I personally am running Seventeen Or Bust (and being the Team leader kinda helps in matters too ) after running Seti exclusively for 2+ years!
What drew me in, is the fact that even with just a couple of computers, you can appear to make much more of an impact than with a project like Seti.
This project has so far found 5 new Serpeniski Primes, one by our very own SeanD (Sean2002). There is also a "subset" to this project, which is the ability to "sieve" more numbers, looking for possible candidates for the rest of the project to work on. At this exact moment in time, this would be more suited to the slightly slower CPU if you do not feel the actual client is making as much as an impact as you wish. Your UserID and TeamID are currently being noted as you Sieve, however there are no stats for this at the moment, but soon there will be. This is yet another side to the project, being able to do some of the "background work" as well as the down-and-dirty work of actually factoring looking for Primes.
The project is optimised for each individual processor, and enhancements, such as MMX, SSE, SSE2, 3D-Now! and Enhanced 3D-Now!, which covers the appropriate extras on each CPU series since the original Pentium! This means that the project will use the best of whatever processor you throw it at
If you do decide to run Seventeen Or Bust, I will let you know that the speed (cEMs/sec) is dependent on the n value you are factoring. In general, the higher the number, the faster it will go. However, all systems seem to move up/down the n values together, so you will always be in line with what numbers everyone else is doing.
Of course, the best thing to do, of course, is to try out the projects that you like the sound of, and then run them for a bit, and see which you prefer.
Garry