No fair...
If you want to know how powerfull each quad cpus is, why use very different memory settings to compare? If you want to compare full computers, that's another story but that's not the case here.
How is 299 $ reasonable price, and 235 $ not? I dont buy the "The Phenom X4 9850 is currently AMDs fastest shipping quad core CPU, but it's definitely a budget processor". Definitely a biased Pro-Intel review.

I'm really starting to think that the quad core design has a data bottle-neck of some kind.

I remember Microsoft having an affinity filter that could be used in multi CPU Servers. I got the feeling that it worked by locking a core to a particular job/ process.

Remember back to when the VIA Apollo-133a chipset came out there was no interleaving on the memory and a dual would run slower than a BX board until the fix was implemented.
It was soon learnt that a dual was not faster than a single, but it could do more at the same time

It it with the thoughts above that have me thinking that there is possibly too much confusion of data going in and out of a Quad core CPU.
If only a single application was running you would get maximum results, but our PC's have lots of threads from all sorts of system processors and other software running.
There must be contention in there slowing things down while decisions are made of where its going or where its coming from?

Maybe The Intel design of sticking two dual cores together is the winning point, that dual core is the biggest we can go without loosing too much performance?

After reading Tom's Ultimate Ram Speed Test, I was reminded yet again of how insignificant of a role memory plays in the grand scheme of things. These bleeding edge memory reviews are getting a bit annoying. If people only realized that a low-end $45 2gb kit will trail that $150 set by only 3-5 percent in performance they'd be buying the cheap kit everytime.

If we think logically, most of the people who will buy Phenoms are the ones who already have AMD CPU and have Mb that supports Phenoms. I am one of those,and being that it is my media PC would like to know what is the gain from one X2 4400 to 9550 or 9750?

Not a bad review, except when the decided to do the read and write speed tests.

Graphs starting at non-zero always put me off reviews, they are typically used to show bigger differences than what actually exist, in this case they make a 3% difference in speed and latency look about 20-50%.

They are comparing two chips, one at $299 and one at $290, to one at $235.

The Phenom only costs 80% of the Intel chips, and the performance is way more close than that in some cases, and much worse than that in other cases.

So the thing you are really interested in is what is most cost effective for what you are doing, and the answer changes between the two vendors depending on what it is you want to do with it.