Native Quad Core?

[Donnie27]

http://www.theinquirer.net/default.aspx?article=35442

So I was having the *talk with the AMD guys and they argued over Inte's Two Duals Vs. AMD 4 Single Cores Processor being truly Native. Now without starting a Flame War, I'd love to hear how or what Native or what's Native when Quad = 4 and 4 Cores = Quad Period?

My idea of Native Quad Core would be 4 Execution Cores with one L1 and L2. It would feature a BUS wide/fast enough so none of its units would starve for Bandwidth. Something like 2000MHz X 256Bit or 1000MHz 512bit.

It is pretty silly for Mr. Ruiz to be running around talking about native Quad Core while calling a two Socket system "Quad Father", how FOS can you get?
Crap and after all this time I thought Jerry Sanders was a "Blowhard".

 

[Poncho]

How many cores are in a single CPU? If there are 4... then guess what, it's a quad core CPU. Why the fuck should I or anybody else give a shit if it's 2 dies on a package or 1? If the thing performs well.... that's all I care about. This native/non-native BS is just that... bullshit. Are there better ways to do quad core than 2 dies? Maybe, hell probably, but if that way kills the competition then whats the problem? People need to realize that while Intels first solution to these things (first dual core, quad core, etc) may not be the best way to do them.... they are still doing them first. So while Intel's first quad core may not be the "perfect" solution (as if there ever was one)... they will still have quad core on the market LONG before AMD will, not to mention dual core and 65nm.

 

[Lolento]

Hector Ruiz is pitching is sell....just like the guys at Best Buy

so take it like a grain of salt

 

[Donnie27]

How many cores are in a single CPU? If there are 4... then guess what, it's a quad core CPU. Why the fuck should I or anybody else give a shit if it's 2 dies on a package or 1? If the thing performs well.... that's all I care about. This native/non-native BS is just that... bullshit. Are there better ways to do quad core than 2 dies? Maybe, hell probably, but if that way kills the competition then whats the problem? People need to realize that while Intels first solution to these things (first dual core, quad core, etc) may not be the best way to do them.... they are still doing them first. So while Intel's first quad core may not be the "perfect" solution (as if there ever was one)... they will still have quad core on the market LONG before AMD will, not to mention dual core and 65nm.

To me it is much simpler than that. 4 Cores on one package equals Quad core =P Doesn't matter if that's 4 AMD cores or 4 Intel Cores on one package. Doesn't matter if they're glued by a FSB or a Crossbar.

The problem, minor or not, is to have the Head Guy at AMD trying the BS the public like this. Never mind the Fan$ and etc...... What else would he try to mislead the public on if he lies about something this simple?

 

[PaHick]

IMO, quadcore means all on one socket. If native quadcore doesnt mean one socket, then its been here long ago.

 

[Donnie27]

IMO, quadcore means all on one socket. If native quadcore doesnt mean one socket, then its been here long ago.

Sure it's one socket. Trying to say two Dual Cores is not as Native as 4 single cores (K8L) is a Joke though

 

[PaHick]

Sure it's one socket. Trying to say two Dual Cores is not as Native as 4 single cores (K8L) is a Joke though

With computers getting smaller all the time, I cant see why any manufacturer would setup 2 sockets for quadcore.

 

[Poncho]

The quad-core you've heard about is just two dual-core chips glued together, it's not really quad-core

I like this quote. And it's quite accurate. A couple years ago they went through training at all the fabs where they taught the people how to glue CPUs together properly. It took some time... and A LOT of super glue... but I think the result was well worth it. I've actually seen reports on the money saved if Intel would get into the glue making business. That's the real outcome of the efficiency action taking place within intel right now and the costs that would be saved are insane.

 

[AACDIRECT]

Native quad core means a single piece of silicon containing 4 cores. Connectivity between the cores has nothing to do with if it is a native quad core cpu (Both companies will do it differently). So multiple chips in one package or multiple packages on a motherboard is not native. That means that currently there are no native quad core solutions. Core 2 Quad is not and 4 x 4 is not. But both Intel and AMD will have native quad core solutions next year. Which one is available first. Only they AMD/Intel knows for sure but I am sure they both want to be first.

 

[Donnie27]

I like this quote. And it's quite accurate. A couple years ago they went through training at all the fabs where they taught the people how to glue CPUs together properly. It took some time... and A LOT of super glue... but I think the result was well worth it. I've actually seen reports on the money saved if Intel would get into the glue making business. That's the real outcome of the efficiency action taking place within intel right now and the costs that would be saved are insane.

Screw you man LOL! I just spit Sprite on to my Keyboarrd LMMFAO!

 

[Martyr]

intels current 4 core isnt native. intel will be release a native 4 core before amd releases theirs. people like to overlook that

 

[Donnie27]

With computers getting smaller all the time, I cant see why any manufacturer would setup 2 sockets for quadcore.

They did two sockets because they weren't finished or close enough to being finished to glue all four cores together. Then ship it a Day or two before Intel did.

 

[Donnie27]

intels current 4 core isnt native. intel will be release a native 4 core before amd releases theirs. people like to overlook that

Then AMD's isn't either.

 

[DemonDiablo]

Before I start, Im not trying to flame you here at all. Its just that sometimes the fact that you work for Intel really shines through is all. Heck I could tell you worked for Intel from your posting before I knew for fact that you worked for Intel.

I like this quote. And it's quite accurate. A couple years ago they went through training at all the fabs where they taught the people how to glue CPUs together properly. It took some time... and A LOT of super glue... but I think the result was well worth it. I've actually seen reports on the money saved if Intel would get into the glue making business. That's the real outcome of the efficiency action taking place within intel right now and the costs that would be saved are insane.

I love how you have this tendancy to compltely put words into someones mouth and to take a statement out to be something that its not. You and I both know what he meant by that quote and why he stated it as he did. He was just simply dumbing down the techno jargain for less than savy pc literate people to understand.

People need to realize that while Intels first solution to these things (first dual core, quad core, etc) may not be the best way to do them.... they are still doing them first.

So what your saying is that, even though you come out with a half-assed crappy product, so long as you're first thats all that matters? Coolbeans imo. You take pot shots at Ruiz but then you make a comment like that.

For the record though, I believe 4x4 and Kentsfield are both garbage. Neither imo is of true "native" design. My definition of native design is what aacdirect stated which Intel should be out first with next year according to road maps anyways. Thats what Im waiting for from both Intel and AMD, not these current gimmicks that they both plan on coming out with next month-ish.

 

[chrisf6969]

AMD's plan for quad core isn't "native" either... its slapped together with a shared L3 to "tie the cores together".

L3's are band-aid fixes. Like the Xeon's & Gallatins (XE) had an L3 to help compensate for their weak bandwidth issues.

Kentsfield will not be native, but the next revision of Intel's quad core at 45nm will be native with a shared L2 cache.

 

[Poncho]

I love how you have this tendancy to compltely put words into someones mouth and to take a statement out to be something that its not. You and I both know what he meant by that quote and why he stated it as he did. He was just simply dumbing down the techno jargain for less than savy pc literate people to understand.

You do realize that it was a joke. I know what he was saying... everybody here knows what he was saying. I just found it funny.

So what your saying is that, even though you come out with a half-assed crappy product, so long as you're first thats all that matters? Coolbeans imo. You take pot shots at Ruiz but then you make a comment like that.

Who says it's half assed? If it beats the competition.... then how is it half assed? These things don't happen over night, and yes in the real world of business "First to Market" is HUGE. They release the first dual core... then build on that technology. That's what most people do. Now you could go the other way and make something a different way, that may take more time, but then you'll lose out on the time that your competitors have had a similar product on the market. Intel IS releasing quad core this month... and AMD will have a years worth of momentum to regain when they come out with theirs. That is huge and it will hurt AMD. Native/non-native is simply marketing BS that is being used by AMD and !!!!!!s to slow Intel's lead down. It won't work... believe me.

Another thing... I would really like to know why is it that just because Intel goes about things a different way (MCM, FSB, etc) that they are automatically half assed solutions? Who says that you have to have an IMC, or 4 cores with a crossbar and shared L3? If you can beat your competition with the product you have.... that sounds like the better way to me.

 

[DemonDiablo]

Thats your opinion and you are certainly entitled to it. I for one personaly believe the K8L to be a true quad core design.

I fore one eagerly await Intel's native quad core design in the not to distant future. After seeing the conroe in action, I see good things for their quad core.

 

[Donnie27]

Native quad core means a single piece of silicon containing 4 cores. Connectivity between the cores has nothing to do with if it is a native quad core cpu (Both companies will do it differently). So multiple chips in one package or multiple packages on a motherboard is not native. That means that currently there are no native quad core solutions. Core 2 Quad is not and 4 x 4 is not. But both Intel and AMD will have native quad core solutions next year. Which one is available first. Only they AMD/Intel knows for sure but I am sure they both want to be first.

That might be true but here's my point. K8L isn't anymore NATIVE Quad Core than Kentsfield. I ask the Question again. Which is more native? Four Single Processors, not just Cores, tied together with Crossbar/HTT or Two Native Dual Core tied with a FSB?

No, what AMD better have is a better Core Architecture. How they're Glued together, Native or not, will not matter otherwise! Without improvements to each of the Four cores on K8L, it would get its ass handed to it, just as they are now.

Real Native or not, Kentsfield is kicking ass right now. It's kicking ass because Core 2 Duo is kicking ass. If K8L was just two FX-62's native or NOT, it would still be getting its ass spanked. Just as if Intel Glued 4 Northwoods together with a Crossbar and an IMC. Kentsfield would still kick its Booty. Core is better than Netburst and that is the real story here. Well, K8L better be better than K8 Per Core. Then THAT is what in the hell Hector SHOULD be pushing.

 

[DemonDiablo]

You do realize that it was a joke. I know what he was saying... everybody here knows what he was saying. I just found it funny.

Sorry my sarcasim meter doesnt work to well on internet forum boards. I wanted to give you the benefit of the doubt that it was a sarcastic joke but well you never know.

 

[DemonDiablo]

Whoa...on second thoughts maybe I should have finished reading your post before hitting submit.

Who says it's half assed? If it beats the competition.... then how is it half assed?

I dunno does it beat 4x4? Havent really been keeping up with info on either of the two (espically 4x4) as its just a half-assed (thats rigth you heard me) concept to get quad core out before the competition. Do we really need to bring up the Pentium D here?

These things don't happen over night, and yes in the real world of business "First to Market" is HUGE.

Nope the certainly do not. It takes time to make a well designed product, as shown by amd's and intel's roadmaps to have native quad core out later next year. But to come out with a half-assed product, that certainly doesnt take any time at all. Slap that togehter with that. Glue it up and your ready to go. First to market isnt huge. Its PR spin.

Intel IS releasing quad core this month... and AMD will have a years worth of momentum to regain when they come out with theirs. That is huge and it will hurt AMD. Native/non-native is simply marketing BS that is being used by AMD and !!!!!!s to slow Intel's lead down. It won't work... believe me.

Like I said to Chris, you're without a doubt entitled to your own opinion. If you truely believe kentsfield to be quadcore then good for you. I believe it to be quad core as well, I just dont believe it to be of native design. This just simply turns into a tomate/tomato (sp) debate though. If you call a quadcore a tomato and I call it a tomate then I just say whatever.

After all I believe the whole point of this thread was to find out whatother peoples conception of a quad core to be. Not to say your wrong Im right.

Another thing... I would really like to know why is it that just because Intel goes about things a different way (MCM, FSB, etc) that they are automatically half assed solutions? Who says that you have to have an IMC, or 4 cores with a crossbar and shared L3? If you can beat your competition with the product you have.... that sounds like the better way to me.

LOL this right here is what really got me to post a response. Whoa talk about coming out of left field with that one? Pent up anger much? Where did IMC/FSB come into the picture for what makes a true quadcore or a dualcore or a single core for that matter? In fact I do believe I said that I look forward to Intel's native quad core thats coming out next year and correct me if Im right or wrong here but I dont believe it has an imc, crossbar or shared L3. So how does that whole blurp have anything to do with me or pertain to anything in this disscussion so far? Please dont get me wrong here, I dont have anything against intel or amd here im completely non-baised non-favorist here. In fact I've been just as harsh about amd's marketing here as I have been about intel's in my posts in this thread so I dont get where all of this im an amd !!!!!! talk is coming from.

 

[(cf)Eclipse]

It is pretty silly for Mr. Ruiz to be running around talking about native Quad Core while calling a two Socket system "Quad Father", how FOS can you get?

read more carefully, he didn't specicially state that 4x4 is native quad core. he said that amd will probably be the first to launch a native quad core.. aka K8L

Ruiz predicted AMD would be first to introduce a native quad-core processor. "The quad-core you've heard about is just two dual-core chips glued together, it's not really quad-core," he said.

to me, native quad core means 4 cores on one die. yorksfield is quad core in my eyes, but not NATIVE, since it's two dual core dies on one package

 

[Poncho]

I dunno does it beat 4x4? Havent really been keeping up with info on either of the two (espically 4x4) as its just a half-assed (thats rigth you heard me) concept to get quad core out before the competition. Do we really need to bring up the Pentium D here?

Jesus Christ.... I have no idea if it beats 4x4, nor do I care. And let me tell you why..... THEY ARE COMPLEATLY DIFFERENT PLATFORMS. To compare a DP and a UP system to each other is simply stupid. A CPU doesn't make a system perform better than another for the most part. Yes it factors in, but it's not the biggest thing. I'm sure 4x4 will beat Kentsfield, I would hope it would. A DP system will do things better for the most part in regards to benchmarks. Now... the reviews that I'll be looking to for 4x4 are the ones where they compare it to Bensley which is.... oh that's right ANOTHER DP PLATFORM. Fuck... this 4x4/kentsfield debate is getting fucking annoying. It's an apples to oranges comparison.

First to market isnt huge. Its PR spin.

Actually it is huge since it.... helps you market your product better. If you have a product on the market a year before your competitor then that's an entire year of sales that they are losing out on. That's a good thing for a company. The PR spin is this native/non-native BS.

LOL this right here is what really got me to post a response. Whoa talk about coming out of left field with that one? Pent up anger much? Where did IMC/FSB come into the picture for what makes a true quadcore or a dualcore or a single core for that matter? In fact I do believe I said that I look forward to Intel's native quad core thats coming out next year and correct me if Im right or wrong here but I dont believe it has an imc, crossbar or shared L3. So how does that whole blurp have anything to do with me or pertain to anything in this disscussion so far? Please dont get me wrong here, I dont have anything against intel or amd here im completely non-baised non-favorist here. In fact I've been just as harsh about amd's marketing here as I have been about intel's in my posts in this thread so I dont get where all of this im an amd !!!!!! talk is coming from.

My whole point here is get some understanding on you overused "half assed" comment. I really want to know how something that beats the competition can be considered half assed. It seems like you think it's half assed because they didn't design theirs like AMD. Just looking for an answer to your use of language.

 

[Donnie27]

read more carefully, he didn't specicially state that 4x4 is native quad core. he said that amd will probably be the first to launch a native quad core.. aka K8L

to me, native quad core means 4 cores on one die. yorksfield is quad core in my eyes, but not NATIVE, since it's two dual core dies on one package

No, not this time. I didn't say he said 4 X 4 was Native Quad Core, so you read more carefully next time. I said they're calling 4 X 4 "Quad Father". I didn't say they're calling 4 X 4 native Quad Core. Link to where I did? I said I didn't think K8L was any more "Native" Quad Core than Kentsfield is=P

Again, if Two Dual Cores ain't Native Quad Core, then Four Single Processors sure as hell ain't IMHO, both are Quad Core and Poncho is right, the whole Native thing is BS! This is nothing more than; Since AMD has no Quad Core right now, they're trying to downplay Intel's. Please don't fall for it?

I clearly said if AMD doesn't upgrade the core, and we both know they will, the native stuff is shown for the BS that it is. You can Natively connect the current K8 all you like and it still gets pimp slapped IMHO.

Then AMD is setting themselves up to Fail. If AMD squeezes out a 1 to 7% advantage over Intel or just matches them, it'll still seen as if they failed. Folks will say stuff like; "You say you have the real Native QC support, IMC, better Architecture and that's all you can do?".

 

[NulloModo]

Personally, I don't see 4x4 beating Kentsfield at all. The midrange E6600 beats t he hell out of the top end AMD FX CPU in every benchmark, so as long as Intel can clock the best Kentsfield at equal or better levels than the E6600, they are golden.

To me, quad-core 'native' vs. 'non-native' means all of nothing, it is the numbe of CPUs on the chip that counts.

 

[As I Lay Dying]

actually 4x4 might not beat kentsfield in terms of cpu power but 4x4 might beat it in gaming benchmarks, and I really think thats all 4x4 is for, e-penii, and those who need to have 120FPS with full HDR, AA, AF at 16000x9000 widescreen. I hate to see the price tag for a quad 8800GTX SLI, 4x4 system.

 

[(cf)Eclipse]

No, not this time. I didn't say he said 4 X 4 was Native Quad Core, so you read more carefully next time. I said they're calling 4 X 4 "Quad Father". I didn't say they're calling 4 X 4 native Quad Core. Link to where I did? I said I didn't think K8L was any more "Native" Quad Core than Kentsfield is=P

Again, if Two Dual Cores ain't Native Quad Core, then Four Single Processors sure as hell ain't IMHO, both are Quad Core and Poncho is right, the whole Native thing is BS! This is nothing more than; Since AMD has no Quad Core right now, they're trying to downplay Intel's. Please don't fall for it?

true, i fail as well

what has four single cores on one package? no chip i can think of off the top of my head at least. native quad core means there are 4 cores on one die, and one die on that package. not two cores per die and two dice per package.

indeed it is marketing, but you asked, what is NATIVE quad core. i'm giving my opinion. not telling you if it's worth anything, just defining a term

and if you ask me, sharing L1 and L2 between all cores is dumb beyond belief. cache coherancy and thrashing ftl. if you're gonna draw the line between native and non-native as 4 "cores" that share the same cache, so be it. just be aware that multithreading performance will lack, as shown with conroe, there is a relatively large drop in performance when running two tasks at once vs a single task, when compared to K8.

but then brings the question.. i guess K8 isn't native dual core either, since it has seperate L1 and L2? gotta draw the line somewhere
and disregarding the 'native' bit, 4 cores on one package. how "large" are you letting the package be? realistically speaking, 4x4 will give the same performance as a quad core revF K8 part... if not a smidge better due to the extra ram bandwidth. you're not gonna call that quad core, despite it having 4 cores in one "package", just the package would be considered from an external standpoint. if you look performance only and treat the computer as a black box if you will....

 

[Killroy]

While a "native" quad-core design is better than a "non-native" quad-core design with the same basic architecture, I see Kentsfield as being a kind of stop-gap until Intel releases their "native" quad-core. That being said, I don't see Kentsfield as being any kind of hack or half-assed design. I just see it as Intel offering the most powerful (in some but not all applications and situations) processor they can, giving the consumer more options, and so that they can say they were the first to market with four cores in one package. Kentsfield is a quad-core processor regardless of whether or not it's "native".

 

[Donnie27]

true, i fail as well

what has four single cores on one package? no chip i can think of off the top of my head at least. native quad core means there are 4 cores on one die, and one die on that package. not two cores per die and two dice per package.

indeed it is marketing, but you asked, what is quad core. i'm giving my opinion

and if you ask me, sharing L1 and L2 between all cores is dumb beyond belief. cache coherancy and thrashing ftl. if you're gonna draw the line between native and non-native as 4 "cores" that share the same cache, so be it. just be aware that multithreading performance will lack, as shown with conroe, there is a relatively large drop in performance when running two tasks at once vs a single task, when compared to K8.

but then brings the question.. i guess K8 isn't native dual core either, since it has seperate L1 and L2? gotta draw the line somewhere

Quad Core is just that Quad Cores. Four cores on one Package. No, what's truly Dumb is his NOT talking about improvements made to the Core K8 Architecture while Bullshitting folks with this whole Native BS! I didn't try to say how I thought Native Quad core would perform. Hell, I don't know. With all due respect, I don't think YOU know how it would perform either.

First, I respect your Opinion and unlike that other person on the AMD side of the forum, I didn't say anyone was "wrong" because I disagree with them. I just posted a big *if about my own Opinion. I can't state it as fact, too much is unknown.

K8L is four single complete processors with a Core, L1 and L2 that could connect to an external Memory controller and run alone. IMHO, this wouldn't be the case with Native as it is described by AMD. I agree with the other guy that K8L's use of a L3 helps make that point. Now, the connection might be native LOL!

I'm really tired of that whole AMD thinks everyone is stupid so they must simplify by lying CRAP. Hell, tell the truth and maybe their customers will learn something? Maybe I'm tired of being lied to for the sake of simplicity or for simplicity's sake.. Yes, that goes for Intel as well. I'll go back to saying both are Quad Core and each has advantages, one we know of and the other we're waiting to see, fair?

 

[Donnie27]

but then brings the question.. i guess K8 isn't native dual core either, since it has seperate L1 and L2? gotta draw the line somewhere
and disregarding the 'native' bit, 4 cores on one package. how "large" are you letting the package be? realistically speaking, 4x4 will give the same performance as a quad core revF K8 part... if not a smidge better due to the extra ram bandwidth. you're not gonna call that quad core, despite it having 4 cores in one "package", just the package would be considered from an external standpoint. if you look performance only and treat the computer as a black box if you will....

Just for this part. I'm not so sure if a Dual Core K8L on a Dual Socket Board wouldn't be faster than a single socket Quad Core K8L. I'd put my money on the Dual Core Dual Socket, same goes for Two Woodys vs. One Kentsfield. Personally, I'd take the Two Woodys.

I'm thinking 200mm-squared and it would be smaller if we're talking Four exec units united with a Single LARGE L1 (might be very expensive). But what about 4 L1s and just one 32MB L2? 8MB of us poor folks? Again dude I don't know.

 

[Poncho]

No, what's truly Dumb is his NOT talking about improvements made to the Core K8 Architecture while Bullshitting folks with this whole Native BS!

You are going to be seeing a lot more of this in the next year. AMD is in scramble mode right now, "revamping" their message since all their old marketing strategies don't hold true for them anymore. You've seen the beginning of it when they started saying that benchmarks don't really matter and now the native BS. The simple fact is that they are lagging behind Intel by a great deal. They are a year behind in 65nm, they'll be a year behind in quad core, and by the time they catch up to those points... intel will be onto the next thing. Not to mention the fact that Intels current CPU is beating their asses across the board, and more than likely AMDs next generation won't compete either. They've got their work cut out for them that's for sure.

Oh... and before any says it yes I know that Intel used these same tactics in the past when they had an under performing CPU and I don't fault either company for what they are doing. This is all business/marketing/strategy.

K8L is four single complete processors with a Core, L1 and L2 that could connect to an external Memory controller and run alone.

You know this is interesting. If you think about it what is the difference between AMD and Intel's quad core designs really? They both connect to the memory controller in a similar fashion... the only difference is that AMDs is on the die, whereas Intel's is external. Really... take the IMC/FSB out of the equation and you have very similar designs. At least for a package perspective.

 

[Donnie27]

While a "native" quad-core design is better than a "non-native" quad-core design with the same basic architecture, I see Kentsfield as being a kind of stop-gap until Intel releases their "native" quad-core. That being said, I don't see Kentsfield as being any kind of hack or half-assed design. I just see it as Intel offering the most powerful (in some but not all applications and situations) processor they can, giving the consumer more options, and so that they can say they were the first to market with four cores in one package. Kentsfield is a quad-core processor regardless of whether or not is "native".

But as always, it's the end results that matter most. Again, the "Native" thing is nothing but a Marketing angle. "We don't have it, so let's knock it until we do". How do we know? By saying it is a poor implementation, not Native, not well thought out, Rushed, quick fix, stop -gap and etc..... Meanwhile, it goes right alone kicking ass and taking names. As the old say goes;

"If you can't dazzle em' with fancy footwork, baffle em' with Bullshit!"

I'll say it now and please hold me to it? If K8L is kickass and kicks Intel's ass, it will be due to Core improvents, not because of any cock-a-mammy Native BS! I remember this same Crap when the so called Native Dual Core X2 launched. X2 kicked S/DC Netburst's ass because it was the better Core Architecture, NOT because it was More Native LOL! The single core versions proved that.

 

[Donnie27]

You are going to be seeing a lot more of this in the next year. AMD is in scramble mode right now, "revamping" their message since all their old marketing strategies don't hold true for them anymore. You've seen the beginning of it when they started saying that benchmarks don't really matter and now the native BS. The simple fact is that they are lagging behind Intel by a great deal. They are a year behind in 65nm, they'll be a year behind in quad core, and by the time they catch up to those points... intel will be onto the next thing. Not to mention the fact that Intels current CPU is beating their asses across the board, and more than likely AMDs next generation won't compete either. They've got their work cut out for them that's for sure.

Oh... and before any says it yes I know that Intel used these same tactics in the past when they had an under performing CPU and I don't fault either company for what they are doing. This is all business/marketing/strategy.

You know this is interesting. If you think about it what is the difference between AMD and Intel's quad core designs really? They both connect to the memory controller in a similar fashion... the only difference is that AMDs is on the die, whereas Intel's is external. Really... take the IMC/FSB out of the equation and you have very similar designs. At least for a package perspective.

I see AMD at least closing the Gap though.

I'll just say YES you got it and post again later! It's kind of nice to post like this and not have Lame Flamers screwing it up.

 

[InorganicMatter]

I find it funny that you guys are arguing over core connection techniques. Each has it's advantages and disadvantages. K8L is four cores on one die connected with a high-latency L3 cache. Kentsfield is two dies on a piece of silicone connected by a Netburts FSB. Each has it's strengths and weaknesses. Why is it that AMD's is automatically "native" and (therefore superior) if the 4 cores are on one die? Four cores on one die connected with a high-latency cache is worse than two dies connected with a slow but highly efficient FSB. What makes "native" so desirable? You guys sound like politicians: "AMD's way of doing quad-core automatically defines what quad-core is, and nothing Intel does is "native" because it's different than what AMD does and therefore automatically wrong."

 

[As I Lay Dying]

I think what, AMD is claiming is that two dies slapped onto a package isn't native. This is what Intel did with Pentium D, and are doing with Kentsfield, AMD is claiming that two cores on one die is native dual core, and 4 cores on one die is native quad core. I do not think AMD has claimed core duo and core 2 duo which are two cores on one die, as not being native, just Pentium D's and Kentsfield.

 

[spectrumbx]

What some here are trying to get at is the efficiency issue of native-multiple-cores vs the non-native ones.

We've all seen how conroe performs better and uses less power than the P4 D (yes, I understand the difference in core archicture, but something about the dual-core nativeness is helping conroe).

Now, the P4 D had its merits and so does Kentsfield.
I prefer two chips packaged under one socket to two sockets.

However, at 120watt, people (me in particular) are looking forward to a quad-core solution that will use far less power.

 

[Donnie27]

What some here are trying to get at is the efficiency issue of native-multiple-cores vs the non-native ones.

We've all seen how conroe performs better and uses less power than the P4 D (yes, I understand the difference in core archicture, but something about the dual-core nativeness is helping conroe).

Now, the P4 D had its merits and so does Kentsfield.
I prefer two chips packaged under one socket to two sockets.

However, at 120watt, people (me in particular) are looking forward to a quad-core solution that will use far less power.

But that's a point as well. That's one 120W Quad core vs. two Dual core FX-7x processor that are speced for 125W* worse case, EACH. Even if they run at Half that temp, that's still higher than Kentsfield.

Not that AMD dynied that Core was Native Dual Core, they never said it was either. Folks at Intel said Core was made to scale natively more cores. They even bragged about it being built right from the start to go to 32cores.

You know, a little Honesty never hurt Intel or AMD. Here's an example.

http://www.eweek.com/article2/0,1895,1974778,00.asp

I E-mailed this to a Bud. He said too late, we already E-Val-ed Woody vs Opteron and Woody won every power test we performed.

I still thought this comment on page two is what we need more of.

Tulsa, Skaugen said, will allow Intel to get back into the game in multiprocessor servers, an area where AMD's Opteron has been strong of late. Even Dell, once an Intel-only shop, has chosen to offer multiprocessor servers based on the Opteron.

"We have lost share on Xeon MP particularly in the performance space. We got behind on performance, shame on us," he said.

But, "With the performance we're seeing out of Tulsa…we're confident that we have a significant lead on transaction performance…the primary criteria on which I think we lost share is addressed by the Tulsa product."

 

[MrWizard6600]

please for the love of god, lets stop this quoting of the inq!

 

[Lolento]

please for the love of god, lets stop this quoting of the inq!

Seriously, some of the articles in the Inq are really ridiculous, lacks logic and most are purely speculative....

If they allow readers to comment on the articles and post the comments below the articles like other news sites, I'm pretty confident they'll get miles and miles of flames.

 

[Donnie27]

I find it funny that you guys are arguing over core connection techniques. Each has it's advantages and disadvantages. K8L is four cores on one die connected with a high-latency L3 cache. Kentsfield is two dies on a piece of silicone connected by a Netburts FSB. Each has it's strengths and weaknesses. Why is it that AMD's is automatically "native" and (therefore superior) if the 4 cores are on one die? Four cores on one die connected with a high-latency cache is worse than two dies connected with a slow but highly efficient FSB. What makes "native" so desirable? You guys sound like politicians: "AMD's way of doing quad-core automatically defines what quad-core is, and nothing Intel does is "native" because it's different than what AMD does and therefore automatically wrong."

It's marketing BS, that's all! If I'm AMD, I talking up and Doing Core improvements. The weak point in Hector's argument is K8 and it's Crossbar was built to be scalable right? Native becomes a Gimmick because if you could do a so called Native Current K8 FX-60 for example, it would still get spanked by a so called Non-Native Kentsfield. I'm not talking about a K8L. If K8L is faster and that's possible, it will because of its Core Architecture Improvements NOT because of its so called a "True Native Quad core" nature.

 

[brucedeluxe169]

AMD's tech is so goddamn far above Intel's head its just ridiculous...

AMD has IMC, HT, NATIVE DUAL CORE and NATIVE QUAD CORE designs

Intel is so goddamn far behind tech wise.....







.... Conroe is faster than anything AMD has?

.... Dothan coulda bested AMD 3 years ago on the desktop if Intel didn't hold it back for BS marketing reasons?

.... Intel beats AMD's tech with totally inferior technology that cannot compete at all to AMDs far superior engineering and know how?

.... AMD's supposed winning streak for performance was directly the result of Intel's managers listening to the marketing guys instead of common sense telling them to use the superior tech they *already had* ???



*amd f@nboy heads explode*