis doing that and someone else has an Apple and who knows what
they are doing. So when we crunch the unit other things are happenning
at the same time, the numbers are also skewed because there is no
standard on how each type of OS should track of those clock cycles.

Also the Statistics tab doesn't exactly 
produce detailed information, just a graph with (at present) two points 
on it.  What is this supposed to tell me?
This will expand as you return more units to show the general
direction that your computer is heading in the different fields.

Presumably there will soon be an updated version of Boinc just like seti 
classic addressed bugs with new versions when it first came out. 
New versions are coming out about once a month, a major one is
supposed to be coming soon. The highest right now is 4.45, go here to
see the list: http://boinc.berkeley.edu/download.php?dev=1
4.19 is the lowest, in Seti, that you can currently use.

Also 
it encourages you to join other projects when seti is having an outage - 
projects which do not have the same appeal or subscribers as seti but 
may be producing just as significant results (I note the climate change 
model that was mentioned in the paper a few months back).
Now I've got it running I'll stick with it to see if things improve.  It 
certainly isn't as high maintenance as my old floppy network was ... 
plus the overhead of waiting several minutes for a WU to finish before 
changing it are now gone.
That was the whole idea.

It would help if when a WU finishes processing 
the next one could start automatically but it seems some regular 
checking is required to keep the system working.
If you give it a chance it should go ahead and start a new unit
automatically, as long as there are units in the cache. Mine do.
If you do not have enough units in the cache consider extending it out
to a few days or even up to the max of 10 days if needed. I currently
have mine set at 4 days and some machines have over 20 units in its
cache while other have much less.