AMD Unveils "Barcelona" Architecture

[3NZ0]

They knew BARCELONA was coming and would eclipse everything they've EVER created so threw some PIII's together and came up with crap-2-duo

Wrong, heard of the pentium M?

And crap2duo? Not worth responding to that.

They took the same approach with C2Q's......fire a couple of crap-2-duo's together and launch them as a quad-core

And the problem there is?
It works and in the G0 stepping, clocks like a mofo under air and is a very cool (and relatively cheap) cpu.

"Look everyone.....we've beaten AMD to a pair of DC's on one die....awesome!"

Meanwhile, back in AMD labs.....some of the techs were designing BARCELONA and not skimping on it either.

to AMD PWNAGE

Mofo, i had quite a bit of respect for you when you started those benching results threads with detailed lists in but that post makes you look like a blatant fanboy and as a result you have lost a bit of respect from me, unfortunately.

[largon]

I wonder how revolutionarily ingenius does SOLDNER-MOFO64 declare the MCM approach when AMD introduces Montreal octacore...
You know, Montreal, the dual die Shanghai MCM.

Oh the irony...

[madcho]

I wonder how revolutionarily ingenius does SOLDNER-MOFO64 declare the MCM approach when AMD introduces Montreal octacore...
You know, Montreal, the dual die Shanghai MCM.

Oh the irony...

Montreal use HT3 link, one die to the other die. The first die don't send data to chipset before go in the 2nd die. They're going directely from the first die to the 2nd one using HT3 links.

A lot faster than fsb way, a very short behind the native way.

fsb is the past, Hypertransport's way is the futur

[Nasgul]

Well, a whole year+ to come up with a new architecture to fight against a year-and-a-half old architecture.

[SOLDNER-MOFO64]

Intel arent stupid for releasing 2 dual-core dies on 1 chip to get a Quad- 95%+ of people dont give a **** how they get 4 cores onto 1 CPU, as long as they do. From an engineering POV, its a little crude, but in terms of making money and getting something onto the market, its beautiful.

I didn't say INTEL were stupid for doing so.....rather smart if you ask me. They managed to get a load of guys to buy 2xC2D's in one package by telling them it was a quad-core.

All cars have 4 wheels ('cept for maybe RELIANT ROBIN), but they ain't all 4 WHEEL DRIVE.

At the end of the day INTEL chose that route in order to beat AMD to the punch....nothing more.

And the problem there is?
It works and in the G0 stepping, clocks like a mofo under air and is a very cool (and relatively cheap) cpu.


Mofo, i had quite a bit of respect for you when you started those benching results threads with detailed lists in but that post makes you look like a blatant fanboy and as a result you have lost a bit of respect from me, unfortunately.

Aww come on mate, your not gonna go all biscuit arsed on me are u? I'm not flaming YOU personally....I merely used a non-flattering term toward your dearly beloved cpu is all. No need for us to fall out over it.

I wonder how revolutionarily ingenius does SOLDNER-MOFO64 declare the MCM approach when AMD introduces Montreal octacore...
You know, Montreal, the dual die Shanghai MCM.

Oh the irony...

Yeah, one main difference......AMD ain't fallin' over themselves to release it before INTEL do. AMD ain't cutting corners and creating FAKE cpu's with FALSE names in order to sell them.

INTEL made C2Q's which at their own admission ARE NOT TRUE/NATIVE quads. I'm simply pointing that out....I didn't ask them to make them that way. It's not my fault C2Q owners are running dual-C2D's.

AMD (you know that little sucky company that's about to go broke) have managed to engineer a TRUE/NATIVE quad-core cpu with 4 independant cores and INTEL fanboys just can't take it.

Nice SHANGHAI argument there but sadly it's died as quickly the C2Q's (dual-C2D's) have.

Well, a whole year+ to come up with a new architecture to fight against a year-and-a-half old architecture.

Yeah, like I said.....AMD design QUALITY products that actually ARE what they say they are. AMD don't shove half-assed FANBOY cut&shut CPU's out the door just 'cos someone wants a faster super pi time.

Besides, how long did it take INTEL to glue 3xPentium III's together to fight off A64?? 3yrs?

[FghtinIrshNvrDi]

mofo64, I'm an amd fanboy, but I'll be honest... I'm glad they released these glorified pentium 3's. My pair of pentium 3's @ 3.4ghz are a substantial upgrade.

Ryan

[SOLDNER-MOFO64]

mofo64, I'm an amd fanboy, but I'll be honest... I'm glad they released these glorified pentium 3's. My pair of pentium 3's @ 3.4ghz are a substantial upgrade.

Ryan

Yeah, we know a C2D has an impressive Pi time.....and have good performance in lots of apps, but I merely touched upon INTEL's decision to call their CPU's QUAD-CORES, when in reality they're dual-dual-cores.

Glad your happy with your CPU though.....that's all that matters anyway really

[JumpingJack]

I didn't say INTEL were stupid for doing so.....rather smart if you ask me. They managed to get a load of guys to buy 2xC2D's in one package by telling them it was a quad-core.

All cars have 4 wheels ('cept for maybe RELIANT ROBIN), but they ain't all 4 WHEEL DRIVE.

At the end of the day INTEL chose that route in order to beat AMD to the punch....nothing more.

And Intel did a good job at it wouldn't you say? Getting 4 cores into the market sooner rather than later allowed them to take all sorts of crowns on the performance front.

2 x 2C2Ds at 2.66 Ghz as you put it did indeed out peform 2x2 Optys at 3.0 Ghz. Not to mention the server market has shifted significantly back toward Intel's quads, Intel gaining some 5-6% MSS back from AMD and continuing on that trend....

Oddly, 2x2 C2Ds shows 4 cores in taskmanager, multithreaded code shows all 4 of those cores being utilized, and multithreaded apps are scaling with core count as if it were a quad core.... so, 4 cores working must mean quad cores working (at least that is what most people associate with the word quad).

The truth of the matter is that integrating 4 cores onto the same die as Barcey did is certainly technically more elegant; however, there is a cost associated with that... AMD has not had a quad solution for almost a year after Intel put their quad into play... Barcey is, by Hector's admission, late by 6 months and, if IBM's leaked PDF is any indication, will not show real volume until November.

Add to that a larger die and lower yields, and Barcey will be more expensive for AMD to mass produce as opposed to the MCM approach. So while latency will not allow an MCM to exploit the full potential, the real question is how much of that potential is lost... 1%, 2% ...20% ... and while Barcey will scale (per core not per socket) better in multithreaded code, the quesiton is will the absolute performance over take the MCM... the answer to that is likely no.

We have a hint from IBM's leaked data that suggests how the desktop version of this may play out clock for clock. Barcey is not out performing clovertown in the non-rate SPEC scores, which is more indicative of the multitude of single threaded apps that will run on the DT workload. Barcey will be competitive in server, it will smash Intel in FP intensive HPC applications, but the likelihood that your stipulation that Barcey is the next best thing since peanut butter is probably not going to be accurate.

[rozzyroz]

All cars have 4 wheels ('cept for maybe RELIANT ROBIN), but they ain't all 4 WHEEL DRIVE.

actually its more of a Ford 4 wheel drive vs. a Jeep 4 wheel drive. they do things differently, but arrive at the same outcome.

4 cores are 4 cores, its not like you cant use 2 of the 4 cores on an intel quad right?

[Nasgul]

actually its more of a Ford 4 wheel drive vs. a Jeep 4 wheel drive. they do things differently, but arrive at the same outcome.

4 cores are 4 cores, its not like you cant use 2 of the 4 cores on an intel quad right?

No matter which way you shape it, he won't understand.

It would be funny though, to see a car with four wheels and the moment you hit D, two wheels get suspended in the air and there, you take off on two wheels, that's how he probably thinks.

A wheelie!

[SOLDNER-MOFO64]

Very good point you make.

IMO though there was no real need for quad-cores a year ago anyway, we're just starting to see apps which utilise all 4 cores and actually need to...so AMD being 6months late is neither here nor there. It's not like there were loadsa games/apps you haven't been able to use if you haven't had a quad. Sure the CINEBENCHES of the world complete faster on a quad but gamers have seen no real performance boost.

I appreciate what INTEL have done, but at the end of the day it only looks like a stunt they pulled to beat AMD to quads...considering they are by far the largest and most dominant of the two companies...you'd have thought THEY would've taken their time and designed a native quad themselves instead of rushing out a cut&shut quad which still uses a FSB.

If they'd taken another 6months+ and launched a native quad then perhaps AMD would have a much tougher job trying to surpass them. INTEL are the BIG company remember.....AMD are the small guys with -loads of errors to their name....delays after delays.....facing bankruptcy, low yields, etc etc etc.....yet INTEL are the guys with all the knowledge and wonga, yet they rush out cpu's as if they're desperate for money!!!!

IMO I thought INTEL woulda went about this whole saga in a different way. It's like they knew they had one up on AMD with C2D and thought "Let's run with this as fast and far as we can before AMD reltaliate".

[SOLDNER-MOFO64]

No matter which way you shape it, he won't understand.

It would be funny though, to see a car with four wheels and the moment you hit D, two wheels get suspended in the air and there, you take off on two wheels, that's how he probably thinks.

A wheelie!

Now see you're opening your mouth but all I hear are your BOWELS.....pls move along......kindergarden is THAT way >>>>>>>>

[Hornet331]

Very good point you make.

IMO though there was no real need for quad-cores a year ago anyway, we're just starting to see apps which utilise all 4 cores and actually need to...so AMD being 6months late is neither here nor there. It's not like there were loadsa games/apps you haven't been able to use if you haven't had a quad. Sure the CINEBENCHES of the world complete faster on a quad but gamers have seen no real performance boost.

why do you think we see now apps that make use of quad core ?

right because they are there, there would be no consumer apps that would make use of quads right now.

[JumpingJack]

Very good point you make.

IMO though there was no real need for quad-cores a year ago anyway, we're just starting to see apps which utilise all 4 cores and actually need to....

Except in server applications.... the majority of those are multithreaded. Hence, the success Intel has enjoyed pushing the MCM approach, driving 4 execution engines into one socket.

On top of that, someone has to break the ice and provide incentive for software to push into the multithreaded arena.

[freeloader]

Yeah, we know a C2D has an impressive Pi time.....and have good performance in lots of apps, but I merely touched upon INTEL's decision to call their CPU's QUAD-CORES, when in reality they're dual-dual-cores.

Glad your happy with your CPU though.....that's all that matters anyway really

You say "tomato", I say, "tamato". Let's see how AMD's native quadcore compares to two glued together C2Ds.

[SOLDNER-MOFO64]

Except in server applications.... the majority of those are multithreaded. Hence, the success Intel has enjoyed pushing the MCM approach, driving 4 execution engines into one socket.

On top of that, someone has to break the ice and provide incentive for software to push into the multithreaded arena.

Granted, but what success would they have had with a native quad-core and an even greater performance increase over opteron? Then they'd have less to worry about from AMD with theirs.

As it stands, the 5%-6% market-share they may have recovered is gonna all but dissapear overnight when AMD launch they're OPTY-QUADS.

So now INTEL have an even bigger fight on their hands than they would've had they gone the native route. IMO they'd have got there before AMD with all they're $$$$$ and know-how....and would have left AMD in a very precarious position. Alas they didn't and decided to rush out C2Q's still clinging onto the FSB like nobody's buisness and raking in as much $$$$ as possible....and as a result now find themselves in this position, by which I mean staring at BARCELONA.

[savantu]

You say "tomato", I say, "tamato". Let's see how AMD's native quadcore compares to two glued together C2Ds.

Even if it'll lose they'll resort back to " but it's a native QC and that's all that matters " !



There is a term which defines this : zealots.

[mstp2009]

....AMD ain't fallin' over themselves to release it before INTEL do. AMD ain't cutting corners and creating FAKE cpu's with FALSE names in order to sell them.

That "fake" CPU at almost 12 months old looks like it is going to put the smack down on the "real" quad except in a very few specific benchmarsk.

And that is ALL I care about. PERFORMANCE.

[SOLDNER-MOFO64]

Even if it'll lose they'll resort back to " but it's a native QC and that's all that matters " !



There is a term which defines this : zealots.

Much the same as when it out-performs your pentium 6 (2xPENTIUM-3's) you'll moan "it's supposed to, ours are only pentium III's glued together in a fake quad-setup "

So what's the term for someone who buys a product which they are lead to believe is something which it is not? DUPED?

What's the term for someone who cries into they're cornflakes when BARCELONA is released? INTEL-USER

[frankR]

I don't really care about the native quad core then. What matter is performance. It could be 4 single cores in MCM for all I care.

[savantu]

Granted, but what success would they have had with a native quad-core and an even greater performance increase over opteron? Then they'd have less to worry about from AMD with theirs.

As it stands, the 5%-6% market-share they may have recovered is gonna all but dissapear overnight when AMD launch they're OPTY-QUADS.

So now INTEL have an even bigger fight on their hands than they would've had they gone the native route. IMO they'd have got there before AMD with all they're $$$$$ and know-how....and would have left AMD in a very precarious position. Alas they didn't and decided to rush out C2Q's still clinging onto the FSB like nobody's buisness and raking in as much $$$$ as possible....and as a result now find themselves in this position, by which I mean staring at BARCELONA.

Intel's decision to go with an MCM aproach was based on the following factors :

-costs
-performance
-time to market

Barcelona falls flat on all 3 :

-huge die , expensive to make
-looks like it will perform worse than C2Q
-no comment

Intel executives said that single die QC is too much for 65nm , the sweet spot is 80-150mm^2.Barcelona is 2x that and this will bite AMD dearly.

Secondly , AMD looks more and more like 3DFX preparing to release the 5000 and 6000 series ; huge monster that ultimately were the last nails in the company's coffin.

[Hornet331]

Much the same as when it out-performs your pentium 6 (2xp3's) you'll moan "it's supposed to, ours are only pentium III's glued together in a fake quad-setup "

So what's the term for someone who buys a product which they are lead to believe is something which it is not? DUPED?

What's the term for someone who cries into they're cornflakes when BARCELONA is released? INTEL-USER

wth are you talking about ?

you pay for 4 cores on one cpu, you get 4 cores on one cpu.

[JumpingJack]

Granted, but what success would they have had with a native quad-core and an even greater performance increase over opteron? Then they'd have less to worry about from AMD with theirs.

This is where I think people need to really take some time and study the economics of this industry and think through the logic of what it takes to design a CPU.

The cost to manufacture a CPU is wrapped around a whole bucket of details, but ultimately it fits within die size and yield... if the die size is large yield is smaller, just by nature of the way CPUs are made. Example say you have 3 random particle defects that kill 3 random die on a wafer. Just as an example, say one die size puts 10 die to a wafer, but a different approach (smaller die) puts 20 die to a wafer, in one case you can kill 3 of 10 die, and yield is 70%... in the second case you kill 3 die or get 17 of 20 die working, yield is 85%. So to get the same through put and on larger die, a company (in this example) needs to increase capacity (tooling) by 15%... this runs into the hundreds of millions of dollars.

Making a monolithic die at 65 nm is relatively large, both AMD and Intel have the quad die of about 280 mm^2. But Intel will get higher yields as a killer only kills 1/2 the die.

Make no mistake, Intel will undoubtedly continue this trend... their monolithic quad will be at 45 nm, to go 8 core they will MCM it.... the monolithic 8 core will be at 32 nm and the 16 core will be MCM.... Intel takes the cost and design complexity and balances that approach to optimize yield and lower costs. Hence, we have a 270-300 buck quad core available to us.

Now, AMD will get a 250-300 buck quad to you as well... but they will not make nearly the money... this is good for you and me, but bad for AMD as their cost will be higher overall.

The compromise to time to market and costs is slightly lower performance scaling across threads.... in servers Intel is paying a larger price for this though as the BW of a shared bus does not compound with core count, this is the uniqueness of AMDs platform design/approach and has enabled AMD to get throughput performance way way up... just smart if you ask me, I think Intel is making a big mistake sticking with the parallel UMA approach this long.

In terms of how much better would they have done... it doesn't matter, pre-Barcelona -- a 2P server with quads just smashes anything 2P with dualies. The question was.. could Intel bring that in at reasonable costs such that the price/performance is compelling... indeed they did... Intel's Xeon 53XX series were not priced 2x above the dualies, rather some 10-20% overall... this is just killing AMD, if AMD had the resources and the design features capable they would have MCMed K8 to get quads out there asap.

http://www.fool.com/community/pod/2007/070823.htm

Rivas was quoted as saying, "If I could do something different, I wish we would have immediately done a MCM - two dual cores and call it a quad-core."

[mstp2009]

I wonder how revolutionarily ingenius does SOLDNER-MOFO64 declare the MCM approach when AMD introduces Montreal octacore...
You know, Montreal, the dual die Shanghai MCM.

Oh the irony...

By the time Shanghai comes out, Nehalem will be out.

Native quad vs. native quad.
CSI vs. HT
IMC vs. IMC


And they way Intel is humming along, I've got $5 that they get there first and do a better job. The IPC w/ Conroe/Penryn looks like it is already better than K10. Goodness knows what kind of IPC increases are in store for us with Nehalem.

[v_rr]

Intel's decision to go with an MCM aproach was based on the following factors :

-costs
-performance
-time to market

Barcelona falls flat on all 3 :

-huge die , expensive to make
-looks like it will perform worse than C2Q
-no comment

1- Barcelona core is smaller then Intel core. And a native core is much better then gluing cores.
2- That comment about performance is nothing but your personal believe. It´s just Fud what you are saying about performance.
3- Late.