The UD version is becoming very dated since it is not being upgraded. New projects may or may not run on it.

Even if staying with the 2 existing projects, there is at least enough data left for another year's work on FAAH and HPF2. Just keep crunching.

There's some work being done on an improved BOINC throttle, but it's not there yet. BOINC is changing as needed.

Lawrence I am with you on this subject, I prefer that things be handled from the right end.
Being a Windows user I know the situation here, but what about the two other OSs supported by WCG?
Is Mac (which is so often praised by its users) better in this area?
Same question for Linux? And if not, one additional question: has anybody already tried to bring such a general purpose throttling feature in Linux?

Thanks in advance for additional information. Jean.

I've proposed using TM at the BOINC dev forum as source code can be had, but the obstacle is it only works on Windows OS. It needs working for all OSses. Maybe one day I'll write a true step by step for Threadmaster, even understood by the lesser technically inclined. One without ever to go directly into the Registry (there's really no need). The fast majority at WCG crunches on Windows, so at least these members would be helped.

Don't know how the throttle in BOINC is implemented, but have a feeling it has something to do with the 1 second heartbeat over RPC that BOINC core client uses to communicate with the science and other parts. Didactylos will know. At any rate, increasing RPC would be murder and calling for more issues like firewalls, so I'd suspect.

Hi twilyth,
knreed occasionally mentions this to the CAs. I forget his last figures, but I have a feeling that approximately two-thirds use UD. That is a rough guess on my part.

Scheduling software is something that can only be done accurately once a program is debugged. At this point, we expect to roll out BOINC 6 sometime in 2008. BOINC 6 is expected to be Vista-ready, so it will install on Vista without problems.

BOINC 6 will also redo the graphics. By this time, it is obvious that it is unexpectedly difficult to program graphics for BOINC. So a basic change is going to be made. The problem will then be to reprogram applications to run with BOINC 6, running BOINC 5 applications with BOINC 6 and running BOINC 6 applications with old BOINC 5 clients.

So we have exciting times coming -
Lawrence

LH - thanks.

So despite the volume of posts regarding boinc, ud is the dominant client. That's interesting.

Is WCG serious about converting completely to boinc by mid-2008? Considering that rolling out boinc 6 raises issues with backward compatibility, how will they handle switching over so many users?

Plus, unless the version available when the switch is made is much more user friendly, and essentially bug-free, I think a lot of devices will just disappear. I suspect that will happen anyway once ud goes away but it would be nice to minimize the impact.

This would be a major problem for a commercial IT enterprise. I can't even imagine what's going to happen with a volunteer organization running open source software.