Another Vapor Stat

Add Your Comments

We have the following statement here:

“Well-informed sources have confirmed that Shanghai K10.5, a 45nm version of AMD’s native quad core, and it will be about ten to twenty percent faster than the current K10 at 65nm. This is a clock per clock comparison and this is just as much juice as AMD needed to run faster than Core 2 Duo and Quad.”

Hmmmm, that seems to be an awful lot of improvement out of a process shrink and a little extra cache.

OK, time to start looking for loopholes. Are there any funny words in that statement that can a) be easily overlooked by the average reader and b) warp the meaning of the statement.

There sure is (funny word bolded):

“Well-informed sources have confirmed that Shanghai K10.5, a 45nm version of AMD’s native quad core, and it will be about ten to twenty percent faster than the current K10 at 65nm. This is a clock per clock comparison and this is just as much juice as AMD needed to run faster than Core 2 Duo and Quad.”

Isn’t the current K10 the B2 stepping K10? You know, the one with the TLB feature, and the TLB fix that drops performance more than a bit? And Shanghai is supposed to be the successor to Barcelona, so it’s an apples-to-apples comparison.

Am I absolutely certain of this? No, but after all the misleading statements of 2007, only a fool or a grossly-biased person would take an unattributable, unaccountable statement like this at face value. The proper view of such statements should be extreme skepticism.

If AMD knows enough about the processor to be able to make such a claim, then it knows enough to tell us the basis for the claim. Remember those “Fe-Fi-Fo-Fum” comments a year ago about Barcelona blowing away Intel based on a “theoretical” 2.6GHz Barcelona that still hasn’t and likely never will put on the robes of reality?

Given that AMD only got first silicon around a month ago, it’s very unlikely this claim is based on any real test with a real processor. Indeed, if AMD actually got first silicon running at 2GHz+, that would be more than pretty good news, but AMD has been quite quiet about anything its first silicon actually did. They have just said they were “very pleased” by it, but then they said the same thing about Barcelona and Phenom, didn’t they?

Fool us once, shame on them; fool us twice, shame on us. Believe nothing until you see just what this claim is based upon.

Of course, the real question about K10.5 isn’t “How much more clock-for-clock,” outside of fixing TLB, that was never going to be a big improvement. The real issue is “How much more clock?”

Tell me that Shanghai can run well over 3GHz default, and 4GHz+ overclocked over the next three-six months, and then you’ll get my attention. Show me specific benchmarks from real tests of real processors, and you’ll get even more.

But this kind of unaccountable statement about an unspecified, unclear measurement of a hypothetical processor? This is clown material, it’s now dumb clown material, the worn-out routines of a clown who doesn’t notice the act isn’t working anymore.

Ed


Leave a Reply

Your email address will not be published. Required fields are marked *