AMD Announces August Release Date for Barcelona

An anonymous reader writes "Rumors said the release wouldn't be until late Q4 but an August ship date is now promised for AMD's quad-core chips. They're only releasing up to 2.0 GHz processors at first, with the top speed devices coming out later in the year. 'AMD's Barcelona puts four cores on a single slice of silicon, an approach AMD calls native quad-core, and the company has argued that Barcelona will outperform the Xeon 5300. The only problem: that comparison soon will become obsolete. Intel's second-generation quad-core server processors, Harpertown a server member of Intel's Penryn family, will arrive this year, too, with the promise of better performance, lower power consumption and lower manufacturing costs by virtue of a manufacturing process with 45-nanometer features. AMD is only just now moving to a 65-nanometer process.'"

Comments

Lower Manufacturing Cost?

[enjerth]

Lowering the costs of manufacturing may be nice, but I don't think it's a particularly significant factor in their pricing model.

Thanks, anyways.

Re:Lower Manufacturing Cost?

[enjerth]

I should have clarified.

RE: Intel's Penryn family mentioned in the summary. Yeah, AMD may have higher manufacturing costs, but they're cheaper! DUR!

Re:Lower Manufacturing Cost?

[edwdig]

Lowering the manufacturing cost doesn't make a huge difference on the price marketing sets, but it makes a huge difference on what the engineering team does.

The engineering team is given target transistor count based on expected manufacturing costs, which affects the cache size and feature set.

Marketing and hubris may have done AMD in.

[RightSaidFred99]

They were so vocal about "true dual core" and then "native quad core", they made it sound like it actually meant something important. Now they're paying the price - they can't manufacture it. All indications I've heard are that they're having production problems. Compare this with the alternative of just gluing a few dual cores together. AMD can mock this approach all they want, we'll see who's laughing when they're "next gen" chip underperforms (in many benchmarks, I'm betting) a previous gen competitor's chip and falls quite a way behind the competitor's "next gen" chip.

Re:Marketing and hubris may have done AMD in.

[suv4x4]

Now they're paying the price - they can't manufacture it. All indications I've heard are that they're having production problems. Compare this with the alternative of just gluing a few dual cores together. AMD can mock this approach all they want, we'll see who's laughing when they're "next gen" chip underperforms (in many benchmarks, I'm betting) a previous gen competitor's chip and falls quite a way behind the competitor's "next gen" chip.

Looks like you're mocking the outcome of a future event that has not happened yet.

IT is a funny place to be: sometimes when it seems you're a total loser, you are, but sometimes, you come on top and kill the competition.

It's all about the details, details which you don't know.

Re:Marketing and hubris may have done AMD in.

[fitten]

Actually... it's odd, too... because AMD's architecture (HT/etc.) really lends itself nicely to MCMs... it should be almost trivial for them to make an MCM quad core part.

Re:Marketing and hubris may have done AMD in.

[RightSaidFred99]

True enough. I can only judge based on known information and rumor because, well, there's nothing else to judge on at this point. Based on those things it ain't looking good for Barcelona.

Re:Marketing and hubris may have done AMD in.

[Adult film producer]

Hey, come on mods, that was funny, admit it :)

Re:Marketing and hubris may have done AMD in.

[Kohath]

Marketing hype is not relevant. It's not relevant when it's true. It's not relevant when it's false. It's not relevant when your marketing predicts a win and you win. It's not relevant when your marketing predicts a win and you lose.

All the fanboyism and taunting and one-upsmanship and told-you-sos are worth exactly zero dollars.

The chips will perform the way they perform. There will be benchmarks. People will buy based on cost vs. performance decision-making, not cost vs. hype decision-making.

Re:Marketing and hubris may have done AMD in.

[semiotec]

I really hope that AMD can pull themselves out of the current slump.

Their technology have always been competitive with Intel, regardless of whether they are holding the performance crown of the moment, and thus they provide the only true competition to Intel in the mainstream PC market. Unlike Via or the defunct Transmeta and others, which only managed to compete in some niche markets.

we'll see who's laughing when they're "next gen" chip underperforms (in many benchmarks, I'm betting)

Should AMD go down, even Intel fanboys are going to feel the pain when Intel starts ignoring the cheap segments and prices CPU whatever way they feel like. In a way, it'd be a worse monopoly than Microsoft, since it's much easier to create software from scratch than it is to create hardware from scratch. If the unthinkable happens, we can only hope that IBM (or maybe Sun) becomes interested in making x86 chips enough to provide an alternative, or provide cheap Power processors for desktops...

Personally, I don't care who's got the highest performing CPU, as long as I can get cheap CPUs that will do the job adequately.

Re:Marketing and hubris may have done AMD in.

[NateE]

Marketing hype matters when you have ill informed, non-technical people making purchasing decisions.

Re:Marketing and hubris may have done AMD in.

[poopdeville]

All the fanboyism and taunting and one-upsmanship and told-you-sos are worth exactly zero dollars.

If what you said was true, neither Intel nor AMD would spend a dime on marketting.

Re:Marketing and hubris may have done AMD in.

[RightSaidFred99]

How are my comments odd? AMD is releasing Barcelona at a slower than expected speed. Now I know GHz ain't everything, but the fact is that for a chip this delayed they had to come out strong and they're not. My assertion is that this is because they're having problems getting good yields and enough higher clocking CPUs. They're trying to produce a chip that has 4 cores and add to that that it's on a relatively new (to AMD anyway) process and things don't look good.

My other point was that AMD has mocked the "glue 'em together" approach. Well, they can still mock but they certainly won't be mocking all the way to the bank.

Re:Marketing and hubris may have done AMD in.

[geekoid]

hahahahaa...

Marketing sells lots, and lots of crap. Many very good products cease to exist because of poor marketing(usually do to attitudes like yours) or a competitor out marketing you.

Re:Marketing and hubris may have done AMD in.

[Vancorps]

So you're saying AMD shouldn't mentioned the only reason Intel beat them to it is the fact that Intel cut stupid corners to produce a product for the sake of being first?

Stability of AMD server chips certainly shows that AMD has an idea about SMP while Intel is still scratching their collective balls. When things go to 64 bit you really start to see AMD's strengths as their current offerings are more than competitive. The desktop arena is a tough market that Intel current has the crown for but only because of Intel cutting corners. The technology itself is not superior to what AMD is offering. The only difference being that Intel can manufacture and when they want to, they can engineer. For some reason Intel is sticking with the whole FSB idea for communicating among cores. That's just plain silly although admittedly easier than AMD's approach. Which is why Intel is winning right now. It remains to be seen whether this was a short term gain with a long term sacrifice. Intel certainly has a lot of money so they could be doing both capitalizing on what AMD lacks working on their own upcoming products to meet or beat AMD's next gen. Should be interesting to see if Intel learned it's lesson from the Athlon days of glory.

Re:Marketing and hubris may have done AMD in.

[Kohath]

Those people will buy a Dell or an HP. The guys at Dell and HP are not ill-informed.

Re:Marketing and hubris may have done AMD in.

[Kohath]

When the products are essentially the same, then the marketing matters.

But not like the original poster meant -- he was saying AMD would be proven wrong about some hype-full marketing assertion and somehow "lose face" or something. It doesn't work that way. Hype is forgotten. If you remember it, you don't understand it.

Re:Marketing and hubris may have done AMD in.

[jcgf]

The guys at Dell and HP are not ill-informed.

Working on their machines often makes one think otherwise.

Re:Marketing and hubris may have done AMD in.

[2ms]

If your theory that marketing is irrelevant were true then there wouldn't be maketing. The companies spend billions on marketing because it works. Most people who buy computers do not know squat so they buy based on image or reputation, obviously. Marketing and hype, obviously, affect people's perceptions of products.

So, I really do not know what the hell you are talking about.

Of course, all the hype everywhere these days is revolting. That has nothing to do with whether or not it works. How else do you think the PIVs have been outselling Athlons ratios of at least 5:1 for nearly ten years?

Re:Marketing and hubris may have done AMD in.

[RightSaidFred99]

Incorrect. As I clearly explained, my point was that AMD's insistence on "native quad core" may do them in because they can't manufacture the thing. Them losing face is more amusing and gratifying than meaningful.

Re:Marketing and hubris may have done AMD in.

[RightSaidFred99]

If any of what you said was true, you'd have a good point. Unfortunately Intel's chips outperform AMD's in 64-bit as well, just by a slightly smaller margin. What you call "cutting corners" is actually called "making money". A business doesn't release something first for the sake of releasing something first, they do it to make money.

Intel "cut corners" because they know they can get very good yields on the parts that way, hence making money. Now, if this part performed badly then you'd have a point, unfortunately for your non-point Intel chips are outperforming AMD chips in just about every segment. The only exception is in the 4S+ domain where AMD's HT does give them an edge. Unfortunately for AMD, Intel is releasing chipsets with per-socket FSB's.

So explain how releasing a faster product that costs less to manufacture is somehow a negative in some bizarro world? I know, FSB isn't pretty! It's ugly! Nobody cares. And CSI is just around the corner, I wonder what people will find to complain about then.

In short, the statement "the technology itself is not superior to what AMD is offering" is provably and undeniably false. It performs better and it's cheaper to produce. Ergo, it's better technology.

Re:Marketing and hubris may have done AMD in.

[Kjella]

I think that without AMD, Intel itself will actually go down in a slump. For most consumers faster CPUs are now "nice to have". Throw in the cheapest dual-core processor you can find and 2GB RAM (which has also come down a lot in price) and you got a very cheap yet powerful and responsive machine that'll run pretty much everything for a non-gamer. Throw in a video card like the GF8500 with full H.264 acceleration if you want 1080p HDDVD/Blu-Ray playback. Sure, it'll gradually creep upwards but I think the glory days of desktop CPUs where people will eat up that 10% increase in CPU power is over. Right now there's competition, there's agressive pricing, there's innovation and very powerful chips are pushed to the market. I think if AMD failed, Intel would also fail to "compete" with its current generation chips, people would simply continue to stay with the hardware they have.

Re:Marketing and hubris may have done AMD in.

[SacredByte]

You assume that Technologically superior products are always more successful than their inferiors, and that AMD's multiple core solution is in fact, inferior.

for my first point, I would remind you of BetaMax, and Windows; BetaMax had technological superiority than its competition, but lost (Largely because of cost & capacity) To VHS; whereas Windows was vastly inferior to its competitors yet somehow managed to end up with the largest market share.

As to my second point, AMD's design seems to have some advantages over Intel's, most notably in how it manages the cache; With Intel's quad core, to move data from one core's cache to another's, it has to go out the processor's bus, and back in again, thus creating an unnecessary bottleneck on performance. In AMD's quad core it doesn't have to move the data through the bus for it to be accessible by another core. I'm not sure how useful this is in real world applications, but sure looks good on paper.

Re:Marketing and hubris may have done AMD in.

[niceone]

The chips will perform the way they perform. There will be benchmarks. People will buy based on cost vs. performance decision-making, not cost vs. hype decision-making.

I have noticed in the audio world (and I'm guessing in other areas too) it only works like this when Intel is ahead. When AMD is ahead a large number of people carry on buying Intel because it is Intel. It sucks to be AMD.

Re:Marketing and hubris may have done AMD in.

[turgid]

You, and all the other intel staffers, have my condolences. itanic was destined to hit the iceberg the moment it set sail, and now you've been out-engineered by a competitor who was once merely a minor nuisance.

Re:Marketing and hubris may have done AMD in.

[RightSaidFred99]

I wondered when some anonymous jerkoff would catch this. Congratulations, you have officially won the jerkoff of the month award, you caught a typo!

Re:Marketing and hubris may have done AMD in.

[RightSaidFred99]

AMD's solution is provably inferior. It costs more to manufacture, has poorer yields, and doesn't perform as well. In fact, VHS was provably superior to Beta contrary to the urban legend. It had more capacity and hence provided more value to the customers to who it was marketed. You seem to confuse some kind of "engineering superiority in a few narrowly defined categories" for "meaningful technical superiority". The two are orthogonal.

Re:Marketing and hubris may have done AMD in.

[RightSaidFred99]

Ooh, zing! Unfortunately, your baseless arrogance has no basis in reality. AMD is sinking like a rotting turd, and good riddance. They used to be a good engineering company, but they got a little too whiny and complacent and now they're suffering a much deserved smackdown. It's amusing that you would bring up "itanic", which nobody but you cares about one way or the other, to counter AMD's sinking ship.

Have a nice trip back down to $12.

Re:Marketing and hubris may have done AMD in.

[Vancorps]

I have no idea where you get your information. On the 64 bit end of things the Opteron has been hammering the Xeon for years. This is the area I have experience as I run a website which sees millions of users. The Opterons in addition to being more stable also perform a hell of a lot better. The numbers add up too as integer performance has been a strength of AMD for quite some time. Maybe in the gaming arena Intel's 64bit offerings perform faster but certainly not in the real world with IIS/SQL 2005 and Apache/mySQL. I run both environments and both are consistent.

Intel's shared FSB has been a joke for years, I have no idea why you would defend it as being sensible when Intel could do it right from the get-go and not have to redesign the chip later costing them more money. It's all about short term gains which you seem to think is a good idea. AMD made enough money to buy ATI I don't think they are exactly hard up given that their processors are back to being cheaper and performing in line with Intel. The Core 2 Duo is more expensive than the equivalent Athlon. The only place where there is real competition is at the top end of the consumer chip market where Intel holds the performance crown. For the price AMD is clearly a better value which is why more and more OEMs are releasing AMD products. Think Dell, Think Sun

You're statements are pretty much bizarre considering the Core 2 Duo and the latest Athlon X2 products or the Athlon FX series processors are very competitive in 32 bit so there is no smaller margin to make up in the 64 bit otherwise the processors would be performing identically in which case you've already proved the point of why AMD is the smart move.

AMD's most expensive desktop processor? $800 Throw that money at Intel and you will get a processor that is slower than the Athlon FX. If you want to spend $1300 on a chip then sure Intel will be faster, a whole 10fps faster in your game. That's definitely worth the extra $500.

Also, why do you think MS went with an Athlon processor for their low cost PC they are selling in India? Perhaps because AMD is squarely in the best performance for the dollar? That strategy has always worked for AMD. They ran into trouble when they dropped it for a while and raised their prices.

Re:Marketing and hubris may have done AMD in.

[t_ban]

Looks like you're mocking the outcome of a future event that has not happened yet.

that's right. if you must mock the outcome of future events, at least select those that have already happened. :-)

Re:Marketing and hubris may have done AMD in.

[turgid]

My other point was that AMD has mocked the "glue 'em together" approach.

And rightly too. AMD has a proper chip-to-chip interconnect (Hypertransport), which makes the "glue 'em together" approach unnecessary and undesirable. On an AMD system, you just plug in another CPU. On an intel system, the font side bus architecture makes this approach quite ineffective.

Never mind, intel will be bringing out its Hypertransport competitor in 2008, a full 5 years behind AMD.

65nm?

[beavis88]

AMD is only just now moving to a 65-nanometer process

That's a nice thought, except it's totally wrong. All their Brisbane core X2 chips are on 65nm now, and have been for quite awhile.

Re:65nm?

[samkass]

AMD has been shipping 65nm CPUs of one kind or another for about 6 months now. However, the Athlon X2 line still has many 90nm parts in its lineup-- they're still in the process of moving to a 65nm process, as the comment notes. So "totally wrong" is probably less correct than the original statement.

Re:65nm?

[zensonic]

True. But that is most likely a production volume issue. AMD can not afford to turn out new FABs at the same rate as Intel can. Sad but true. They can only win this game by thinking smarter not harder!

Re:65nm?

[John Jamieson]

Huh? Intel still sell processors fab'd at larger sizes as well ... Does that mean Intel is also still in the process of moving to 65nm???

Nope, "totally wrong" is probably more correct than the original statement.

Re:65nm?

[OrangeTide]

My 65nm X2 was dirt cheap and I bought it earlier this year. Good job AMD, I'm another satisfied customer.

Not ruling AMD out

[tjstork]

Moving to native quad core has a lot of advantages and I'm actually excited to see how well this CPU will perform. Critics that claim that AMD lags behind in the process size would do well to note that AMD has ALWAYS lagged behind Intel in that category, and, yet, has managed to not only survive, but prosper.

Re:Not ruling AMD out

[2nd Post!]

Releasing a product in a timely also has a lot of advantages
Releasing a performance commanding product also has a lot of advantages

Currently Intel is king of the hill on both counts. If AMD has a 2GHz quad core, Intel could match with a 4GHz quad core ON THE SAME DAY.

It is very hard to believe that AMD will be able to field a 2GHz part with 2x the performance of Intel's chip... that is what would be required to show a 40% performance advantage.

Re:Not ruling AMD out

[Anarke_Incarnate]

You are erroneously assuming that clock frequency matters across different designs. If I were to tell you I could sell you a working chip at 10Ghz would you want it? What if I told you all it had was an And gate and an OR gate and nothing else? How would that help you? IPC (instruction per clock), Bandwidth, Interprocessor latency, Thermal efficiency and Power draw all matter in the enterprise space.

Re:Not ruling AMD out

[cartman]

Moving to native quad core has a lot of advantages and I'm actually excited to see how well this CPU will perform. Critics that claim that AMD lags behind in the process size would do well to note that AMD has ALWAYS lagged behind Intel in that category, and, yet, has managed to not only survive, but prosper.

AMD has always lagged behind in process technology, however they've usually only lagged behind by a few months. Now, however, the lag is more significant since Intel is moving to 45nm soon, while AMD is still in the transition to 65nm. I can't remember a time when AMD was nearly a full process generation behind.

..AMD has survived, true, but it hasn't prospered. AMD's split-adjusted stock price is about the same as it was in 1985. And AMD has taken significant losses in a great many of the intervening years.

When AMD has prospered, it usually was because Intel management had made some colossal strategic mistake and AMD exploited it. For example, Intel management decided not to design a successor core to the PPro/PII/PIII until AMD had released the Athlon, because of their confidence in Itanium. And Intel strenuously resisted going to 64 bit on x86, again to protect Itanium. And Intel delayed multicore processors. In all of those areas, AMD was able to beat Intel to the punch, not for technical reasons, but because the people who run Intel made strategic mistakes in direction, over and over again.

However Intel can bring colossal resources to bear, which matters because making CPUs is the most capital-intensive industry in the world. Intel has tremendous innate advantages because of their economy of scale and easy access to capital. Whenever AMD gains an advantage, Intel stops doing whatever stupid thing they were doing and re-commits themselves to beating AMD at the x86 game. When Intel isn't on the wrong path and isn't making silly mistakes in strategy, they almost always beat AMD and force AMD into heavy losses.

This time, Intel doesn't appear to be making any silly mistakes, which is terrible for AMD. Not that I think AMD will go bankrupt anytime soon, but I suspect AMD will have a few "lean" years, like they did when they were selling K6's.

Re:Not ruling AMD out

[2nd Post!]

Unfortunately for AMD I think Intel currently has the crown in IPC, thermal efficiency, and power draw :)

Re:Not ruling AMD out

[QuoteMstr]

Okay, for the sake of argument, assume the OP meant a CPU with one instruction, the mythical decrement-and-branch-if-zero. I'd rather have a 2ghz x86 chip than a 10ghz single-instruction processor. :-)

Re:Not ruling AMD out

[Anarke_Incarnate]

but when scaling out to 4+ sockets Intel's memory bandwidth and latency goes to shit

Re:Not ruling AMD out

[Crazy Eight]

I wish there were some wiggle room in your assesment for equality in the chip wars, but I think you've justly captured the scenario. Intel need only refine C2 and avoid screwing up. AMD will play second fiddle for some time to come.

Re:Not ruling AMD out

[2nd Post!]

Too bad for AMD that is such a small, niche, market.

AMD needs something for laptops, the hottest market now, in order to regain profitability.

Oh here it comes!

[Seumas]

Wow, it'll be 6 or 7 usable ghz of computing power! WOW! We're a whole 1.4% of the way to that 500ghz CPU power that we were talking about in yesterdays article! Can you smell it? The future is coming!

*snicker*

I hope...

[Mad_Rain]

I hope they manage to avoid the problems that hit Intel's Core 2 Duo, and squash some of their own bugs while they're at it. Or is it too late to handle those problems at this stage in the manufacturing process?

Re:I hope...

[Bacon Bits]

Of course it won't be. Do people honestly think this Core 2 Duo thing is new and unique to Intel?

What the heck do people think BIOS updates are?

Re:memory bandwidth?

[Joe The Dragon]

Intel maxes out at 4 cpus and there plan to have 4 FSB in there upcoming 4 wat xeon chip set and CSI has been pushed back to 2008

Not for everybody

[Tribbin]

Games can be specialized to use 4 or more cores.
Servers will really use it.

Mr. PC enthousiast who likes to rip DVDs and do other things in the meanwhile can do with 2 cores.

I'm a multitasker who converts audio and video and downloads a lot while intensively browsing the internet. I see no need for me to go more than dualcore. If you are like me; better yet use the money on more happy HD-space, quiet cooling and memory.

Re:Not for everybody

[Seumas]

And at this point, I'm not aware of any game other than Supreme Commander that will actually make use of quad core (and boy does it need it).

Re:Not for everybody

[aegl]

Mr. PC enthousiast who likes to rip DVDs and do other things in the meanwhile can do with 2 cores.

Isn't DVD ripping (video transcoding) just a perfect application for 4 (or more) cores? CPU intensive, trivially parallelizable, takes long enough now that user will notice and appreciate a 4x (or more) speedup.

Re:Not for everybody

[Slack3r78]

Is this a joke? Video encoding is a task that's relatively trivial to parallelize and is one of the more common tasks that receives pretty massive speed ups from it.

The average DVD ripping guy is going to see more benefit from a quad core at this point in time.

Re:Not for everybody

[TheRealMindChild]

Quake 3 was multi-threaded a lot, but in games such as it, I am not so sure that having enemy's and crates and whatever else on their own thread is such a good idea. A primitive, yet effective way to approach such a game can be illustrated with such:

void Gameloop()
{<br>
...
Tick(); //Game engine moves in time one increment

for (Index0 = 0; Index0 < World.Objects.Count; Index+=1)
{
World.Objects(Index0).React(); //Each object reacts to happenings in the previous moment
}
...
}

To put it simply, the world ticks by one moment at a time, and each object in turn gets to react to the happenings in said moment.

Using this model, multithreading only makes sense in that your music can play on one thread, input on another, and the game engine on another. But the core fundamentally makes more sense illustrated above, to me.

Re:Not for everybody

[Tribbin]

The new flight simulator by Microsoft will support multicore for scenery rendering and stuff.

Re:Not for everybody

[wall0159]

Um... ripping video and audio is ideally suited to multi threading, and hence multi-core.

It's just a matter of parallelising software, and once it's done, the more cores the better.

Re:Not for everybody

[MojoStan]

the bottleneck is the dvd read/seek speed and IDE/ATAPI's single read channel limitations. The meaning of the term "ripping" (CDs and DVDs) has changed over the years to include ripping and encoding. I personally don't like this definition, but it seems to be the accepted defn nowadays. "Rip the CD to MP3s." "Rip the DVD to XviD."

I agree with aegl. Many video encoding/transcoding apps have been updated to effectively use 4+ cores.

Re:Not for everybody

[Surt]

Even the most cpu intensive games tend to spend very little time in React(), and even if you do, partitioning that for loop and multithreading it tends to be trivially parallelizable.

The real difficulty in multithreading games is multithreading the render phase. That's most of your CPU and GPU time, and it's hard to make parallel on the CPU side.

Re:Not for everybody

[Surt]

Diablo 2 will make use of up to 8 cores. (Not that it will scale linearly, but it will get faster, of course, it being so far out of date these days means that it's already so fast you won't care).

Re:Not for everybody

[Slack3r78]

Out of curiosity, how many video encoding applications support more than one core? The average user probably uses a one click solution to produce their dvds. Most of those tend to support only 1 cpu (are there any that don't).

x264 is definitely multithreaded. A quick bit of Googling would suggest that Nero's encoder, which is one of the more common ones I've seen among the average Joe crodw, is multithreaded, as well.

Nothing but downhill for AMD

[PlugPlover]

It is unfortunate, because the only reason Intel killed off the poorly designed Pentium 4 was the strong competition, and the only reason you can buy a Core 2 Duo for under $150 is the competition from AMD. AMD is showing that it costs them more to make worse chips than Intel - since they have no choice, they are selling them at price points that often make them a great value, especially if you don't need the best performance per watt. But AMD isn't making enough profit to do the development they need to catch up. Note to IBM: buy out AMD and set them on the right path before it is too late. There is great potential in their CPU designs, but without more cash and technology to speed up development to compete, they will die a slow death.

Re:Nothing but downhill for AMD

[Martin Blank]

Also, Intel can't buy AMD (I think they actually tried once, many years ago) due to anti-trust legislation.

PlugPlover never said anything about Intel buying AMD. He suggested that IBM buy AMD, which would be more natural as IBM has collaborated with AMD on numerous aspects that advanced the CPU world, such as SOI and copper interconnects.

Re:Nothing but downhill for AMD

[5pp000]

While AMD has excellent architects, Intel has a commanding process-technology lead, and has some good architects itself. It was only because Intel got complacent and made some stupid strategic decisions (Itanic) that they let AMD back in the game. We all owe AMD thanks for forcing Intel to wake up... but now that they have, I don't know how much chance AMD has of keeping up.

Also, read again. The GP was proposing the IBM buy AMD, not Intel. I think it would be a good thing too. Well, except that the company cultures are utterly different :(

Barcelona? Phenom? What's the diff?

[Thagg]

I can understand that there would be a difference between chips designed for multiprocessing, but if I wanted to have a single-chip workstation, would there be any difference between the Barcelona and the Phenom?

I'm interested in building new workstations for my company, and the Phenom chips look great except that they don't exist, and won't, for at least six months. Why not build Barcelona workstations, though?

Thad Beier

NPT

[Natales]

I don't think the battle is only on the 45 vs. 65nm arena. There are other interesting technologies in the package that deserve some consideration. Barcelona will include Nested Page Tables (NPT) technology, which could potentially give a significant performance boost to memory intensive applications running on virtual machines once the hypervisors start supporting it.

Intel will also be coming out with a similar technology called Extended Page Tables or EPT, but AFAIK their timeframe is early 2008.

Not as simple as 65 v 45 nm.

[Junta]

AMD is using SOI, which as I understand, produces some benefit over Intels process at a given process size. If the benefit offsets 65 v 45, or comes close, I don't know.

I do know that AMD sales have persisted in high-performance applications, simply because AMD's memory and IO architecture remain *much* metter than Intel's. Intel with Core 2 finally seriously had a competitor in terms of performance (i.e. very good floating point), but it will be interesting if Barcelona essentially matches the performance in terms of floating point and such, but maintains (or widens) the memory performance, it could be an interesting time for AMD once more.

Re:Since when...

[goarilla]

moderators seem to be very harsh today
this should be moderated funny imho (the tomatoes part made it funny)

on another note i hope this chip is killer because it deserves to be !
i saw this chart which iirc eventually meant every core on the amd can be controlled
frequency wise and vcore wise although i'm not sure about the latter

Re:Cores and process are nice, but what about SSE(

[tknd]

The problem is that Intel will not implement anything AMD comes up with unless AMD can convince enough software companies to switch their products to the AMD developed standards. For example way back, they created 3dnow! instructions and had minimal success even though there were some obvious performance benefits. Intel never implemented 3dnow and instead went with SSE. It's pretty obvious who won that battle. Now AMD did have a win with AMD64/x86-64, but progress there has been slow and Intel still doesn't want to admit defeat by relabeling the supported instructions as EMT64. If AMD wants to truly win the instructions standards battle, they would have to focus more effort on gaining market share and software partnerships first.

Another thing to note is that while they do typically lag behind one version of SSE, they did implement a number of improvements to SSE instruction performance in barcelona. Hopefully these improvements will shorten the gap on SSE heavy benchmarks (optimized encoding software, games).

Re:memory bandwidth?

[Anonymous Coward]

Look out Slashdotland, it's the incredible edible run-on sentence!

Re:Barcelona? Phenom? What's the diff?

[Chandon Seldon]

The early Barcelonas will be designed for dual socket servers, and they'll be released at reasonably low clock speeds. If you want to make an 8 core workstation that isn't super fast on single threaded tasks, then they'll be a great deal. If you want a single socket system, you'll be spending more money for lower speeds than if you waited for the Phenom processors.

Re:Cores and process are nice, but what about SSE(

[Chandon Seldon]

When have you ever seen a game benchmark changed by the difference between SSE2 support and SSE3 support? From what I've seen, most game developers don't even consider using the new SSE instructions for a couple years - both waiting for AMD to support them and waiting for people to replace the vast majority of older Intel machines that don't support them.

Even when chips do support SSE type instructions, they rarely produce as drastic a performance improvement as the chip manufacturers hype would imply. Writing a program for SIMD parallelism (like SSE) is just as hard as writing it for a multi-core processor, and works for far fewer workloads.

Re:Cores and process are nice, but what about SSE(

[larry bagina]

The only OS support for SSE(x) is saving/restoring the registers while doing a context switch.

Actually, yes, Intel does create these randomly.

[WoTG]

AMD and Intel have cross-licensing deals that handle the instructions that each company creates... these deals go way back to the mists of x86 time. So, for AMD to implement SSEn there is no legal problem. Ditto for the reverse. Except, the "problem" is that Intel w/80% of the market can pretty much dictate what instructions will survive in the market -- with the big exception of x86-64, and potentially some of the new virtualization stuff.

Now, about releasing chips in a timely manner... the trick is Intel doesn't have to tell anyone about the new instructions until they are well on their way to being in Intel CPUs. AMD finds out about these things at the same time as software developers get the promotional material from Intel. There's no way for AMD to release chips with these functions at the same time as Intel - they have to wait until the next moderate chip revision.

Does it matter? Usually not. Most software lags instruction changes by years. The exceptions are typically where performance really counts. For example, video encoders picked up on SSE2 pretty quick, since it provided dramatic improvements for their code.

Re:memory bandwidth?

[afidel]

Intel doesn't max out at 4 CPU's, at least not for systems you can actually buy. Since the memory bandwidth isn't there the most you can buy is dual quad core (this is from IBM, HP and Dell). Oh and they max out at 32GB of ram whereas you can get a DL585g2 which can economically go to 64GB and maxes at 128GB with 4GB chips. HP and AMD have committed to supporting Barcelona on the DL585g2 so I expect I will have some 16core 64Gb machines by the end of the year =) Oh and if you are into Sun the x4600 will be able to go to 16cores and 256GB of ram once Barcelona ships.

Re:Cores and process are nice, but what about SSE(

[ChrisMaple]

I've written video processing code using SSE, and it makes a substantial speed difference. I haven't looked at the numbers recently, but my recollection is about a 30% reduction in run time. Beyond having to learn more assembly language instructions, writing for SIMD is not terribly difficult. Writing for a multicore processor requires learning about threads and paying attention to data timing, which I find quite difficult.

Re:memory bandwidth?

[Wavicle]

Oh and they max out at 32GB of ram whereas you can get a DL585g2 which can economically go to 64GB

Gosh, I sure wish HP would put an Intel processor in the DL580g4. That way there'd be a platform that would max out at 64GB of ram. Oh, wait...

Re:Cores and process are nice, but what about SSE(

[Chandon Seldon]

Yea, video processing code is the poster child for SIMD - a 30% runtime improvement there over no SIMD is quite reasonable. On the other hand, the difference between SSE2 and SSE3 for the same code is probably somewhat smaller. In other areas, SIMD doesn't help at all, or only helps if you use techniques that are much more complex than multi-threading code.

Another interesting development is GPUs as general purpose SIMD processors...

Re:Actually, yes, Intel does create these randomly

[Hal_Porter]

AMD and Intel have cross-licensing deals that handle the instructions that each company creates... these deals go way back to the mists of x86 time. So, for AMD to implement SSEn there is no legal problem. Ditto for the reverse.

There is a diffrence. AMD has to pay to license Intel patents...

http://news.com.com/2100-1040-257059.html

A source familiar with the deal said it is essentially similar to the last one, which calls for Intel to receive royalties from AMD. Intel has patents covering aspects of the x86 instruction set used in processors for Windows-based PCs.

...but Intel doesn't have to pay to license AMD ones.

http://news.zdnet.co.uk/hardware/0,1000000091,3914 6227,00.htm

The lawsuits started in 1987. Rich Lovgren, former assistant general counsel for AMD, recalled that AMD founder Jerry Sanders sat through "every second" of one of the trials. "There were certainly bridges that were burned," he said.

Under the terms of the settlement, both companies gained free access to each other's patents in a cross-licensing agreement. AMD agreed to pay Intel royalties for making chips based on the x86 architecture, said Mulloy, who worked for AMD when the settlement was drafted. Royalties, he added, only go one way. AMD does get to collect royalties from Intel for any patents Intel might adopt.

AMD also agreed not to make any clones of Intel chips, but nothing bars Intel from doing a clone of an AMD chip, Mulloy added.

So everything AMD invents can be adopted by Intel for free. But AMD has to pay license fees to Intel. Pretty indefensible really, since Intel has a much larger market share, and much higher proft margins.

Re:memory bandwidth?

[afidel]

Wow, and the DL580 doesn't support the 5300, so no quad core with large memory support. So unless you are doing one of the small subset of HPC problems that needs pure CPU performance there's very little need. The fact is the Intel systems run out of ram before they run out of CPU power for most real world datacenter needs.

Re:memory bandwidth?

[MojoStan]

Intel maxes out at 4 cpus and there plan to have 4 FSB in there upcoming 4 wat xeon chip set and CSI has been pushed back to 2008 DigiTimes reported yesterday that Tigerton (along with the quad-independent-bus Clarksboro chipset) is scheduled to launch in September (according to sources at server makers). Clarksboro will have four independent point-to-point connections to the CPUs. These new chipset-to-CPU connections are not CSI, but they are supposedly more efficient than the front side bus connections Intel currently uses. The Tech Report has an interesting photo of a Tigerton/Clarksboro demo from October that shows four Tigertons lined up right next to each other with the heatsinks touching (not possible without the new point-to-point connections).

Currently, Intel's Core-based Clovertown is only available for 2-way motherboards (with two independent FSBs). Intel's 4-way boards still use the Netburst architecture and two FSBs for four CPUs. Tigerton and Clarksboro (the Caneland platform) will update Xeon MP to the Core architecture and four point-to-point buses.

Re:Actually, yes, Intel does create these randomly

[Hal_Porter]

I think that's a misprint. It should be "AMD doesn't get to collect royalties from Intel for any patents Intel might adopt."

AMDs 65nm process big let down

[edxwelch]

No one has voiced it yet, but AMD's 65nm process is a failure. It's 65nm parts overclock worse than processors at 90nm process and that's probably why AMD are still producing all there high end parts at 90nm

Re:AMDs 65nm process big let down

[Crazy Eight]

When did overclockablity become a valid metric?

Re:AMDs 65nm process big let down

[RzUpAnmsCwrds]

No one has voiced it yet, but AMD's 65nm process is a failure. It's 65nm parts overclock worse than processors at 90nm process and that's probably why AMD are still producing all there high end parts at 90nm

Then tell me why the Athlon 64 X2 3600+ Brisbane is an overclocker favorite?

My 1.9GHz Athlon 64 is now running at 2.85GHz, limited by the maximum FSB on my cheapshit $50 motherboard. It's overvolted by 0.1v, and it's cooled by a $9 heatsink and powered by a 250W SFX12V 2.0 power supply. I don't have any case fans.

It's 100% stable, at least according to my 2-process 36-hour Prime95 stress test, 6-hour 3DMark06 stress test, and 12-hour Memtest86+ run.

Oh, and I paid $65 for the CPU.

Re:When will Intel and AMD catch *Sun*?

[canuck57]

8 hyperthreading cores running 8 threads each, with each core having 2 ALUs and 1 FPU.

Processing power over cost makes this expensive. Besides, notice how they say "hyperthreading" and threads... this isn't the same as "real" cores. Try loading up those threads and you will watch single thread performance drop like a rock on the T2000. We have this Java app, opens up 1GB RAM and 96 threads of execution. A E240 was more than 1.5 times as fast with 4 cores than the loaded T2000.

I am looking forward to AMDs 4 core...hyper threading is a joke.

Re:When will Intel and AMD catch *Sun*?

[owlstead]

First of all, there are true 8 cores on a T1 (Niagara) chip (T2000 is the server if memory serves me right). There are 4 threads per core, which make sure that the ALU's and FPU's are used to the full. Now, there was some problems with floating point operations and crypto operations because not every core possessed these functionalities. The Rock processor, with 16 "processing units" (4 cores, 4 processing units each), 16 FPU's and crypto-processors will set this right.

Note that these chips are created to deliver a good application server performance, while still showing a strong price point and power/performance ratio. If the data I found it correct, each T1 processor uses something like 56 Watt to do what it has to do. I wonder if your little test took this seriously, using two AMD CPU's running full speed. Also, if anywhere possible, loading the chips with 96 threads will make the scheduler go ape. Better use a thread pool of 8 threads (or 16 or 32) instead. So even if AMD comes out with this 4 core

But I agree on one thing: hyperthreading does very little. I would be just as happy with a T1 using only two threads. Two threads on a single core sometimes can make a 10-20% performance difference. If Intels hyperthreading is any indicator, this difference can go either way. I created a crypto cracking application which run much *slower* when hyperthreading was enabled (using 8 threads out of convenience). If hyperthreading is not (or cannot be) done correctly, it's better to leave it out altogether. It seems that the notion of processing units on the Rock processor targets this issue.

Re:When will Intel and AMD catch *Sun*?

[Wdomburg]

8 hyperthreading cores running 8 threads each, with each core having 2 ALUs and 1 FPU.

That's 64 concurrent threads, 16 ALUs, and 8 FPUs. And probably only needs a 150- or 200-watt power supply. There's a reason why Sun is getting something like $20K per UltraSPARC T1000 or T2000 rack-mount systems and can't keep up with demand...

Um, if you're talking about T1000 and T2000, that's 32 concurrent threads, 8 ALUs and 1 FPU. And the T1000 and T2000 start at $3995 and $9995, respectively. And lead time isn't any worse than their traditional single-core UltraSparc III based systems.

It's an apples to oranges comparison, anyways. Niagara has a wholly different design philosophy and a different set of trade-offs. The T2 cores are far more capable than the T1, but they are still relatively primiative compared to anything Intel or AMD are putting out, latency on common instructions and memory access are (often drastically) slower, and the clock speeds are significantly lower. The high concurrency successfully counter-blances the relatively low speed and high latency in certain workloads, but got spanked in real world tests (including web servers and databases, their target market) in benchmarks.

Apples to oranges. Niagara's design is based on lower speed, relatively primative cores using register switching to allow a lot of "concurrent" threads (most of which aren't actively executing at the time). They went with masking latency instead of reducing it (which, ironically, is pretty well the opposite of the UltraSparc III).

AMD from Intel's eyes

[gosand]

However Intel can bring colossal resources to bear, which matters because making CPUs is the most capital-intensive industry in the world. Intel has tremendous innate advantages because of their economy of scale and easy access to capital. Whenever AMD gains an advantage, Intel stops doing whatever stupid thing they were doing and re-commits themselves to beating AMD at the x86 game. When Intel isn't on the wrong path and isn't making silly mistakes in strategy, they almost always beat AMD and force AMD into heavy losses.

This is exactly right. One of my best friends is an engineering manager at Intel, and he's been there for 9 years now, and he survived the 1000 manager cut recently. A couple of years ago, when AMD was kicking Intel's butt, he said "just wait about 3 years". I thought he was just blowing smoke, but he was right. He recently told me that when AMD burst onto the scene with the Athlon and started trouncing Intel, Intel took it very seriously. There were many meetings about what to do, and they had people do reasearch.

He said they were in a large meeting with some higher-ups, and he saw a presentation that outlined the move to 45nm. He said after that meeting, nobody there was worried at all about AMD because they knew it was a matter of time. They knew AMD didn't have the financial or research resources that Intel had, and THAT is what it takes to get ahead. So they bided their time, did what they needed to do, and trounced them.

He said there was also talk about how brash AMD was, and how the execs would drive their flashy cars around, and be just blowing money right and left. The Intel execs were more subdued because it wasn't new money to them. Tortoise and the hare. There was some back and forth, but it will be a LOT harder for AMD to compete now than it was for them to come up. They woke Intel up, and they aren't stupid. My friend acknowledges, and said everyone at Intel does as well, that AMD coming onto the scene was a great thing for Intel. It made them get off their lazy asses. Not to mention what it has done for the computing industry.

Re:Cores and process are nice, but what about SSE(

[kolbusa]

Well, SSE matters much in HPC code. Intel compiler has quite decent vectorizer which helps a lot. And writing vectorizeable code is not that hard, actually.