mine is set to 1450 minutes (24 hrs and 10 min), for a specific reason of course. on my office machine, i run 2 CPU projects - LHC@Home & Test4Theory@Home. b/c the new T4T@H tasks are limited to running on a single core (for now), that means 5 LHC@Home tasks are running concurrently at any given time. if i leave task switching set to 60 minutes, this host will crunch 1 T4T and 5 LHC tasks for 60 minutes, and then switch to 6 LHC tasks. sometimes the T4T task resumes after 60 minutes, but more often than not, it remains in the queue "waiting to run" while 6 LHC tasks continue to crunch away. having changed my task switching to 1450 minutes, my T4T task now has the full 24 hours it needs to complete, plus an extra 10 minutes for any calculations that might (and occasionally do) run over the 24-hour mark...usually it only runs a few seconds over, and more rarely a few minutes over, so i thought an extra 10 minutes might be a nice safety margin. having the task switching set this way also allows a new T4T task to be downloaded and started almost immediately after completing, uploading, and reporting the previous task. with task switching set to 60 minutes, although a new T4T task might get downloaded just afrer the previous task is reported, there's a good chance it won't start running right away, and that it'll sit in the queue behind a decent chunk of LHC tasks.