Hello whocrazy,
The Athlon64 broke up the x86 instructions into 3 streams of micro-code. The Core2 breaks them up into 4 streams - - even faster. The poor Pentium 4 is left behind in the dust.

Lawrence

Imagine if all those quad cores were i7's:

Intel Core i7: The Essential Guide

Some benchmarks of the i7 running HCC
Movieman has posted some early results crunching HCC on the i7 965 he has been trying, over at XtremeSystems
The length of HCC WUs is fairly constant, so I compared times from one of my Yorkfield (LGA775 45nm) quads here
From these very limited statistics, when the i7 was running with HyperThreading disabled, it ran 4 WUs at once, and was about 12% faster than the Yorkfield, clock-for-clock. With HT enabled, it ran 8 WUs at once, but they ran more slowly. However, the overall throughput was then 27% faster than the Yorkfield, clock-for-clock.
I've read that HT on the P4 makes WCG throughput slower (but JimBoullier - 2 posts down - has found otherwise for HCC), but on the i7 it seems to be the biggest factor in increasing WCG productivity, at least on HCC. And HT may work even more effectively when running a mix of different projects.
Don't be disheartened by your choice of a Q6600, though. They still give about most crunching bang for your buck.

I have done such a test recently with my P4 HT 2.4 GHz and I have found 10.4 hours when running HCC WUs one at a time, and about 15.5 when running two concurrently. Which would show that HT is finally giving a better global WCG throughput when running two HCCs at the same time. I was not expecting that, so my measurements are not biased. Note that HCC is much different from other projects (more integer calculations than usual, I think), so what is true for HCC might not be true for those other projects.
HCC is also running much faster (25 %) in a Q6600 in 64-bit than in 32-bit, which is not true for other projects.

Regarding crediting (I have looked at the threads that Rickjb has given) I confirm that the way HT is screwing up measurements remains the main problem for grid computing. However it appears that what is finally granted seems close to what it should be. Remains the problem of runtime (for WCG only) since this P4 HT is now returning 2 days of runtime per day while it is doing "only" about 34 % more work.

Cheers. Jean.

I run several 1.6 and 1.7 gz P4's. They do HCC right around 17 hrs. very consistently. None have HT. They were pieced together from several throw away machines. I can only look longingly at the quad cores at the moment. Those i7's look mighty nice, but at what price ? Perhaps at tax refund time ???????????? (Yeah, I know I could save on the electricity, but they help to heat the basement )

Cheers