Inquirer: "AMD's K10 delayed or dead"

[Diasper]

I just came across this article on the inquirer and I was wondering what people made of it.

Is that really likely to be true?


Could AMD be suffering the same delay problems as they did with K8? While they don't have to figure out hypertransport anymore, perhaps with AMD's smaller R&D budget and resources they're suffering?

Or are AMD being so tight lipped about it that no-one will know until shortly before it's released?

Or is the author talking alot of crap (it's Charlie Demarjian if anyone knows more) as he mentions about the

Like the horrendously complex eight issue K9 before it, I would bet good money that it will slip beneath the waves without a hiccup.

When there was no such difficulty with the K9 - in actual fact I thought the single-core AMD64/Opteron was the K8 while I think the AMD X2 was loosely dubbed or at least AMD were thinking of calling it the K9 - or I could be wrong.

Alternatively, what the hell is K8L - sounds like an weak/obscure name that no marketting department would give the go ahead perhaps indicating this is all BS..

Does anyone actually have any real details on the K10?

 

[Hacp]

x2=k8

 

[biostud]

Originally posted by: Hacp
x2=k8

A64 = k8

 

[Diasper]

bump for more thoughts on the topic

 

[Leper Messiah]

The Inquier spreads a lot of FUD, so I take what they say with about a truck load of salt.

 

[Icepick]

Inquirer = Conjecture, extrapolation of the facts mixed with a pound and a half of BS.

I wouldn't put much stock into this article until a reputable site publishes this as fact and cites it's sources. I don't care what the inquirer says. The Inquirer is the National Enquirer for techies.

 

[Stoneburner]

they were right about tejas though weren't they?

 

[Hard Ball]

We need to keep in mind that:

There has not been an official story about the K10 that we have seen in the last year or so;

It's highly unlikely that someone would suddenly have a backdoor connection to AMD internal planning operations to suggest that it is cancelled.

We can't even be sure that there is such a thing as K10 to begin with, let alone it being cancelled. AMD is probably developing something; but since there has been no news on the next-gen chip from any reliable source over a long period of time, I don't think this type of speculation is going to amount to anything.

Yes, they were right about the Tejas, but that was a well established chip under development, while we are not even sure if the code name K10 ever existed.

 

[Hard Ball]

Originally posted by: Diasper
Could AMD be suffering the same delay problems as they did with K8? While they don't have to figure out hypertransport anymore, perhaps with AMD's smaller R&D budget and resources their suffering?

Yes, very possible, we know how chip-makers almost never stick to the timelines that they promise to deliver; there is really no telling of what type of problems developments can bring, and what amount of time it would require to solve those with contingency plans. With AMD's history, you almost have to expect 1-2 year delays. The K8 was supposed to be out to compete with the first Willamettes (not that the Will was a good chip in any way, shape or form), and it ended up on an almost 30 months delay.

Or are AMD being so tight lipped about it that no-one will know until shortly before it's released?

AMD has been very tight lipped about the replacement of K8 and K9 (the code name for DC Opterons and X2s); no one is really even sure what the code name would be. There was an article back in 2003 from some financial journal, that mentioned that AMD was hiring for the development of K9; but since they decided to use K9 for the DC AMD64 chips, no one is really sure what it's called anymore.

Or is the author talking alot of crap (it's Charlie Demarjian if anyone knows more) as he mentions about the

Like the horrendously complex eight issue K9 before it, I would bet good money that it will slip beneath the waves without a hiccup.

When there was no such difficulty with the K9 - in actual fact the AMD64 was the K8 while I think the AMD X2 was loosely dubbed or AMD were thinking of calling it the K9.

The "complex eight issue K9' is complete BS, there are not too many CPUs that have 4 issue wide core (Merom is said to have 4), let alone developing an X86 processor with eight issue wide core. Plus, K9 is know to indicate A64 X2s.

 

[rgreen83]

Just remember, TheInq also said that merom/conroe would be VLIW.

 

[Stoneburner]

That was just conjecture based though. They were theorizing, here they are reporting a source which may not give it complete validity but more so than that article you referenced.

 

[DrMrLordX]

Judging by the contents of the article alone, the Inq author seems to think that a second-generation Merom/Conroe/what have you will be a big thing, and that AMD needs a new core to fight it.

However, based on what I've seen thus far, I don't know that to be true.

The K8 is scaling nicely in many different forms, and it seems to be able to handle multi-core configurations without difficulty. Could it be that AMD is taking their time and/or retooling future cores in light of the fact that K8 variants will keep them competative well into 2008 or later?

Also keep in mind this article:

http://biz.yahoo.com/bw/051101/316192.html?.v=1

AMD may be scrapping core designs in development to incorporate tech they're getting from IBM.

 

[Ronin]

Slightly OT, but Charlie runs The Inq (the guy who posted the article), and he generally spews less FUD than the rest.

 

[Furen]

Well, canceling a CPU project doesnt really mean anything, after all Intel slaughtered half its CPU projects (granted, it probably had more concurrent ones than AMD) and it looks like it will be in a pretty strong position (let us say Intel SHOULD be competitive, though I doubt it will out do AMD) next year. Canceling a project just means that AMD didnt think that its outcome would fit what the market wants, so why weep over something that was found deficient before it even made it out the door?

From the article's own wording it sounds like the K10 either didnt scale too well on multiple cores, meaning that it was probably aimed at better single/dual core performance than the K8/9, or that it used up too much power. This makes sense if you consider that low-power will probably be the focus over the next few years, since this is very helpful if you plan on putting many cores onto a single chip and for notebooks, the market sector that has the highest growth potential.

About the K9 being an 8-issue CPU: It actually sounds very interesting in a way, but I think most of its execution units would probably be as useful as the Cell's 8 SPUs (which is to say not very useful)...

 

[obeseotron]

Maybe AMD is doing a new low power focused core as a basis for future products, or maybe they've decided a 4 core k8 will be enough. Intel can afford to design several different architectures at the same time, AMD can barely afford one so this concerns me a lot.

 

[Aenslead]

I do beleive the story. So far, I have not found a single article that has not come to pass or that is FUD.

Maybe someone can correct me here?

There is an old saying in Mexico that goes: "Si el rio suena, es porque agua lleva", which means "if the river sounds, its because it carries water", which means that there IS a reason for them to write such article.

On the other hand, AMD does not have any problem AT ALL. They are waaaaaaay ahead of Intel in Server, Workstation and Desktop processors. AMD's main goal is to achieve productivity, profits, money, increase market share, AND, besides... by Q2'06, when Intel FIIINAAALLYYY releases their new archuitecture... well.., AMD already has M2. So that's that.

No need to rush on what's not needed.

 

[Diasper]

I wonder if anyone has got the figures for what AMD is spending on R&D as it'd help indicate what sort of level R&D is at - it'd be especially interesting to compare R&D numbers over time say comparing these last few quarters to the year or so before they brought out the K8/Opteron/A64 architecture...

 

[imported_michaelpatrick33]

Originally posted by: Aenslead
I do beleive the story. So far, I have not found a single article that has not come to pass or that is FUD.

Maybe someone can correct me here?

There is an old saying in Mexico that goes: "Si el rio suena, es porque agua lleva", which means "if the river sounds, its because it carries water", which means that there IS a reason for them to write such article.

On the other hand, AMD does not have any problem AT ALL. They are waaaaaaay ahead of Intel in Server, Workstation and Desktop processors. AMD's main goal is to achieve productivity, profits, money, increase market share, AND, besides... by Q2'06, when Intel FIIINAAALLYYY releases their new archuitecture... well.., AMD already has M2. So that's that.

No need to rush on what's not needed.

Are you kidding? The Inquirer gets as much wrong as right. They stated that the X1800XT would have 32pipes then 24pipes then 16pipes just to name one thing they got wrong multiple times.

 

[SynthDude2001]

Originally posted by: michaelpatrick33

Originally posted by: Aenslead
I do beleive the story. So far, I have not found a single article that has not come to pass or that is FUD.

Maybe someone can correct me here?

There is an old saying in Mexico that goes: "Si el rio suena, es porque agua lleva", which means "if the river sounds, its because it carries water", which means that there IS a reason for them to write such article.

On the other hand, AMD does not have any problem AT ALL. They are waaaaaaay ahead of Intel in Server, Workstation and Desktop processors. AMD's main goal is to achieve productivity, profits, money, increase market share, AND, besides... by Q2'06, when Intel FIIINAAALLYYY releases their new archuitecture... well.., AMD already has M2. So that's that.

No need to rush on what's not needed.

Are you kidding? The Inquirer gets as much wrong as right. They stated that the X1800XT would have 32pipes then 24pipes then 16pipes just to name one thing they got wrong multiple times.

Indeed. Another example is their "nForce4 will have Soundstorm 2" claims from about a year ago. They're not wrong all the time, but they're it's important to remember that they're not right all the time either.

 

[Hard Ball]

Originally posted by: SynthDude2001

Originally posted by: michaelpatrick33

Originally posted by: Aenslead
I do beleive the story. So far, I have not found a single article that has not come to pass or that is FUD.

Maybe someone can correct me here?

There is an old saying in Mexico that goes: "Si el rio suena, es porque agua lleva", which means "if the river sounds, its because it carries water", which means that there IS a reason for them to write such article.

On the other hand, AMD does not have any problem AT ALL. They are waaaaaaay ahead of Intel in Server, Workstation and Desktop processors. AMD's main goal is to achieve productivity, profits, money, increase market share, AND, besides... by Q2'06, when Intel FIIINAAALLYYY releases their new archuitecture... well.., AMD already has M2. So that's that.

No need to rush on what's not needed.

Are you kidding? The Inquirer gets as much wrong as right. They stated that the X1800XT would have 32pipes then 24pipes then 16pipes just to name one thing they got wrong multiple times.

Indeed. Another example is their "nForce4 will have Soundstorm 2" claims from about a year ago. They're not wrong all the time, but they're it's important to remember that they're not right all the time either.

Yeah, its like throwing spitballs at the wall, sooner or later, some will stick.

But that doesn't mean speculations based on limited information will always turn into reality.

 

[Greenman]

Even a broken clock is right twice a day.

 

[dmens]

TheInq isn't bad... even the director of Intel EMG recommends it.

 

[stratman]

From wikipedia, which could also be wrong, but I think it's probably closest to the truth:

The AMD K9 generation of CPUs was the codename used to refer to dual core K8 processors. However, AMD has since decided to stop referring to upcoming CPUs by K numbers. While many people thought the K9 was a complete core redesign, that is reserved for what would have been the K10. Some claim the cancellation of the K nomenclature is likely an attempt by AMD to avoid ridicule caused by the similarities between "K9" and "canine", but this is only conjecture, considering K7 and K8 were only internal codenames for the processors that AMD marketed as the Athlon and the Athlon 64, respectively.

 

[n7]

Meh, so much speculation.

I don't really care whether this is true or not.

They will have something good coming out soon enough.

Maybe they'll call it the IK10 (Intel Killer 10) :laugh:

 

[Viditor]

Normally, when Charlie is fairly sure of the story he quotes a source (though this is often "sources inside XYZ company"). In this case, he's not quoting anyone or saying where the info comes from...
That is usually a good indication that he's flying the story and hoping someone will deny or confirm it.