This is probably not the best thread to ask your questions in, but I'll try as best I can to answer anyway.
Right now Team 2630 is producing slightly above
500000 points per day. If you ran a Q9450 at 3200 MHz as a dedicated folding machine with no GPU help you could produce over 4000 Points Per Day before the recent updated folding cores were released. Now, easily in excess of 5000 PPD, probably in the range of 5500-6000 PPD. That would mean your machine alone would be more than 1 per cent of our current output. Of course, this would be running 2 SMP clients with each getting 2 cores.
As for the 4870 GPU, I think they produce over 1000 PPD, but I forget the numbers people in this forum have reported. You can search on 4870 if you want. Over at the folding forum there's
a chart that makes claims of around 2000 PPD, but i'm somewhat suspicious of it because of the numbers on the nVidia chart in the same forum (different subforum).
The only way to use the GPU client is in Windows. The best way to do that presently appears to be under Vista, because it seems to use a lot less CPU time than Windows XP does.
Pegasus has been running a Q9450 and an nVidia GPU, and I think he's using VMWare to run the Linux SMP client on cores 0 and 1, and then another Linux SMP client on cores 1 and 2, leaving core 3 free for the GPU. Or something like that...
Your other Intel box with the E6550 in it, you could probably fold on it and lose little responsiveness, but it would produce more heat and draw more electricity. It's up to you. You can always test it and see how you like it. If you're running Linux it is much more reliable than what people report about Windows SMP clients.
You could run the standard console client on your workstation. They have long enough deadlines (31 days for the ones I'm currently folding, for example) that they'd still be useful, if not a huge difference. But many small contributions can make a difference. Every bit helps.
