JETSOLVER,

What Sid2 says is indeed correct.

Let's presume your laptop spends most of it's time working on fictitious project x. You know from past experience, your WU's take between 1 - 4 hours to complete one project x WU.

You know your machine will only be powered up say 4 hours / day. So this dictates how many additional days of buffer you can handle.

Presuming your WU's take 4 hours to complete, should your laptop run continuously, you could process between 5-6 WU's / 24 hours of time.

So if you can anticipate how much time your laptop will be able to process whilst not connected to a network, then setting a buffer to some quantity of days will do you nicely.

Of course, presuming you get a WU that is on a short leash (meaning needs to be back withing something like 2; 4; 6 days or shorter) you might process the WU and return it only to discover it was returned later than needed.

I'm sure one of our nice helpers here would be willing to step in and articulate all the things (good / bad) that happen under these conditions.

have fun on your trip

1. HCMD2 Parent tasks: 14 day deadline,
2. HCMD2 Child tasks: 10 day deadline,
3. HCMD2 repairs tasks: 40% of deadlines 1 or 2 above

Latter only send to clients that return tasks continuously within 48 hour basis (not confirmed!) and are known from last ~15 results to produce valid results.

Never take more than you can chew. If switching from a 24-7 crunching pattern,m short buffer to a part time pattern with larger buffer, BOINC needs to learn, thus set additional buffer to the total days the PC is thought to be on times the number of cores times the factor of throttle. Example:

Away 5 days
On, using mains, 10 hours
Throttle 60% (% of CPU time while on)
When in use: Yes

5 x 10/24th x 0.6 = 1.25 day

BOINC automatically will multiply this with the number of cores in a computer allowed to BOINC. 2 cores in the example would fetch thus ~60 hours of work, based on the known CPU efficiency and Duration Correction Factor (plz don't roll eye's)

If you did take more than the computer can chew, you might see eventually High Priority processing. With other safety factors that BOINC has compulsively build in, such as always wanting to return work 24hours before deadline, best not exceed 7-8 days.

How not to get rush jobs for a planned trip: Increase buffer to 2.5 to 3 days several days before departure.

Note to self: hmmm, I think WCG might want to add an option to the device profile for those with 'mobile' crunching behavior: Don't send rush work. (Maybe knreed reads this)

Steve, don't get confused between WUs deadlines and the cache size.
As far as I know the max cache size is still 10 days, whatever the deadlines and the projects.

And by the way, Sek and Steve, the HCMD2 parents deadline has apparently (no formal announcement) been set to 10 days too, since June 28 or 29.

Cheers. Jean.

So, all we need is a scheduler that works as expected.

Not much to ask for, right?