Friday, September 15, 2006

K8L desktop chips to launch in 2Q07

See news.

"Very similar picture, all were interested(interested) a theme, already met by way of illustration the changes planned in a kernel, known under code name K8L. Thus, practically all of them (at least, that, anyhow, were sounded and mentioned) should enter already into the first generation четырехъядерных the processors planned to release (to be exact — commodity availability) in 2-nd quarter 2007! "

39 Comments:

Anonymous Anonymous said...

Do you know if K8 will be socket compatible to existing AM2 mobos?

Thanks

3:18 PM, September 15, 2006  
Anonymous Anonymous said...

http://madmodmike.blogspot.com/2006/09/impossible-has-happened.html
http://img232.imageshack.us/img232/193/omgze1.jpg

Too bad MMM can't wait:)

We haven't seen any retail 65nm K8s yet and we're supposed to believe that K8L will come not too long from now? THG's preview of Kentsfield expects K8L to come H207, second half of 07.

4:12 PM, September 15, 2006  
Anonymous Anonymous said...

Why not join up on AMDzone with your username Sharikou:)

4:57 PM, September 15, 2006  
Anonymous Anonymous said...

i see them talking but don't understand anything they said.

5:32 PM, September 15, 2006  
Anonymous Anonymous said...

K8L will fit in AM2 says AMD.

5:40 PM, September 15, 2006  
Anonymous Anonymous said...

Yeah.. like INTEL's launch of COre2 in Q2 and Kentsfield EOY 2006.

Expect them to make it just in time. They will have NO signficant volume till a quarter later. It took INTEL many weeks to get signficant volume with 3 300mm factories, imagine AMD with one little factory doing all their products. THey won't ramp this puppy till they are good and sure they have a real health product/stepping, then it'll take a full quarter to get it out, assembled, tested. Best case schedule now when they have nothing and say Q2 means this is the most optimistic.. IE everything goes perfect. I doubt it. If the A0 stepping is healthy they'd be saying Q4'06. Since they are't saying that you can be sure the A0 is not healthy.

That is the fact!

Somebody said "i see them talking but don't understand anything they said. "

Same thing here

Dude you can get a Dell with an AMD chip.. DUDE its only 399!

5:55 PM, September 15, 2006  
Anonymous Anonymous said...

There were people who didn't believe AMD's profit margin would surpass that of Intel's, but that happened; there were people who didn't believe AMD could buy ATi, but that happened; there were people who didn't believe AMD has the capacity to supply Dell, but that happened.

Some said AMD could not lower its dual-core price due to limited capacity, but AMD did, and when that happened, these people start to blame AMD's previous high prices. Some said AMD could not roll out socket F on time, nor would it make a difference amid Woodcrest; but in fact socket F came up as scheduled and was very well-received.

These people were wrong and have been wrong for a long time; it seems being wrong about AMD is one of their marks and specialties. Yet we keep seeing such people claiming "facts," which are at best their lunatic imaginations.

AMD will start 65nm production by the end of this year and release quad-core "K8L" in 2Q 2007. BTW, that is not fact yet; I'm not like those who always predict facts, and worse, predict them wrongly. But that is the closest we know will happen so far.

6:40 PM, September 15, 2006  
Blogger Sharikou, Ph. D. said...

There were people who didn't believe AMD's profit margin would surpass that of Intel's, but that happened; there were people who didn't believe AMD could buy ATi, but that happened; there were people who didn't believe AMD has the capacity to supply Dell, but that happened.


There are people who don't believe Intel will BK in 2008, but it will happen.

7:03 PM, September 15, 2006  
Anonymous Anonymous said...

"There were people who didn't believe AMD's profit margin would surpass that of Intel's, but that happened;"

"It" instead of "that" would make more grammatical sense. And last time I checked, Intel's still ahead in terms of profit margin(18% vs 7%)

"there were people who didn't believe AMD could buy ATi, but that happened"

Ok, now you're just making up imaginary people that never existed.

"there were people who didn't believe AMD has the capacity to supply Dell, but that happened."

They DIDN'T have the capacity, but now do.

"Some said AMD could not lower its dual-core price due to limited capacity, but AMD did, and when that happened, these people start to blame AMD's previous high prices."

I'm sorry, what? o.o

"Some said AMD could not roll out socket F on time, nor would it make a difference amid Woodcrest; but in fact socket F came up as scheduled and was very well-received."

They didn't make it on time, Core 2 came before their original Q3 target. And last time I checked, Woodcrest is winning in the reviews.
"Yet we keep seeing such people claiming "facts," which are at best their lunatic imaginations."

Please expand on this because I don't understand:)

"AMD will start 65nm production by the end of this year and release quad-core "K8L" in 2Q 2007."

If any K8L will be released as early as Q207, it'll be the Opterons. K8L is looking to be Q3/Q407 on desktop at earliest.

7:27 PM, September 15, 2006  
Blogger pointer said...

sharikou said:
There are people who don't believe Intel will BK in 2008, but it will happen.


didn't you realize that in the same post edward said:

I'm not like those who always predict facts, and worse, predict them wrongly. But that is the closest we know will happen so far.


:)

7:43 PM, September 15, 2006  
Blogger Sharikou, Ph. D. said...

didn't you realize that in the same post edward said:

I'm not like those who always predict facts, and worse, predict them wrongly. But that is the closest we know will happen so far.


I think both Intel and DELL folks must have a lot of respect for me. I told them what will happen even before they could forsee it. You can imagine they were very much annoyed by my projections. However, as events unfolded exactly as I predicted, I look like an Oracle to them. That kind of power to look into the future is formidable. People always wonder what future holds for them, that is especially true for Intel folks. The sense of uncertainty, the sense of insecurity of their jobs, their technology, their survival torment their minds all the time.

For me, it's simple logic. Its like playing chess. Intel folks can see two steps ahead, I can see five. That's why I was right.

A lot of folks debated with me on many issues, such as whether DELL will go AMD. I pinpointed the exact time of that and I gave a full analysis why that is the time. You just have to learn to think with mathematical rigor and consider all relevant factors in the equation.

8:24 PM, September 15, 2006  
Anonymous Anonymous said...

We seen it all happen, we thought it would happen and it did. AMD is growing ever stronger and is a serious threat to Intel.

Doesn't matter what Intel fans don't beleave or wish to beleave. They just can't accept the truth around them. Sad really...

AMD said what they will do and have done it. So why is it so hard to be opened minded about it? I fail to see where they get their logic and views from. They want to beleave AMD can't do things when it can and did. Intel fans just like to lie until they are blue in the face.

Just don't hold your breath! If you can't understand any of the technical parts of anything then why say Intel is better automaticly or this won't happen because Intel is the best. Your perspectives are childish and silly because you can't understand anything they say at all to begin with.

That shows pure 200% Intel fan right there.

8:52 PM, September 15, 2006  
Anonymous Anonymous said...

"And last time I checked, Intel's still ahead in terms of profit margin(18% vs 7%)"

I was talking about gross profit margin, which of AMD was 57% in Q2 2006, higher than Intel's 49%.

"
"there were people who didn't believe AMD could buy ATi, but that happened"

Ok, now you're just making up imaginary people that never existed."


See this page and realize how ignorant you are.

"
"there were people who didn't believe AMD has the capacity to supply Dell, but that happened."

They DIDN'T have the capacity, but now do."


So AMD didn't have the capacity in 2Q, but does in 3Q?

Those people KNEW AMD had a deal with Chartered; they KNEW AMD is ramping fab36 and upgrading fab30; still they said AMD wouldn't have the capacity to supply Dell.

Do you actually agree with them?

"And last time I checked, Woodcrest is winning in the reviews."

Oh, reviews. But is Woodcrest winning the sales? How about high-end server market? How about performance per watt and per price after factoring in the north bridge and FB-DIMM?

"
"Yet we keep seeing such people claiming "facts," which are at best their lunatic imaginations."

Please expand on this because I don't understand:)"


You don't need to understand it; just don't claim facts when you don't have them.

"
"AMD will start 65nm production by the end of this year and release quad-core "K8L" in 2Q 2007."

If any K8L will be released as early as Q207, it'll be the Opterons. K8L is looking to be Q3/Q407 on desktop at earliest."


Actually the official time was mid-2007. Make it 2Q or 3Q as you wish. And it will be the native quad-core with shared L3 cache, double 128-bit SSE, double L1D cache width, and improved branch prediction & OOO memory access.

AMD's processor lines (notebook, desktop, server) are beginning to diverge since Rev.G, which is supposed to be the first 65nm processors by the end of 2006. The name "K8L" will be less well-defined as previous K[5-8].

12:24 AM, September 16, 2006  
Anonymous Anonymous said...

I am sorry to inform that K8L already in October 2006 will be "dead before its arrival" in 2Q07, or DOA, unless its performance will be better than Intel's Core 2 Quadro Kentsfield(see the link below). Do we have any reason to believe that K8L will be better?

http://www.tomshardware.com/2006/09/10/four_cores_on_the_rampage/page10.html

1:11 AM, September 16, 2006  
Anonymous Anonymous said...

Henri Richard, EVP AMD (DigiTimes - june 2006):
"We're now introducing AM2, and of course AM2 will be quad-core compatible.
[...]
And the fact that the AM2 socket allows us to have a seamless transition to quadcore tells me that I can buy a 4×4 platform, in the later part of this year, and then take it to eight cores sometime in 2007. And that's phenomenal.
[...]
Our plan is to have both dual-core and quad-core K8L-based processors within the same timeframe. The K8L development, from the ground up, is a quad-core development."

6:09 AM, September 16, 2006  
Blogger Unknown said...

Speaking of the fanbois... what happen to them... i notice almost total silence for the 3 recent posts you make Sharikou.....

7:51 AM, September 16, 2006  
Anonymous Anonymous said...

"I was talking about gross profit margin, which of AMD was 57% in Q2 2006, higher than Intel's 49%."

Did you say gross profit margin? Either way, Intel still makes more in the end, percentage wise.

http://www.theinquirer.net/default.aspx?article=32234

So the Inquirer counts as 'people'..

"So AMD didn't have the capacity in 2Q, but does in 3Q?"

Not specifically, but they can't just start selling to Dell right away just because they can.

"You don't need to understand it; just don't claim facts when you don't have them."

Yes, because 'lunatic imaginations' are facts -_-

"Actually the official time was mid-2007. Make it 2Q or 3Q as you wish. And it will be the native quad-core with shared L3 cache, double 128-bit SSE, double L1D cache width, and improved branch prediction & OOO memory access."

Mid07=Q2/Q3 Was there a reason you listed K8L's features? I don't think I contradicted that:)

Intel could have 2 socket desktop platforms if they wanted. They don't need it though unlike AMD.

8:20 AM, September 16, 2006  
Blogger Sharikou, Ph. D. said...

Speaking of the fanbois... what happen to them... i notice almost total silence for the 3 recent posts you make Sharikou.....


I was wondering about that too. Suddenly Intelers become very quiet...what happened?

8:29 AM, September 16, 2006  
Blogger Ho Ho said...

"I am sorry to inform that K8L already in October 2006 will be "dead before its arrival" in 2Q07, or DOA, unless its performance will be better than Intel's Core 2 Quadro Kentsfield"

From what I know so far the two will be roughly equal when the clock speeds are similar. The final winner will be the one with higher clock speed.

"Suddenly Intelers become very quiet...what happened?"

I think your posts have became rather boring since there is little new information in them.

9:55 AM, September 16, 2006  
Blogger S said...

"I was wondering about that too. Suddenly Intelers become very quiet...what happened?"

Easy to guess - you are losing your credibility. Your are depending more and more on convoluted arguments than real analysis.

10:22 AM, September 16, 2006  
Anonymous Anonymous said...

An Inquirer article here: AMD quad cores: the whole story unfolded

10:50 AM, September 16, 2006  
Blogger pointer said...

"there were people who didn't believe AMD could buy ATi, but that happened"

Ok, now you're just making up imaginary people that never existed."

See this page and realize how ignorant you are.


Edward, there is a significant difference between AMD 'will' buy' and 'could' buy. There were a lot of people say that AMD 'will not' buy ATI, not because of its ability to buy, but in term of the economical sense and benefit, etc.

10:54 AM, September 16, 2006  
Blogger Altamir Gomes said...

Bruno, Sharikou, methinks K8L makes them kind of remember K7, K8, and the spankfest they've promoted on Intel at their respectives TOA.

History tells that Intel counter'd AMD with Cumine and P4 spec'd to their knees. Core2 for sure will hit its wall, Intel will release overclocked Extreme Editions, however they won't be able to regain the lead for a good time.

Methinks CSI will get outdated before its arrival too, just like FB-DIMM has. Intel's engineers do have a problem in coming to a common sense (in the sense that it should do the job of regaining the technological crown for Intel, but always couldn't - even Core2 is still technologically inferior to K8 in many respects).

10:56 AM, September 16, 2006  
Blogger Christian Jean said...

"Doesn't matter what Intel fans don't beleave or wish to beleave. They just can't accept the truth around them. Sad really..."

Can't blame them really!

So many years of Intel's false claims, propaganda, paper launches, twisted benchmarketing, etc. that they have all come use to subconciencly NOT believing what they are told, read and hear.

But please AMD friends...

Don't point the finger at them. Don't laugh or redicule them for what they say. Don't even hold a grudge for their disbelieve. Cause it's not their fault really. They have a rare desease called FUDOIS (Fear, Uncertainty and Doubt Over Intel Syndrome)

11:44 AM, September 16, 2006  
Anonymous Anonymous said...

Quite?
"I was wondering about that too. Suddenly Intelers become very quiet...what happened"

If you would post all my post I will take more time to educate you flies who hover around Sharkiou's $hit.

Post it all and I'll give you all some tidbits... but you gotta post it PhD pretender.


You know who..

12:06 PM, September 16, 2006  
Anonymous Anonymous said...

"There are people who don't believe Intel will BK in 2008, but it will happen."

Sharikou I assume you mean by end of 2007 according to your previous predictions? Or are you just trying to slowly migrate the prediction?

1:41 PM, September 16, 2006  
Anonymous Anonymous said...

" Anonymous said...

I am sorry to inform that K8L already in October 2006 will be "dead before its arrival" in 2Q07, or DOA, unless its performance will be better than Intel's Core 2 Quadro Kentsfield(see the link below). Do we have any reason to believe that K8L will be better?

http://www.tomshardware.com/2006/09/10/four_cores_on_the_rampage/page10.html

1:11 AM, September 16, 2006 "
}actually, it will have way more oportunity being released at THAT Time than current time, since NO PROGRAM USES QUAD CORES
so its a total WASTE to buy one at these moments, pcs can barely do dual core. In short words.. Intel is desesperately hyping their crap.

and come with bullshit "kentsfield is not for games, its for servers!"
but thats what "Clovertown" will be.

so Intel is shotting himself in the nutz, while hyping his "intelligence"..
honestly AMD as always been on schedule on everything since Hector Took office.
when its not need, they deliver nothing, when its need, they do!.

1:53 PM, September 16, 2006  
Anonymous Anonymous said...

Way too much speculation and not enough facts in this blog. No point in arguing with tech extremist.

2:12 PM, September 16, 2006  
Anonymous Anonymous said...

Edward - I think you are talking about gross margin (don't confuse that with profit margin) and you are ignoring the fact that AMD no longer sells flash products so there margin is determined by only CPU sales where Intel's margins are affected by chipsets and flash sales, both of which are significantly lower margin products than CPU's.

You will also likely start to see a dip in AMD's level due to depreciation for F36 starting to hit the books and obviously both companies will likely be hit in the margin area by the desktop pricing wars.

2:18 PM, September 16, 2006  
Blogger Ho Ho said...

"actually, it will have way more oportunity being released at THAT Time than current time, since NO PROGRAM USES QUAD CORES
so its a total WASTE to buy one at these moments, pcs can barely do dual core. In short words.. Intel is desesperately hyping their crap."

So what do you call 4x4?

2:38 PM, September 16, 2006  
Anonymous Anonymous said...

"Edward - I think you are talking about gross margin (don't confuse that with profit margin) and you are ignoring the fact that AMD no longer sells flash products"

Yes, I knew that. Being a much smaller company with 1/3 to 1/4 market share, AMD has much more operating overhead and thus lower profit margin.

However, even on gross profit margin, many people wouldn't believe AMD could surpass Intel. That was my point. Anyway, as much as I dislike Intel's monopolistic marketing strategies and bad technologies (Netburst, Itanium, FSB, all the way back to x86/286/386), I have never contested that it made more money than AMD (by far).

While I don't know whether Intel would BK as soon as Sharikou suggested, I won't be sorry if it does go Chapter 11. IMO, it will actually be good for the industry.

8:50 PM, September 16, 2006  
Anonymous Anonymous said...

"So what do you call 4x4?"

I don't know how you call 4x4, but officially AMD says 4x4 is for the enthusiast market.

K8L, IIRC, will be mainstream by the end of 2007 on AMD's roadmap (and 4x4 will be applicable to K8L-based processors, too).

See the difference?

8:54 PM, September 16, 2006  
Anonymous Anonymous said...

What about this, is this true ? any comments? I got this from

http://www.tomshardware.com/2006/09/10/four_cores_on_the_rampage/

There is a fascinating element to this combination of multiple processing cores, a fast architecture and even more performance: The new Core 2 Quadro processors are out to beat the pants off everyone else in the x86 field and to extend Intel's lead over the competition. Check out the summit of the processor elite as we compare the upcoming Core 2 Quadro CPUs with Core 2 Duo/Extreme, Pentium Extreme Edition and the AMD Athlon 64 FX.

Are there any limits to the performance frenzy? Even expert opinions are deeply divided, ranging from "more cores are absolutely necessary" to "why do I need something more than my five-year-old PC system?" Although the Core 2 quad-core processors are not expected to hit retail channels before October, Tom's Hardware Guide had the opportunity to examine several Core 2 Quadro models in the test labs. We would like to make it clear that these samples were not provided by Intel.

Basically, Intel packs two Core 2 Duo processors into one package. This poses several questions: How fast are four cores versus the not-exactly-a-weakling dual-core version? What will the heat dissipation and power consumption figures be like in a PC system with a quad core processor? Which applications can truly benefit from the increased core count? Is the current platform still sufficient? And finally, the question real enthusiasts are sure to pose: What can be expected in terms of maximum clock speed?

11:32 PM, September 16, 2006  
Anonymous Anonymous said...

" Kalle said...

"actually, it will have way more oportunity being released at THAT Time than current time, since NO PROGRAM USES QUAD CORES
so its a total WASTE to buy one at these moments, pcs can barely do dual core. In short words.. Intel is desesperately hyping their crap."

So what do you call 4x4?

2:38 PM, September 16, 2006 "

is 4x4 out already?
plz inform me where to buy, thx k?

11:34 PM, September 16, 2006  
Blogger Unknown said...

Sharikou, I think we can safely say all those Intelbois were engineers enjoying their vaction time. That is until the AMD quad core appeared.

Now their bosses are cracking the whips on their backs to get their quad native instead of glued.

12:08 AM, September 17, 2006  
Anonymous Anonymous said...

I think it would have made a lot more sense for AMD to buy Nvidia than ATI. Nvidia makes chipsets for AMD! Nvidia/AMD has been a long time alliance of sorts. I would hate to see Nvidia stop making nforce chipsets and stop doing SLI support for AMD.

I am running a 3700+ on an nVidia chipset in my DFI mobo. I love it!

8:31 AM, September 17, 2006  
Blogger Ho Ho said...

"is 4x4 out already?
plz inform me where to buy, thx k?"

Is Kentsfield out already? If it is, please tell me so I could upgrade my CPU. I don't even need a new motherboard for it :)

From what I know, Kentsfield will be here just before 4x4. They both will give a total of four cores. So my question would be, if Kentsfield is pointless because nothing uses 4 cores then why would anyone need to get 4x4?

Btw, for my little research, every extra core will provide linear performance scaling. No 60%, 80% or 95%. Its practically 100%, even with Intel dual-die solutions.


"I think it would have made a lot more sense for AMD to buy Nvidia than ATI"

Perhaps but NV is roughly twice as expensive as ATI.

1:00 PM, September 17, 2006  
Blogger Sharikou, Ph. D. said...

Perhaps but NV is roughly twice as expensive as ATI.


Buying ATI is no brainer. NVidia is already in AMD camp, buying ATI gets another guy in AMD camp.

1:04 PM, September 17, 2006  
Anonymous Anonymous said...

"So my question would be, if Kentsfield is pointless because nothing uses 4 cores then why would anyone need to get 4x4?"

You conveniently skipped my comments, Kalle. You can plug in 2 K8L into the 4x4 platform and make it 8 cores. Can you do that for Kentsfield?

Kentsfield is Intel trying to catch up Opteron's superior scalability in the server arena. 4x4 is AMD's way to compete with E6700/X6800 in the enthusiast market before K8L is available. They are utterly different.

One more difference: 4x4 lives well into the future after K8L is out, while Kentsfield will obsolete in 1 year after its release. 4x4 is a platform approach, while Kentsfield is just a temporary stopgap.

"Btw, for my little research, every extra core will provide linear performance scaling. No 60%, 80% or 95%. Its practically 100%, even with Intel dual-die solutions."

Then you didn't do your "research" properly. I guarantee you that your claim above will not stand in any academic or professional discussion of computer architecture.

6:07 PM, September 17, 2006  

Post a Comment

<< Home