You don't say which interface you are running. For the UD interface, there is an add-on product called UDMonitor to generate multiple devices on the same computer. You can run several workunits offline before calling the internet. They are swapped in/out of the active UD directory from/into cache slots in subdirectories.

UDMonitor takes a bit of babysitting. The BOINC interface is even easier.You can tell it how many days to stack up workunits. The maximum is 7 but no more than 4-5 is recommended. 1-3 works well. Uploads of previous workunits/downloads of new workunits even take place in parallel while the active workunit is running.

I also have dial-up. My own computers run offline 24/7. They could hibernate for several days if I weren't using them for something else. I can go away for the weekend and a bunch of completed workunits will be waiting to upload when I return. The FightAids@Home project is the most compatible with dial-up since the workfiles are very tiny.

When you set up a separate directory for UDMonitor, you only have to tell it where UD.EXE is.

When you start playing with UDMonitor for slot #2 and above, remember to tell the WCG registration box you are already a member for each new device (cache slot). It's just like adding a new computer.

Don't make more slots than you can finish by the deadline. Start with 2 or 3 to watch the function of the program for a few days. There are a lot of little options in there, and the position of the cursor in the UDMonitor display influences the list of activities you can get. You'll figure it out as you go along. A right click of the mouse is used frequently.

The only other wrinkle is that UDMonitor must be started for execution each time with the WCG BAT file since it defaults and was written for grid.org.

If you want to try BOINC, all you have to do is exit UD and install BOINC and add worldcommunitygrid.org as a project. Don't run UD and BOINC at the same time. There's a large BOINC id in your WCG my grid that you may need to copy and paste the first time only.

There are a lot of threads around about ud monitor, but especially the UDMonitor thread on the http://forum.grid.org They have a lot of experience with questions and I have used UDMonitor myself. The internal filenames are different between WCG and grid.org, but the function of UDMonitor is exactly the same.

There's another gotcha'. After awhile you may forget some of your old device names. Names must be unique across all slots, devices and computers for your membername. The registration process will gripe if you try to pick a duplicate, and it is guaranteed to cause problems. It defaults to only a week or two. When looking at your WCG device list, use the picklist to select all devices ever registered and all devices that ever sent in anything.

I just thought of something else. You can set UDMonitor to do nothing after the set of workunits is finished. That might keep UD from calling in and ruining the hibernation....or maybe not. I liked to have several workunits available at all times and call in more often so it would actually never complete. I would then switch the slots in and out and send stuff in at that time for the ones that finished. I don't recall exactly what happens if UDMonitor ran out of workunits since I tried to never let it get there. It would be easy enough to test.

With UDMonitor, it is especially easy to see whether you have passed a checkpoint. You don't even have to watch the pretty screen lines and graphics to tell. If the flask icon is red, it has been awhile. If the flask icon is yellow, it has been a little bit, but not a lot. When the flask icon turns blue, that's the checkpoint and disk saves were just done. That is the best time to turn off the computer without wasting a lot of time resuming work later. All of this is also reflected in the nice UDMonitor log, and the wait times can also be adjusted.

The current version of BOINC will not tell you where the checkpoint is, but future versions are working on it.