AMD's best is still to come

[mito]

nice article...

 

[Tahir2]

A reader recently shot me a note saying that after studying some of my work related to Advanced Micro Devices, he has spotted a pattern: I always side with the underdog. This reader crystallized the prevalent viewpoint, one expressed by most of my colleagues, that AMD achieved its market success by dumb luck, with the “dumbâ€￾ supplied by Intel. Intel discovered too late that NetBurst was, well, everything I’ve always said it is, and hubris born of market domination left Intel without a Plan B.

To say AMD's achievements are dumb luck is oversimplifying tings and not giving AMD credit where it is due. AMD have been executing brilliantly over the last 3 years.

 

[Bouncing Zabaglione Bros.]

To say AMD's achievements are dumb luck is oversimplifying tings and not giving AMD credit where it is due. AMD have been executing brilliantly over the last 3 years.

Read the article - that's not what the author is saying. First he says that colleages and readers think that AMD prospered because of Intel's failure, and then the author goes on to refute that view:

AMD’s got years worth of ammunition already locked and loaded. It hasn’t even played the 65-nanometer manufacturing process card, one that Intel had to play just to get Core Microarchitecture out the door. I’m certain that AMD is truly ready for 65-nanometer and other mind-blowing things, and I’m just as certain that Intel’s claims of Core Microarchitecture’s technological lead over Opteron will prove baseless. AMD’s road map is guided by IT’s needs and the capabilities of enterprise applications, rather than Intel’s provocations.

 

[Tahir2]

I was in the middle of reading it but got called away.. just wanted to post something from the article rather than just a link and title.

Edit:

Intel shot its entire wad on Core Microarchitecture. From here, the only place Intel can go is bigger cache, more cores and faster clocks.

I agree with this sentiment and right now I am not keen on the infrastructure Intel has in place for the Core Architecture, specifically the motherboard chipsets. AMD's platform has excellent variety, stability and features. It really does feel like a step backwards from Intel.
However I do think Intel will in the short and middle term scale better than AMD on the clockspeed front which is important but nowhere near as much as it used to be at one time. However if you now look at the price of an AMD Athlon X2 3800 (£90 or less in the UK) Intel has not got a reply apart from the E6300 which still requires expensive motherboards. Quad-core is something AMD fans are relying on to win back the performance crown but this time Intel has a much more robust CPU architecture to reply with.

I understand the article is focused on the server and workstation markets but my experience is in the Retail side.

 

[Tahir2]

http://www.theinq.com/default.aspx?article=33906

The down side is that AMD is not going to have a new core in 2007, or at least have it until the waning days of 2007 if everything goes swimmingly. This means that Intel will pretty much own the 1S 2C market, where the meat of the sales lie. At 2S 4C, Intel will have a slimmer lead, but should still have a lead. At 4S, AMD all the way.

Another view placing "K8L" into Q4 '07 or Q4 '08 and also giving the performance advantage to AMD when dealing with Quad Core CPU architectures.

Perhaps it would be prudent of Intel to license and implement HyperTransport on its motherboards.

 

[SugarCoat]

Xbit did a preview (all text) of K8L a couple days ago.

http://www.xbitlabs.com/articles/cpu/display/amd-k8l.html

 

[Gubbi]

http://www.theinq.com/default.aspx?article=33906


Another view placing "K8L" into Q4 '07 or Q4 '08 and also giving the performance advantage to AMD when dealing with Quad Core CPU architectures.

But to be fair, that was the quad core K8L that just recently taped out. They could have had a dual core K8L in the works for some time now. It leaves a lot of room for speculation.

Cheers

 

[Blazkowicz]

when there was that K8L tape out, I remember reading that we'll have the K8L opteron on sale by mid 2007?

 

[ANova]

Intel still uses a front-side bus architecture in which memory, inter-processor and peripheral data compete. Core Microarchitecture now uses two such buses, but broken plus broken does not add up to breakthrough.

There is more to a chip than the memory controller and thus why we see the kind of performance we do from the Conroe/Woodcrest. Besides, no one said anything about Intel not integrating the controller in the future.

Study Intel’s rhetoric closely: Core Microarchitecture’s big numbers, like the “40 percent faster,” represent Plan B’s lead over Netburst Xeon in cycle-per-cycle performance and power efficiency. Intel has indeed left its prior best in the dust, but Intel’s presentation of this fact is almost indiscernibly blended with Intel’s far less impressive claimed lead over Opteron.

That depends entirely on the application, the Xeon's lead can be as much as 40% over the Opteron in some cases while others not so much; overall the lead is fairly significant.

Here’s the truth: The direct performance-per-watt numbers that Intel has published to bolster its claims of lower power usage pit a high-power Revision E Opteron against a low-power Core Microarchitecture Xeon. AMD ships 35 and 55 watt Opteron CPUs and always has. AMD’s PowerNow! run-time power management has been standard in Opteron for a long time; it is not Intel’s invention.

The newer Opterons may have power consumption equal to that of the Xeon but the fact remains that they are not at all equal in performance per watt.

Intel shot its entire wad on Core Microarchitecture. From here, the only place Intel can go is bigger cache, more cores and faster clocks. That sounds like a grand triple play, but it isn’t. Mark my words: Core Microarchitecture will not scale.

No, Intel can do more than that and they will. Core microarchitecture won't scale? Take a look at the performance and overclocking records being achieved on the first generation of the technology.

AMD’s got years worth of ammunition already locked and loaded. It hasn’t even played the 65-nanometer manufacturing process card, one that Intel had to play just to get Core Microarchitecture out the door. I’m certain that AMD is truly ready for 65-nanometer and other mind-blowing things, and I’m just as certain that Intel’s claims of Core Microarchitecture’s technological lead over Opteron will prove baseless.

I can't even begin to explain what's wrong with this statement. As far as the technological lead being baseless, all the evidence to the contrary is available as I type this.

If this article were any more biased I would have thought it came straight out of AMD's PR department. You know what's even funnier? That site is loaded with Intel ads.

 

[INKster]

The article totally fails to mention intel's superior unified L2 Cache vs AMD's separate L2's, intel's current 12-month lead on manufacturing processes, 2008's new "Nehalem" architecture (about the same time as the so-called "K8L" will be available on the desktop) and, of course, the sheer manufacturing volume differences.

Extremely biased, completely non-sense article.
What matters are the practical results, and right now, integrated memory controller or not, intel clearly has the lead on performance (even i, currently owning a single intel cpu system -a Pentium D small server- among a handfull or two of desktop and server pc's, promptly admit the fact, because i tend to buy what is best in the price/performance category, regardless of brand names).

 

[Zaphod]

first generation of the technology.

Is it really first generation, though? I'd think that the current Core 2 would have been in the pipeline since before Banias tapes out, and that happened some time during 2002.

 

[crypto1300]

Is it really first generation, though? I'd think that the current Core 2 would have been in the pipeline since before Banias tapes out, and that happened some time during 2002.

Well if you count the Core Solo's & Duo's as first generation wouldn't the Core 2's be the second generation?

They share very many characteristics and architectures don't they?

 

[SugarCoat]

Well if you count the Core Solo's & Duo's as first generation wouldn't the Core 2's be the second generation?

They share very many characteristics and architectures don't they?

yes to characteristics. no to architecture.

 

[crypto1300]

yes to characteristics. no to architecture.

By chance would anyone happen to know where to find a side-by-side comparison of the Core & Core 2's?

 

[Accord1999]

By chance would anyone happen to know where to find a side-by-side comparison of the Core & Core 2's?

This one's quite nice:

http://www.realworldtech.com/page.cfm?ArticleID=RWT030906143144&p=3

 

[karlotta]

Is it really first generation, though? I'd think that the current Core 2 would have been in the pipeline since before Banias tapes out, and that happened some time during 2002.

well its firstish, its still a tuly at heart. That would be a p3.

 

[Zaphod]

well its firstish, its still a tuly at heart. That would be a p3.

I think my point is that, while it's certainly a new architecture in terms of product lines in general (and particularly in the desktop space), it is not new in the sense that: we have this completely new core (no capital C) here, and in a couple of revisions we'll have ironed out the crinkles and then we'll be able to see what this thing can do (like from the original P4 to the C revision Northwood).

They've already more time than that to experiment, tweak and refine in the mobile space. While I certainly don't buy the argument from the original article that: "Core Microarchitecture will not scale", on the other hand, saying that it's "first generation" implies too much of the other extreme, IMO. I'd be surprised if Intel doesn't have a much better idea if how far they can go with Core than they had with Netburst in the first couple of years.

 

[Blazkowicz]

"Core architecture won't scale" : I took that statement in the meaning of scaling up a SMP system
K8L will scale very nicely to 16 cores and 32 cores ; you don't even try to imagine 16 "Core Xeon" cores fighting for the same FSB.


Inkster : that superior Intel unified cache doesn't seem that superior when Intel puts two dual core dies on the same chip ; cache is no more that much unified and the two dies talk other the FSB.

 

[ToxicTaZ]

AMD Quad Core (K8L) vs Intel Quad Core (Bloomfield)

All i can say is:

Q4 2007
AMD AM3 & K8L (65nm SOI)

vs

Q4 2007
Intel Bearlake-X/G+ & Intel Wolfdale 45nm Dual-Core 3MB 1333MHz FSB
Intel Bearlake-X/G+ & Intel Ridgefield 45nm Dual-Core 6MB 1333MHz FSB

Q1 2008
Intel Bearlake-X/G+ & Intel Bloomsfield 45nm Monolithic Quad-Core 6MB 1333MHz FSB

Q2 2008
Intel Bearlake-X/G+ & Intel Yorkfield 45nm Octal-Core 12MB 1333MHz FSB (multi-die)

Q3 2008
Intel Core 3 (Nehalem) Duo/Quattro/Octal

 

[Colourless]

Xenon has unified L2 Cache and everyone says it's the worst thing ever.... Core has unified Cache and people think it's great