Martin 53N 1W wrote:

jahinton@omuonline.net wrote:

I xfered over to BOINC and checked my stats and noticed that they
were almost 200 WU's behind. Have me down for 1289 and I'm @ 1421
using the old client (via SetiCommander). Why is my BOINC data
tranfered in seconds instead of WU's and what if I go past my
deadline? (ie, computer can't connect 24/7)?

Phew... Questions, questions!

The 'classic' WUs listed on Boinc are from a 'snapshot' taken some
time ago. The latest up-to-date numbers will appear whenever another
credits snapshot is taken from the database. One will be taken at
least when s@h classic finally closes.

The Boinc scheduler makes requests in whatever number of seconds of
work is needed to keep to your working 'cache' settings. The servers
then decide whether or not to give you a WU to keep you busy. (WUs
across the projects can be vastly different in the time required to
process.)
Too late past your deadline and your result is assumed dead. You get
zero credit. Also, someone else will have been given your WU to work
through. Set your cache size to the /minimum/ for what you need.
Roughly, this should be x2 the longest you'll be not connected, but no
greater than 7 days worth. Reduce the numbers if your machine is not
powered 24/7!

I beg to disagree. I had it set to 7 days. Then they unplugged the Ethernet 
cable and forgot to plug it back in again - I was on leave. It then ran out 
of WUs within 4 days. Some of the WUs were abandoned after a few minutes - 
rather than taking the usual 4 hours. (This is quite normal.) You have 14 
days to return the results. Now I have it set to the maximum 10 days. Of 
course the downside (?) is that most of my WUs will now just get the average 
credit of the previous three results which arrived earlier.

The other problem with a minimalistic cache is that if there are problems at 
their end, you will soon run out of WUs. One solution of course is to run 
another BOINC project as well (e.g. CPDN or Einsein).

Brian