Intel Stepping Up to Combat AMD's 4x4

Grooves writes "Intel has said that the company is stepping up the pace of its Core 2 architecture rollout to compete with AMD's 4x4. Two "quad-core" parts originally slated for release in the first half of 2007, Kentsfield for the desktop and Clovertown for servers, will make their debut as early as the end of this year. The Ars article warns that per-core bandwidth problems could end up giving a performance advantage to AMD's 4x4 approach."

Whichever... Competition is a good thing!

[ackthpt]

So I'm pricing a new mobo+CPU combo for a friend. I bought an AMD64 about 14 months ago for $350. Now I see I can't even get that model anymore unless I buy the parts separately as "replacements" A few steps up from what I run is now $150. It's a good thing.

Maybe in a couple years I'll consider a Conroe or AMD 4x4 type system if I need any heavy rendering done, but for now It's astounding the bang for buck we get.

I'm seeing a greater demand for smp...

[rivaldufus]

in consulting. One of our clients wants to have at least 4-way SMP on each new box. With virtualization becoming so popular, those additional cores are going to help.

I wonder if AMD is going to focus on 4+ cores to maximize its hypertransport bus - and focus less on 2 core and less systems.

Re:Oooh.... core wars

[DeathKoil]

I'm not entirely sure that throwing more cores at the problem is neccessarily a great solution for combatting Intel's Core 2 Duo chips. Wouldn't this make AMD systems cost more money (2 FX chips on one motherboard) than an Intel gaming box would?

Don't get me wrong, I am a huge AMD fan but I'm not conviced that, "let's just put more cores into the box" is a great response to Intel's Core 2 Duo. The announcement of a new core from AMD would have been more exciting. I guess I'll have to wait for that.

Latencies and more

[cnettel]

The fact that the two dies (with two cores on each) will communicate over the FSB is of course limiting, but we also have to remember that each of those dies will have 4 MB of L2, 8 MB in total. We've already seen what the Core 2 prefetching can do in hiding the memory controller latency, so if things are good it will work equally well in prefetching data from the L2 on the other die. Then, the memory bandwidth is irrelevant, while the FSB bandwidth is still relevant. I seem to remember reading that either Kentsfield or Clovertown would carry some kind of dual-bus solution (with support in chipset), but maybe that was further ahead.

Let's also not forget that the NUMA properties of the AMD solution, with less advanced prefetching, can actually be a more significant latency problem in latency-sensitive applications. The bandwidth, on the other hand, will absolutely be there.

Is itanium officially dead?

[acomj]

Now that intel is finally throwing research and marketing on 64 bit x86 to compete with AMD, is its intel's other 64 bit chip itanium officially dead?

Re:Who is paying?

[Massacrifice]

You know, not only gamers use high end computers. Some of us use them as tools to actually work and earn money. in which case, we (or our employer) accumulates enough money to buy such things as quad-cores CPU, and eventually make them profitable, that is, use them to earn more money than they cost!

What's even better is that these machines, once work is over, can still be used to play games!

Fantastic isn't it? Work - maybe you should try it sometime.

Re:Oooh.... core wars

[October_30th]

With the advent of cheap dual (soon quad) core setups these days, developers will be holding back progress if they don't adapt to the new reality. Parallel algorithms are well researched. It's just a matter of taking what's available and building from there.

Re:Who is paying?

[Wesley Felter]

People who use computers for work and who want four cores already bought Opteron workstations. That's why 4x4 and Kentsfield are targeted at the enthusiast (aka more money than sense) market.

Re:Gotta love CPU wars

[Roody Blashes]

But what if consumers don't need better processors? There's a huge disparity between this new "war" and the early P4 and late P3 days. Back then, it was a foot race to provide the newest, most exciting processors around. Now? Meh.

Here's the rub: everytime a new breed of chips comes out, I find myself just buying the increasingly cheap leftover stocks of prior generations. The last big thing that got me up off my butt to buy a CPU was 64-bit on the x86 platform, and the only reason I even did that was because my LAST chip, the innovation of which was an 800MHz bus, got a nice little roasting from a shorted motherboard.

When it comes down to it, processing power and innovation, outside a few specialized sectors, isn't really doing it for me anymore. I find myself focusing very little on processors when shopping for parts because, frankly, I literally just need "A" processor and I can practically grab just about anything I want as long as it's compatible with my chipset.

Intel and AMD need to branch off into something more interesting. As a consumer, I don't need all this stuff. Give me something truly innovative, don't just drag concepts over from other platforms and hype them up in a press release. This is all well and good, but it's just not exciting and, by extension, it's not creating anything exciting as far as applications and use go. There is niether a NEED for the latest and greatest (short of neuroses some of us suffer) nor is there any exciting drive fueled by innovative new concepts.

Bah. Wake me when the first consumer-level quantum processing unit is getting ready to be released, I guess.

Re:And so it begins

[Mr. Flibble]

Do most chip sales happen at the release date, or do most people wait for the competitors product to come out spurring price drops to compete? I know I seldom buy anything at the alpha-expensive stage, usually preferring to wait a few months for the inevitable price drop.

It is not a question of inital chipsales, it is more a question of marketing. Back when both companies were trying to hit the 1 Ghz mark, AMD got there first. That was a big win for them, as consumers could now say 1000 Mhz! WOW! Even though intel quickly came out with faster chips thereafter. It was a win for AMD because the name AMD got into the minds of customers. The same thing with the 64 bit. Now, most people here on slashdot know what a 64 bit chip is, and does, and does not do. But the public does not. And since AMD had the 64 bit chip out first, consumers wanted it, even if it had no real benifit for them initally.

The same goes with this technology. Whomever gets it out of the gate first wins the "mindshare war" as we call it now. IIRC, the book "Predatory Marketing" covers how this works in detail - but they don't use the "mindshare" term in it.

Re:Oooh.... core wars

[fimbulvetr]

Which came first, the SMP or the SMP apps?

Personally, I think this is the best way to go about solving the chicken and the egg problem. Just doing it. Just start releasing the cores. I have absolutely no doubt that many, many applications will catch the drift and hop on board. It will take some time, indeed, but so did other software with hardware advancements (MMX, SSE, Graphics solutions, etc). Historically, the hardware has become before the software.

Re:4x4 is an inaccurate name...

[kupan787]

Here is what I don't get...why is Ars comparing a single, quad core Kentsfield to two, dual core Athlons? Wouldn't a better comparison to the 4x4 platform be two Woodcrest cups? Or am I missing something?

From the article:
AMD's 4x4 system features two coherent HyperTransport (cHT) AM2 sockets, each of which can gluelessly support a dual-core Athlon CPU and a pool of DDR2. So a 4x4 system gives you two cores and one DDR2 bank per socket.

Sounds an awful lot like what Intel is doing with Woodcrest (and its platform, of which I can't recall the name). 2 sockets, each with a DIB, and can hold a dual core part. So how is 4x4 new? Hell, Apple is doing it currently with their G5 [apple.com]

Name just one really good home use for all this...

[Anonymous Coward]

Name just one really killer, gosh I just can't live without it, my life depends on this etc etc application at *HOME* that an average person whose hobby is *not* computers actually needs this kind of computing power for RIGHT NOW!

I did a head count - it's ZERO! Worse still - that's the core market right there - ZERO!

Sure, we can all see a day when our computers are able to intelligently discuss life issues at length with a voice interface (Hello, Dave) but we're not there yet. And there are the enthusiasts who are always willing to pay for a little more 'go' in their machine to get one extra fps in the latest, errr, FPS. Then there are server configurations that are actually more bogged down by storage transfer rates than CPU usage.

There I said it, the Emperor is naked - let's all have a good laugh at him and think about something more important.

Re:Oooh.... core wars

[Andy Dodd]

It isn't. The original linked article pointed out that while Intel's quad-core approach may suffer performance penalties compared to 4x4, BUT will likely provide more "bang for the buck".

AMD (sadly) seems to have forgotten that x86 SMP was around for at least a decade before the Athlon 64 X2, and due to cost issues, it was always a niche technology.

Dual-core-in-a-single package chips have managed to change that in the span of 2-3 years... SMP has gone from a a niche technology installed in probably less than 1% of computers sold to something present (in the form of dual-core CPUs) in what is likely 75% or more of new machines in a VERY short time. Simply put, multiprocessor in multiple sockets does NOT sell except to the extreme high end. Keep in mind how well dual-core has done despite the fact that it has clear performance penalties in most situations compared to two seperate CPUs.

In short, 4x4 isn't really going to get AMD anywhere in my opinion. Unlike dual core technology, it'll stay as a small market share niche item just like classic SMP systems did.

Hopefully for AMD, they can remain profitable even after the massive price cuts they're going to have to do (and apparently will be doing shortly) in order to remain competitive.

Re:And so it begins

[jejones]

The Core 2 Duos aren't "alpha expensive" -- they're significantly cheaper than AMD's prices as of right now.

Agreed, but... July 24th, the date when AMD is going to cut some CPU prices almost in half, is barely over a weekend away, and there is the question of supply and demand. Will demand be sufficient to drive the price up?