“Just as the bragging rights for dual-core chip supremacy are dying down, Intel gave the first glimpse of a quad-core chip coming next year,” Michael Kanellos reports for CNET News. “Clovertown, a four-core processor, will start shipping to computer manufacturers late this year and hit the market in early 2007. Clovertown will be made for dual-processor servers, which means that these servers will essentially be eight-processor servers (two processors x four cores each).”
“Core expansion will be a dominant theme for Intel over the next few years, said Chief Technology Officer Justin Rattner. By the end of the decade, chips with tens of cores will be possible, while in 10 years, it’s theoretically possible that chips with hundreds of cores will come out, he added,” Kanellos reports. “Rattner showed off a computer running two Clovertown processors… Clovertown and Tigerton are members of a new chip architecture coming from Intel at the end of the year. A notebook chip called Merom and a desktop chip called Conroe coming out around the same time will be based on the same architecture. Intel will give the architecture a name at the Intel Developer Forum taking place in March. Rattner indicated that Merom and Conroe will only be dual-core chips, as many analysts expect.”
“At around the same time, after all, Intel will release Woodcrest, a dual core server chip based around the same Merom-Conroe-Tigerton-Clovertown architecture. It will contain only two cores and consume 80 watts of power, less than the 165-watt server chips Intel sells now,” Kanellos reports. “A large financial institution is currently running servers on an experimental basis with Woodcrest chips, Rattner said.”
Full article here.
[Thanks to MacDailyNews reader “Andy C.” for the heads up.]
Advertisements:
• MacBook Pro. The first Mac notebook built upon Intel Core Duo with iLife ’06, Front Row and built-in iSight. Starting at $1999. Free shipping.
• iMac. Twice as amazing — Intel Core Duo, iLife ’06, Front Row media experience, Apple Remote, built-in iSight. Starting at $1299. Free shipping.
• iMac and MacBook Pro owners: Apple USB Modem. Easily connect to the Internet using dial-up service. Only $49.
• iPod Radio Remote. Listen to FM radio on your iPod and control everything with a convenient wired remote. Just $49.
• iPod. 15,000 songs. 25,000 photos. 150 hours of video. The new iPod. 30GB and 60GB models start at just $299. Free shipping.
• Connect iPod to your television set with the iPod AV Cable. Just $19.
Related MacDailyNews article:
New IBM Power chip breaks speed barriers, promises to be twice as fast as Sun, Intel, AMD chips – February 08, 2006
Pretty cool. The next few years aare going to be fun.
Haha…
“Tiger-ton…”
Yeah ‘R’! And us Mac-heads can get to hang around with the cool kids for a change! – Gosh!
All those that still think Apple made a mistake switching to Intel, sound out below. IBM and AMD aren’t the only chip vendors who can innovate…
What are you talking about? I AM one of the cool kids.
” width=”19″ height=”19″ alt=”raspberry” style=”border:0;” />
Assuming Apple builds these quad core ships into the XServe, if Virginia Tech upgrades their 1100 XServes to new Intel based ones at the beginning of 2007, they’ll be able to have 8800 cores in the same amount of space 2200 take up now.
If Apple puts 4 quad core Tigerton processors in the XServe, Virginia Tech could build a 17600 core machine in the same amount of space.
Is it just me, or is Moore’s law falling behind?
Next up: Apple blade servers. If they are serious about the server market, they’ll release XServe blades in 2007. For those who don’t know what a blade server is: http://en.wikipedia.org/wiki/Blade_server
Anyone look at the http://www.top500.org supercomputer ranking recently? Apple has 4 systems in the top 500 now, up from 0 a couple years ago. Although they no longer hold any of the top 10 spots, they have 2 in the top 20. It’s nice to see, but they are going to have to push a little harder in the server space if they want to stay relevant. The current XServes are looking a little dated…
Tom Ferguson–
What are you talking about? We are the cool kids. We’re just making the other kids a little bit cooler by allowing them to hang out with us.
” width=”19″ height=”19″ alt=”wink” style=”border:0;” />
By the way, I hope this quiets down some of the people that are crying about IBM’s Power 6 announcement.
Are any of these chips 64-bit?
http://musobs.blogspot.com/
Next–Leopardton.
Does anyone have the foggiest idea how low the yield will be and the increased costs and breakdown potential of these new quad cores will be?
I’ll wait until a dual – quad core processor Mactel Tower Pro comes out before I give up my PPC machine.
Why just Xserves? Sounds like these chips would be ideal for the PowerMac replacement. 8 processors in one machine would go a long way to revive the professional line for Apple.
“Rattner showed off a computer running two Clovertown processors.
Multiplying the number of cores brings distinct advantages. First, it cuts down overall energy consumption for equivalent levels of performance. If the recent Core Duo chips released for notebooks from Intel had only one core, the chips would consume far more power, he said.”
Sounds counter-intuitive, but if true maybe the iBooks will also use core-duo chips. Doesn’t make sense to put a single core chip in a laptop if it is going to consume more power.
I would read that as Duo Core is more power efficient than having two single core chips.
Who cooks up these processor names at Intel?
To Tom Ferguson: You’re mistaken. Mac users have ALWAYS been the coolest people around. If you consider Alienware DOS-Boxes cool… well…
ROFLMFAO! TWICE.
It certainly WILL be interesting to see what Apple cooks up to house these new processors.
A laptop or PDA iPod with 200 CPU’s? Cool!
” width=”19″ height=”19″ alt=”grin” style=”border:0;” />
Apple is ahead of the game. They’ve already designed a computer designed to handle “hundreds of cores,” it’s called the G5 tower.
MW – addition
I want to know who comes up with Intel’s chip code names. Whoever it is should be flogged. Clovertown??
Hmmm.
I guess they skipped right over Jaguar-ton.
Voyager
The Clover thing ties in nicely with it having four “leaves”.
To the tune of “Funkytown”:
“…Won’t you take me to….Clovertown….
….Won’t you take me to….Clovertown….
….Aaaaaahhh, talk about it, talk about it, talk about moo-ooo-vin…”
MDN magic word, “Dumb”
As in “Dumb and Dumber”
I recently watched the Megahertz Myth video demonstration (yes, again), and I have a question.
For those who haven’t seen it, you can see it here:
http://www.asia.apple.com/g4/myth/
Now, I am not a chip architecture designer, but I am curious about one thing in this video.
First, Jobs states that for the test the G4 outperformed the Pentium by 80% for the same task. I have no reason to doubt this.
In the demo, John highly emphasizes the length of the pipeline. I come away from watching it thinking shorter pipeline = better performance. What the video doesn’t show is that for every ONE processor cycle or ‘tick’ the G4 takes, the Pentium should take TWO (given that the Pentium has twice the clock speed). The demo basically shows both chips running at the same clock speed – tick for tick. Should it not be ‘tick’ (G4) for ‘tick-tick’ (Pentium)? I am curious as to what others think about this or if anyone can clarify it.
Since higher clock speeds = more heat dissipation. I assume that is why the trend in the industry is now to go with multi-core chips with slightly lower clock speeds. But, doesn’t a program ITSELF have to specifically take advantage of a dual core design, or does the chip take care of this on it’s own? The reason I ask is because there have been issues with Pentium D processors which are also dual-core.
I would be grateful for any insight into this. Thanks.
Not a chip designer either, but a simple tick/tick graph like John showed can’t explain the whole, complex process.
In addtion to tick speeds down a pipeline there’s also, I believe, a difference as the G4 ran using RISC (Reduced Instruction Set Code) versus CISC (Complex Instruction Set Code) in the Windows world. RISC runs faster as it doesn’t have to call other code as often as CISC does. It’s something like a staightforward code running down 7 pipelines is faster than a complex code running down 20 pipelines at a faster speed. And you’re looking at a video explaining the difference of chips, two generations past. We’re now in the world of multiple processing cores on a single chip which introduces a whole new way of assessing performance.
As far as “doesn’t a program ITSELF have to specifically take advantage of a dual core design,” I believe OSX (this isn’t Windows after all) takes care of this, unlike the situation where a program specifically has to be written to take advantage of a 64 bit processor. Hence, The Other Steve’s comment above that Apple has “already designed a computer designed to handle “hundreds of cores,” it’s called the G5 tower.” OS X juggles the cores, not the program.
If I’ve said anything misleading, I’m sure I’ll be corrected.
As far as “doesn’t a program ITSELF have to specifically take advantage of a dual core design,” I believe OSX (this isn’t Windows after all) takes care of this,
OSX is no different than Windows in this manner–both have a built in scheduler that assigns threads across multiple processors (whether dual core or dual processor). A program just has to be multithreaded (not physically determine if a dual processors is there), the scheduler (whether OSX or Windows) takes care of the rest.