Ok, I've read both of Rom's attempts to dazzle with a lot of gobblygook and fancy footwork and I don't think immediate reporting costs would be all that expensive. What's the difference between reporting after transmission, and thousands of users forcing reporting by clicking on Update? Not much in my opinion... I'm half tempted to forego half the points and use cruncher3's BOINC version. He had the foresight to see the value of reporting immediately to avoid possible loss of result s credits due to any of several possible gliches.

Thanks again for all your help...

It's not impossible, because it has happene to me. I have lost credits due to weather changes affecting my satellite connection and reports being launched off to the great bit bucket in the sky... Those will and do show in "My Grid" as "Not Reported". As I said before, there are numerous possible problems the "could" arise that has the potential of lost credits. Thanks again for your comments.

I see this subject has been brought back again. Can anyone recommend what one business PC running 24/7 should have for network communcations?

I believe that I am using the default ...
.01 days.

As I post this, I notice only one work unit being shown. Before a Boinc feeder went down a couple of days ago, I could see two work units. Once being crunched and the other waiting to be crunched.

Currently, my work unit shown will be completed in 1 hour and 20 minutes.

I am still waiting for the second work unit to be downloaded.

Thanks for any suggestions.

Johnny Cool,

The BOiNC scheduler is slow to adapt to quick WU's and quick to adapt to slow work units, a protection mechanism to prevent overcommitment. If a work unit is nearing completion, BOiNC calculates forward and sees if enough work is there, which is primary. Even if your schedule is set to 1 day, BOiNC would pull new work before the current WU is finished.

It is essential to not change the settings every other day. Some of the statistical tuning takes weeks to settle. Anyway, I've withstood the itch to interfere and I'm always crunching. The Sunday case was a bad rare exception.... WCG has an uptime of > 99.8%.

The 'Ready To Report' (part 2 of a Result reporting cycle) will happen... often before the next WU is requested, but if your buffer is set to 1 day it will take much longer than 0.1 days..... likely just before it pulls the next batch of work. The unofficial BOiNC wiki explains in detail.

Note, that because we have 8 hour FAAH, 2 hour HDC, 1.25 hour GC and 3 Hour HPF2, its very difficult for BOiNC to settle fully. Do only one type like FAAH or GC and it will become very constant....eventually when it realizes u e.g. only do GC, it will pull enough work to cover the 0.1 days or 1.0 days or whatever buffer size u set.

cheers

okay, for the overkill.... the DCF correct's the Total Time to Completion projection of all the WU's in the buffer. If the DCF is going from 1 to 1.1 the TTC's get increased for all WU's by 10 percent, but when a fast one comes thru the reduction is much less than the increase.... it's all to prevent not meeting deadlines. My box has been learning of the 10 or so GC's and the projections are now very close to the actual achieved times.

Anyway, too much detail for most who want to just switch it on, 'set and forget' and truly approach it from the idle cycle view.