I don't think the multiplier is high enough. Why don't we just get rid of the fsb (or what ever you want to call it now) and just use the multiplier.
This way we can make it as high as we want.
30, come on you got to be kidding me.

I'm not so sure about your "ray tracers where each thread takes up the whole core without much benefit from HyperThreading" : As far as I know, ray-tracing can become quite memory bound for high complexity scenes and/or incoherent rays. In that case, HT could help alleviate cache-misses, wouldn't it ?

Congratulations to Intel.
It is a first big step towards cluster 4 core supercomputing on a new level making a great possibility to visualize by medical doctors a human internal organs with new enhanced resolution.


Go back only 2 1/2 years ago and AMD had higher margins than Intel with their lousy netburst designs. It is truly amazing how far Intel has come in such a short period of time. Either that or it is truly amazing how poorly AMD has performed over the past couple of years.

please do a follow up with some emphasis on HT. heard HT in Nehalem was good. a confirmation from theinquier would be nice.

@waxwing
.8GHz = .2*4GHz - tho agreed it's scaled up 25% 

what i wonder though is how hard it's going to be to get that 3.2GHz machine to run at 5GHz stable

??

Why compare an overclocked result for i7 to a stock result for a QX9770?

Plus 5544 / (3954 x 1.25) = 1.122

Or 12.2 % faster clock for clock on a single thread, and 16.7% faster on the multithread.

Why is Cinebench reporting 8 cores if HT is disabled?

"Even if you scale these up linearly 20% for 4 GHz"

0.8 GHz is 25% of 3.2 GHz