Sign in with
Sign up | Sign in
Your question

Intel's Clovertown revealed

Last response: in CPUs
Share
April 14, 2006 4:41:26 PM

Recently, IDF in Taiwan revealed Intel's first generation quad cores, Clovertown. It's news and benchmark can be found here

http://translate.google.com/translate?u=http%3A%2F%2Fcp...

The engineering sample of Clovertown scored 362 on Cinbench 9.5 for single operation @ 2GHz. A 2GHz Opteron 246 got the exact same score for the same test.

http://www.2cpu.com/review.php?id=110&page=6

Doesn't Clovertown use Intel's NGMA? How come its test score is exactly the same as a 3 year old single core Opteron?

On a side note, if you take a look at Kentsfield's architecture, it is consisted of two Conroe core, glued together with a single FSB, which is similar to P-D's approach. However, hasn't that approach been determined inferior to AMD's crossbar design? If Intel knows it is inferior than AMD's design, why do they keep on using it for their future generation processors?
April 14, 2006 4:55:36 PM

Lols, 65nm Quad Core, and its practically nexxt Generation Stuff, And a "Classic" Opteron Core can Easly Beat it still hold a finger up^
Loll and its a "Classic DDR" mostlikely

:D  :)  :D  Intel cracks me up, Bleeding Edge next generation can only just barely compare to "classic" AMD systems.
a c 99 à CPUs
April 14, 2006 5:03:33 PM

The only thing going through my mind when I saw the picture of the Clovertown chip diagram was, "That MCM is going to make one HUGE die!" I bet that will be over 300mm^2, easily.
Related resources
April 14, 2006 5:11:28 PM

lolz you bet

conroe already has a large die area because of its unified L2 cache, and now intel is going to put two of them together to become a new Kentsfield... XD

what a brilliant idea
April 14, 2006 5:22:16 PM

quad core hmm

those things must be massive


and i thought my AMD 64 939 was massive with only 512 of cache =/
April 14, 2006 5:41:44 PM

Great american engineering - "This one ain't fast enough, let's bolt a few of 'em together and see what happens"

That screen of task manager with 8 cpu readings was totally stupid.
April 14, 2006 5:45:01 PM

LOL, Cinebech was probably updated a lot since then. And the good thing about it is that it's a two way system quad core . Single core performance simply doesn't need to be high. Because the add of another core(2 in fact) will simply turn the ballance. I don't expect to see the 2.4 ghz barrier passed by quad core CPUs untill 2008. The gains of having such a beast should be considerable. Consider the advantage of having dual over single core. It'll be the same case here. And of course they'll be expensive as hell.
April 14, 2006 5:47:20 PM

I guess that single-thread speed is pretty OK for an early sample of the CPU.
April 14, 2006 6:13:51 PM

2P 8 thread's score is only 4.67 times 2P 1 thread score. i think that said something about intel's FSB approach and Kentsfield's effiency.

we'll see how it really performs when compares to amd's own quad core, featuring on-die memory controller, crossbar, and HTT.

intel is going to lose the server market, again.
April 14, 2006 8:39:26 PM

Well, at this early stage isn't it a bit early for the above to be even slightly conclusive? The end result could be ten times faster... or ten times slower.
April 14, 2006 9:19:03 PM

Quote:
Well, at this early stage isn't it a bit early for the above to be even slightly conclusive? The end result could be ten times faster... or ten times slower.


Ten times faster? Bwhahaha, I will give Intel credit where it's due, but I won't give ANY company 10x better than an ES.

~~Mad Mod Mike, pimpin' the world 1 rig at a time
April 14, 2006 9:23:28 PM

Quote:
Great american engineering - "This one ain't fast enough, let's bolt a few of 'em together and see what happens"

That screen of task manager with 8 cpu readings was totally stupid.


lol

that page doesnt load on my fireofx sigh
April 14, 2006 9:26:08 PM

Considering that Merom, using half the FSB speed and also not using DIB, outperforms AMDs premium desktop offerings, I dont think Opterons will be much of a challenge until the next AMD architecture arrives.
In fact, AMD will need at least 2 years to catch up to Woodcrests "current" performance, which they also confirmed.
Too bad that Clovertown will stay in the processor arena only for a short time, with CSI and the corresponding platform upgrade coming in 2008.
April 14, 2006 9:52:51 PM

Intel isn't American... Apple is (and they suck balls)
April 14, 2006 10:05:26 PM

Quote:
Considering that Merom, using half the FSB speed and also not using DIB, outperforms AMDs premium desktop offerings, I dont think Opterons will be much of a challenge until the next AMD architecture arrives.
In fact, AMD will need at least 2 years to catch up to Woodcrests "current" performance, which they also confirmed.
Too bad that Clovertown will stay in the processor arena only for a short time, with CSI and the corresponding platform upgrade coming in 2008.

*cought*fanboy!*cought*
April 14, 2006 10:07:59 PM

Quote:
Intel isn't American... Apple is (and they suck balls)


words from someone who doesnt understand the tech behind apple *.* OSX *.*

MadModMike along with ALOT of other posters know the joys of a perfect OS sorry but vista in my oppinion will not be as great :( 
April 14, 2006 10:20:42 PM

Quote:
Intel isn't American... Apple is (and they suck balls)


words from someone who doesnt understand the tech behind apple *.* OSX *.*

MadModMike along with ALOT of other posters know the joys of a perfect OS sorry but vista in my oppinion will not be as great :( 

Word.
April 14, 2006 11:04:07 PM

Quote:
Intel isn't American...


Huh? They may not be exclusively based in America, but they are certainly an American company.

http://en.wikipedia.org/wiki/Intel

Quote:
Intel Corporation (NASDAQ: INTC, SEHK: 4335), founded in 1968 as Integrated Electronics Corporation, is a U.S.-based multinational corporation that is best known for designing and manufacturing microprocessors and specialized integrated circuits.
April 14, 2006 11:07:07 PM

Quote:
Intel isn't American...


Huh? They may not be exclusively based in America, but they are certainly an American company.

http://en.wikipedia.org/wiki/Intel

Quote:
Intel Corporation (NASDAQ: INTC, SEHK: 4335), founded in 1968 as Integrated Electronics Corporation, is a U.S.-based multinational corporation that is best known for designing and manufacturing microprocessors and specialized integrated circuits.


Word.
April 14, 2006 11:26:31 PM

It was hypothetical. You know, the organic rocket fuel made from insect feces?
April 14, 2006 11:59:04 PM

Quote:
Recently, IDF in Taiwan revealed Intel's first generation quad cores, Clovertown. It's news and benchmark can be found here

http://translate.google.com/translate?u=http%3A%2F%2Fcp...

The engineering sample of Clovertown scored 362 on Cinbench 9.5 for single operation @ 2GHz. A 2GHz Opteron 246 got the exact same score for the same test.

http://www.2cpu.com/review.php?id=110&page=6

Doesn't Clovertown use Intel's NGMA? How come its test score is exactly the same as a 3 year old single core Opteron?

On a side note, if you take a look at Kentsfield's architecture, it is consisted of two Conroe core, glued together with a single FSB, which is similar to P-D's approach. However, hasn't that approach been determined inferior to AMD's crossbar design? If Intel knows it is inferior than AMD's design, why do they keep on using it for their future generation processors?


If these marks hold up - I wonder what Intel is thinking. I thought they had learned their lesson with the P4's and P-D's. It's like them saying 4 or 8 is better than one. Well in most things in the world 4 or 8 is better than one. Seeing the Dual Core Opteron beat the quad core Intel makes me wonder if Intel will ever turn things around. From what I am seeing is alot of press but not many benchmarks are in favor for Intel (Intel new vs. AMD Old).
:wink:
April 15, 2006 12:31:10 AM

Quote:
Considering that Merom, using half the FSB speed and also not using DIB, outperforms AMDs premium desktop offerings, I dont think Opterons will be much of a challenge until the next AMD architecture arrives.
In fact, AMD will need at least 2 years to catch up to Woodcrests "current" performance, which they also confirmed.
Too bad that Clovertown will stay in the processor arena only for a short time, with CSI and the corresponding platform upgrade coming in 2008.


YCON Are you getting going to become a Senile old man Or woman or anything in between? Woodcrest Line isn't out and about now. "WTF" do you mean "current" Until your sitting with a "bad Axe" mobo with a Conroe 2.66Ghz Processer chip in the wash room Jerknig off feeling and smelling it, It's not a "current" performance. Only "current" performance can come out when you "the Customer" can buy it freely not see 4 megabyte Bits being bench marked. To you>> "current" is like smelling my Fart and telling how big a Turd i'm going to drop. :evil: 
April 15, 2006 12:32:18 AM

Quote:
Considering that Merom, using half the FSB speed and also not using DIB, outperforms AMDs premium desktop offerings, I dont think Opterons will be much of a challenge until the next AMD architecture arrives.
In fact, AMD will need at least 2 years to catch up to Woodcrests "current" performance, which they also confirmed.
Too bad that Clovertown will stay in the processor arena only for a short time, with CSI and the corresponding platform upgrade coming in 2008.


YCON Are you getting going to become a Senile old man Or woman or anything in between? Woodcrest Line isn't out and about now. "WTF" do you mean "current" Until your sitting with a "bad Axe" mobo with a Conroe 2.66Ghz Processer chip in the wash room Jerknig off feeling and smelling it, It's not a "current" performance. Only "current" performance can come out when you "the Customer" can buy it freely not see 4 megabyte Bits being bench marked. "current" is like smelling my Fart and telling how big a Turd i'm going to drop. :evil: 

*raises left eye brow*
April 15, 2006 12:35:32 AM

Quote:
Considering that Merom, using half the FSB speed and also not using DIB, outperforms AMDs premium desktop offerings, I dont think Opterons will be much of a challenge until the next AMD architecture arrives.
In fact, AMD will need at least 2 years to catch up to Woodcrests "current" performance, which they also confirmed.
Too bad that Clovertown will stay in the processor arena only for a short time, with CSI and the corresponding platform upgrade coming in 2008.


YCON Are you getting going to become a Senile old man Or woman or anything in between? Woodcrest Line isn't out and about now. "WTF" do you mean "current" Until your sitting with a "bad Axe" mobo with a Conroe 2.66Ghz Processer chip in the wash room Jerknig off feeling and smelling it, It's not a "current" performance. Only "current" performance can come out when you "the Customer" can buy it freely not see 4 megabyte Bits being bench marked. "current" is like smelling my Fart and telling how big a Turd i'm going to drop. :evil:  8O
April 15, 2006 12:53:26 AM

Quote:
Considering that Merom, using half the FSB speed and also not using DIB, outperforms AMDs premium desktop offerings, I dont think Opterons will be much of a challenge until the next AMD architecture arrives.
In fact, AMD will need at least 2 years to catch up to Woodcrests "current" performance, which they also confirmed.
Too bad that Clovertown will stay in the processor arena only for a short time, with CSI and the corresponding platform upgrade coming in 2008.

well it is the same as saying "Too bad that Conroe will stay in the processor arena only for a short time", since Clovertown is the "double conroe glued together and slap a central FSB on it" version.

but.. i bet most intel fanboyz here disagree with that huh... :twisted:
April 15, 2006 1:01:07 AM

and don't forget, Opteron's score was ran in Dec. 2003, approx 2.5 years ago, and it still beats a quad core which is due in 8 months..

this is like... 2~3 generations of differences and AMD still has the upperhand
April 15, 2006 1:13:44 AM

Quote:
and don't forget, Opteron's score was ran in Dec. 2003, approx 2.5 years ago, and it still beats a quad core which is due in 8 months..

this is like... 2~3 generations of differences and AMD still has the upperhand
:lol: 
April 15, 2006 1:33:10 AM

Intel is nothing since they ditched the shiny spacesuits.
April 15, 2006 1:49:40 AM

Quote:
Intel is nothing since they ditched the shiny spacesuits.
Oh god not Foamy...
April 15, 2006 2:23:20 AM

Do you really have to be so ridiculus. I know people don't trust the Intel benchmarks, which is fine, and that's all based on conjecture and speculation. What you have done is compared Intel results run on Cinebench 9.5 to those Opteron results run on Cinebench 8 (2003), which is even worse than the reliability of Intel benchmarks. Even if the software was the same, we have no idea about the hardware configurations mainly in the case of the Intel system. Afterall, the Conroe test system only used DDR2 667 ram running at 4-4-4. Between the software incompatibilities and the hardware setups this really proves nothing one way or another.

Other considerations include the fact that this is still an early engineering sample of Cloverton who isn't targeted for release until Christmas at the earliest. Giving engineering samples the benefit of the doubt is what people are doing for the Rev F AM2 parts. As well, even though Cinebench is multithreaded, I really have no idea how well it takes advantage of 8 cores. In terms of processor scaling, 4.67 times single-core is good but not particularly impressive. It needs to be noted that that is mainly due to the 1066MHz FSB being used. Intel is supposed to be tweaking Cloverton and its northbridges to allow for a 1333MHz FSB which would be greatly appreciated. Again, given the Christmas - early 2007 launch window they still have plenty of time to make it happen so we'll have to see.

Edit: Ah Jack beat me to the Cinebench issue.
April 15, 2006 2:29:18 AM

Quote:
Recently, IDF in Taiwan revealed Intel's first generation quad cores, Clovertown. It's news and benchmark can be found here

http://translate.google.com/translate?u=http%3A%2F%2Fcp...

The engineering sample of Clovertown scored 362 on Cinbench 9.5 for single operation @ 2GHz. A 2GHz Opteron 246 got the exact same score for the same test.

http://www.2cpu.com/review.php?id=110&page=6

That score is for Cinebench 2003 and isn't directly comparable with scores in 9.5 being lower across the board. Clovertown's score of 362 in 64-bit mode would place it near a 2.2GHz Opteron. Cinebench doesn't scale all that well with core count either so a 4.7x speedup with 8 cores should be similar to what will be seen on Opteron.

There's a ranking page of various computers for Cinebench 9.5 at:

http://www.3dfluff.com/mash/cb95/top.php

But it's down right now, here's the Google cache of it:

http://72.14.203.104/search?q=cache:zBe8ggg7iCMJ:www.3d...

You can see that Cinebench scales ~85% between 1->2, but only about 60% between 2->4 cores.

As well, Cinebench is probably not yet optimized for Conroe's monstrous SIMD capabilities and could see a significant increase in performance, like that seen for the PPC970 when Cinebench added optimizations for it.
April 15, 2006 3:19:39 AM

Quote:
Do you really have to be so ridiculus. I know people don't trust the Intel benchmarks, which is fine, and that's all based on conjecture and speculation. What you have done is compared Intel results run on Cinebench 9.5 to those Opteron results run on Cinebench 8 (2003), which is even worse than the reliability of Intel benchmarks. Even if the software was the same, we have no idea about the hardware configurations mainly in the case of the Intel system. Afterall, the Conroe test system only used DDR2 667 ram running at 4-4-4. Between the software incompatibilities and the hardware setups this really proves nothing one way or another.

Other considerations include the fact that this is still an early engineering sample of Cloverton who isn't targeted for release until Christmas at the earliest. Giving engineering samples the benefit of the doubt is what people are doing for the Rev F AM2 parts. As well, even though Cinebench is multithreaded, I really have no idea how well it takes advantage of 8 cores. In terms of processor scaling, 4.67 times single-core is good but not particularly impressive. It needs to be noted that that is mainly due to the 1066MHz FSB being used. Intel is supposed to be tweaking Cloverton and its northbridges to allow for a 1333MHz FSB which would be greatly appreciated. Again, given the Christmas - early 2007 launch window they still have plenty of time to make it happen so we'll have to see.

Edit: Ah Jack beat me to the Cinebench issue.


:lol:  :lol:  I expected no less from AMD's biggest fan, Mr. Shakirou or whatever the heck his name is..... this guy claims to hold a PhD, in what??? From what school?? I need to know where not to send my kids :) 

Word.
April 15, 2006 4:31:46 AM

Quote:
Do you really have to be so ridiculus. I know people don't trust the Intel benchmarks, which is fine, and that's all based on conjecture and speculation. What you have done is compared Intel results run on Cinebench 9.5 to those Opteron results run on Cinebench 8 (2003), which is even worse than the reliability of Intel benchmarks. Even if the software was the same, we have no idea about the hardware configurations mainly in the case of the Intel system. Afterall, the Conroe test system only used DDR2 667 ram running at 4-4-4. Between the software incompatibilities and the hardware setups this really proves nothing one way or another.

Other considerations include the fact that this is still an early engineering sample of Cloverton who isn't targeted for release until Christmas at the earliest. Giving engineering samples the benefit of the doubt is what people are doing for the Rev F AM2 parts. As well, even though Cinebench is multithreaded, I really have no idea how well it takes advantage of 8 cores. In terms of processor scaling, 4.67 times single-core is good but not particularly impressive. It needs to be noted that that is mainly due to the 1066MHz FSB being used. Intel is supposed to be tweaking Cloverton and its northbridges to allow for a 1333MHz FSB which would be greatly appreciated. Again, given the Christmas - early 2007 launch window they still have plenty of time to make it happen so we'll have to see.

Edit: Ah Jack beat me to the Cinebench issue.


:lol:  :lol:  I expected no less from AMD's biggest fan, Mr. Shakirou or whatever the heck his name is..... this guy claims to hold a PhD, in what??? From what school?? I need to know where not to send my kids :) 

Word.
Honored.


word
obor
robo
dorw
!