>>You mean this is like the 8K memories for the old Spectrum - actually 16K where one side didn't work so Sinclair got them really cheap? I.e. a quad core processor where one of the four doesn't work? Wouldn't be anything new - even to me. I wonder why is that, they lock one to have a processor which fits the price niche, or they are selling a series of quarter-faulty quads as trios?
>
>A bit of both, I suspect. http://www.extremetech.com/article2/0,2845,2320531,00.asp
>
>One supposed theoretical "feature" of a 3-core CPU is that there is always exactly 1 communications hop between any 2 cores. This can simplify cache coherency and other inter-core communications. This may not be implemented in the real Phenom triple-cores if they're actually quad-cores with one disabled.

Well, frankly I don't really care whether it beats Intel in this game or that - it's irrelevant for what I do. It's not my dream machine (that'd be a Mac :), but it's the one I'm happy about - quiet, does what I need, doesn't draw attention to itself and looks grand jury rigged. If its heart is a three-quarter, so what :).

BTW, nice tech website - just about 20% of the raw text (as _cliptext measures) on the page is actual content, the rest is headlines, menus, links to other parts of the website where you can find more links to other parts of the website where you can find more links to other parts of the website where you can find more ... and that's with AdBlock+ on. Actual length of the source (again, with most of the ads filtered out, and not counting the graphics) is 60K of HTML per page, where only 1460 bytes were the article itself - and the article has seven parts. So that's half a meg of html text to read 10k of actual text. IOW, about 2% of the bandwidth is the info we came to read; the rest is website's self-advertising, markup, eye candy and... did I say that I didn't count the ads?