AMD to counter Nehalem in Q4

[Andreas G]

AMD launched its first quad-core architecture, also known as Barcelona, back in September last year. Later on we also got a desktop version which we now know as Phenom. Much of the pre-launch information focused on the poor frequency scaling, which unfortunately turned out to be correct. Phenom is far behind its Intel counterparts in terms of frequency, and now that Intel is on par in terms of IPC (Instructions Per Cycle) AMD has to compete on other terms than pure performance.

AMD managed to beat Intel, and NetBurst, with the K8 architecture, but over the last year, the roles have been reversed and Intel is back on top. With another microarchitecture coming this fall, Intel is certainly enjoying being in the driver seat, but AMD intends to take over with the launch of the 45nm Shanghai core later this year.

AMD sources indicate that since the biggest updates of Nehalem are the integrated northbridge and the new interface QuickPath Interconnect, it has merely moved up Intel to where AMD has been for years. They did admit though that the improvement going from NetBurst to Core was an impressive one, although expected.

Even though we did not get to see any actual performance data, or even estimates, we were told that AMD is not especially scared by Nehalem. It considers HyperTransport to be equal or even better than QPI, while what it now has to do is find ways to make its processors run faster, not just higher frequencies but also get more work done per cycle.

Both are suppose to come when AMD starts shipping its 45nm processor Shanghai. Shanghai is the server processor and should go out to first and second tier partners in Q4. The new processor will sport 6MB of L3 cache which should prove beneficial in many scenarios. Previous tests with the Deneb core has at least pointed to this.

Unfortunately other sources indicate that Shanghai, and Deneb, will not hit significantly higher frequencies than Phenom. It should reach higher than Phenom did, but the roadmaps we've seen have only shown as high as 2.8GHz and that's not enough if we don't see more than what the Deneb preview showed us, which we very well may when server benchmarks are posted.

[KTE]

2.8G is also what I'm banking on as highest by Q1.

Nice overview Andreas, very accurately worded (unlike most)

[Delph1]

Thank you, coming from you it means a lot

AMD is surprisingly confident actually, which surprised me.

//Andreas
_________________
"I'm sort of a postmodern vegetarian, I eat meat ironically"
"You are our Swedish superhero "

Forget everything, knowledge is your enemy, it's only going to make you second-guess yourself."

[KTE]

Ah, you're welcome mate, its well deserved

Sifting through all the technical and convoluted online crud is very technically challenging to ascertain the facts from fiction. Most users would end up failing the task and very very rarely would you have one that doesn't have daily experience of both sides yet can have a balanced and factual view of how everything currently stands in the market. Requires magnitudes of intelligence and impartiality sans sensationalism (i.e. Ed).
Your article, as much as you may not have imagined, spread quickly and far on the net since it was the best, most precise and concise summary I have seen to date without any words which fans of either side or technically knowledged may pick on. To anyone who wants to know the past 10 months of desktop CPU info and the current position with the most possible future outlook, I could not have said it better.

You know I will say what I see as fit and am not one to hold back praise for anyone where I believe it is deserved and is encouraging to the receiver... that's some very good experience and understanding of the IT system around the net you've shown right there in one article Andreas.
It's better than any of my lengthy posts!

Re 45nm.. I'm quite confused. You're right, AMD is acting very confident, but why exactly?
The end of this post will rapidly briefen up my current opinions on their position: http://www.ocxtreme.org/forumenus/showpost.php?p=53435&postcount=306

The only thing I didn't mention for fear of length (as usual) is that AMD will always start low. If they have 3G tops by the intro, that means they will be going upwards a few speed grades before end 2009. They are also rapidly shifting to 32nm hi-k/mg with first samples due out anytime end of 2008.

In honesty though, IDK. What I've heard since 2 months internally re 45nm and what I've seen out in unsubstantiated reports/leaks do not reconcile and I don't have inside visits, so IDK. I will point out though, Digitimes ran a preliminary story in June 2008 about Phenoms delays/cancellations against the commonly accepted and substantiated views, and it turned out near 100% true: http://www.digitimes.com/mobos/a20080611PD204.html

They have ran another story quoted from Xbit Labs of only 2.8G/3G 125W Denebs in Jan '09 now. So it could very well be true. It is plausible but it isn't ground breaking in terms of improvement and with poor delays rumored.

The FX model always existed even with 65nm as MB CPU support lists will show, but they can't appear unless a) they hit higher frequencies than average and b) with high prices. The latter point, as with Intels QX range, is why I would ignore their existence.

[Delph1]

[KTE]

Yep. They won't come unless they oc well.

The last thing AMD wants is to tarnish the FX model rep. The last real DT FX was FX 62 and at the time it released, it had no match to it. I highly doubt AMD would release FX again unless they can match that.

They would require high MHz to start with. At least above 3G 125W... which I doubt will happen before April, if at all. There's only room for a 140W bin above which would be bad enough to cool and clock.

[Delph1]

Alas, true.

AMD to go bulk ... would be interesting if true, but not going to happen

//Andreas
_________________
"I'm sort of a postmodern vegetarian, I eat meat ironically"
"You are our Swedish superhero "

Forget everything, knowledge is your enemy, it's only going to make you second-guess yourself."