backtop


Print 73 comment(s) - last by Calin.. on Jul 27 at 4:05 AM

Native quad-core en route

Yesterday during AMD's Q2'06 earnings conference call, AMD's President and Chief Operating Officer Dirk Meyer recapped the long term plans for the company.  Although the bulk of his comments were already stated in during the June AMD Analyst's Day, Meyer also added the tidbit that the company plans "to demonstration our next-generation processor core, in a native quad-core implementation, before the end of the year."  Earlier this year, AMD's Executive Vice President Henri Richard claimed this native-quad core processor would be called K8L.

Earlier AMD roadmaps have revealed that quad-core production CPUs would not utilize a native quad-core design until late 2007 or 2008. To put that into perspective AMD demonstrated the first dual-core Opteron samples in August 2004, with the processor tape out in June 2004.  The official launch of dual-core Opteron occurred on April 21, 2005.  On the same call Meyer announced that that the native quad-core would launch in the middle of 2007 -- suggesting the non-native quad-core Deerhound designs may come earlier than expected or not at all.

Just this past Wednesday, Intel one-upped K8L plans by announcing quad-core Kentsfield and Clovertown will ship this year, as opposed to Q1'07 originally slated by the company. 


Comments     Threshold


This article is over a month old, voting and posting comments is disabled

Conroe "needs" a memory controller?
By Anemone on 7/23/2006 9:42:36 AM , Rating: 4
Quad core in development for 4 years, probably so. But K8L in development for that long, doubtful. If so they'd have a working sample now. And if they did, they'd have shown everyone who'd listen what it could do and how long till we'd have it.

Now I do know how long this race has been going on, and yes I do very much believe that K8L was the fastest answer to Conroe. They've known about what Conroe was likely to do for almost 2 years now. Everyone knew it was a leap beyond Yonah, and everyone knew Yonah was a mild leap beyond dual P-M's and that being a leap beyond P3's. And if you go back in time enough you'll remember that the old P3 vs Athlon race that was going on, really never died, it just stopped by the sidelines for a breather.

AMD is no dummy as a group, but they didn't exactly invent the onboard memory controller, thank you Alpha. There are tradeoffs to the technology. But that's not what I came here to note.

AMD is behind, partly because they expected Conroe to be a faster clocked P-M dual core and that would have been something to just match clock for clock with the high end A64's. They've been caught napping on their laurels which is something AMD has done before. They come up with a good and superior level of performance at a given price point and then suddenly "advances" either in speed upgrades or process technology go from 6 months out to 12 months out. Just have a look back and you'll watch each time they've been ahead has been followed by a target slip, where sometimes they fall behind and sometimes they don't. But it appears like they take a breather after each rush to leap forward. That's natural and certainly not something other industries don't see as well.

However in this case it came at a bad time, and is compounded by a misjudgement of what Intel's likely next gen was going to be. Intel's chip is better than almost everyone thought it would be and it clocks higher than almost everyone thought it would. It doesn't trip over high latency memory accesses and a quad core version running in excess of 4ghz is already sitting testers hands. Even if you say a 3.2ghz quad core K8L comes in reasonably well, the Athlon itself isn't ready to scale past 3.6ghz or so even at 65nm. They'll manage a few process miracles and perhaps get to 3.8 or 4.0 by the end of 2007 or mid 2008 but that is barely on the edge of what is possible.

Meanwhile you have chips right now on air that can achieve 4ghz status. Can you even concieve what these chips might do under a well underway 45nm process in late 2007? At that point why bother with CSI? Just tap all 4 cores onto the same 8mb dynamic cache at 45nm and run that at 4.5-5ghz. Need more memory bandwidth? If quad pumping could live this long, just move the server issue downstream and put dual 1333 busses onto this chip. Conroe sips memory needs very gently and hasn't been badly underpowered yet by the current structure. It will need more at quad core, but I think folks vastly overestimate how much more it will need. Dual and quad memory busses would do the answer for a year or year and a half and then you could add a built in controller as you had a mature 45nm process or as you moved to the next level. At that point you'd be wondering what to do with all the die space.

Is AMD out of the game? No, not really. Did they misjudge what the Intel response was likely to perform at? Yes, absolutely they did. Do they have designs that will help address this? Yes, if nothing more than taking a few of their own and then copying some of the ideas that went into Conroe. You think Intel would be the only one to copy AMD 64 and built in memory concepts? Nope! It works both ways for both parties. But don't think you are going to see a sudden miracle from quad core. It's going to be expensive and right now at least to the paying customer (even if they are underpricing) Conroe is really cheap. So they have to be able to offer a 3.4ghz quad core that can compete with a $999 quad core E6700 in the form of Kentsfield. And it has to be shipping in Q1 07. I don't know about you, but if I believed I was going to be shipping a product in Q1 07, I would probably have testing samples in the field no later than Sep 06. If you are a wagering person, then you can offer up the odds on that happening.

I think we'll just see.




"If you can find a PS3 anywhere in North America that's been on shelves for more than five minutes, I'll give you 1,200 bucks for it." -- SCEA President Jack Tretton

Related Articles













botimage
Copyright 2014 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki