If my memory serves me correctly, there was a method posted in the forum probably about 10 years ago or more concerning the issue of more profiles. Someone posted a how to list of adding additional profiles, totally unsanctioned by WCG.If I get the chance I may go searching for it and see if it is still possible.
Found it here
Cheers

I knew about that and had created a special profile for my sister's computer years ago, but it didn't show up in the pull-down list since it is static rather than dynamic. I did not feel comfortable hacking a local copy of the web page. I felt like I was doing something prohibited.

Has WCG said that they don't mind us doing that?

Cheers

I found this thread because I'm having the same problem as the OP. Sometimes this is simply because there are other WCG tasks (perhaps even other FAH2 tasks) running at the time of the download, so the newly downloaded FAH2 task has a delayed start that makes it impossible to return by the deadline. Sometimes it's because the FAH2 task is simply too long, even on a fast machine. The net result is that FAH2 requires a lot of hand-holding (I'm not wasting CPU cycles on a task that I can clearly see wont complete by the deadline), disrupts other WCG tasks (because they are pushed too far down the priority list to complete in time), and makes inefficient use of server bandwidth (because of all the resends).

The deadline is not a target, any more than a speed limit is a target. Even if the deadline was set to six years, my PCs would still return many results within a few hours, with some taking a little longer. Surely better that than resending over and over until days later someone completes by the latest deadline. And better that tasks take a little longer to complete than having people give up on FAH2 (I've already turned it off once, and I'm considering doing so again). IMHO, FAH2 is an example of "more haste, less speed".

Return time is a BOINC metric. Why not use that to determine if it is wise to send a task to a host? And then just accept that sometimes a task will get delayed by a few hours even on a fast host.

I don't know. I imagine the scientists who are running the project have considered the implications of the 24 hour deadline and have good reason for it. I think you opt into a project under the terms offered or pass. Although it is rarely necessary for me, I have found that you only need to momentarily suspend projects to start others.

I've got 2 i7 boxes going right now, each one can finish one of these in a couple hours at the most. The problem isn't really with the work units themselves being too long, but rather, anyone who has a queue length (buffer) built in to accommodate system downtime will run into problems.

With the projects being run currently, running with 24hrs worth of work on a machine at any given time does not take up much drive space at all, and allows for nearly any conceivable maintenance window or high severity issue at WCG to take place and be resolved without interruption to processing on the client end.

A 48, or even 36 hour deadline as opposed to 24 would eliminate the issue. Heck even 30 hours ought to work.

For the time being, I've had to reduce my queue length from a full day down to about 3/4 of a day.