Friday, April 13, 2007

Athlon 64 X2 3600+ at $57+free shipping

$57, and you get a Pentium XE 965 killer. Geroge Ou is quiet now. He is waiting for Intel to drop the Pentium D950 to $45.

Looks like AMD is determined to kill Intel at all costs. That's the right move. What da heck. Flood the market with 37 million dual cores per quarter, leaving Intel the rest of the market.

I heard that Hector Ruiz has ordered the K10 troops to commence their attack early. The sharp drop of prices of K8 is just the preparation for much faster chips to be launched soon. The K10 will open a performance gap that will never be filled -- as Intel BKs in 2Q08.

46 Comments:

Blogger enumae said...

Sharikou
Flood the market with 37 million dual cores per quarter, leaving Intel the rest of the market.

Why do you keep saying this?

AMD was at 30% Dual core at the end of Q4 2006, how quick do you think they can ramp to all dual core's?

I heard that Hector Ruiz has ordered the K10 troops to commence their attack early.

Do you have a source?

The sharp drop of prices of K8 is just the preparation for much faster chips to be launched soon.

Do you have a source?

The K10 will open a performance gap that will never be filled.

Do you have benchmarks?

---------------------------------

Now, I understand that your post are to draw hits to your Blog, but at least before you actually made an attempt to debate your claims with your visitors.

In hopes of getting you to voice an opinion on relevant topics I have a simple question for you...

Can you explain how AMD will make less money this quarter than last year when they didn't have ATI?

I would like to hear your opinion.

10:37 AM, April 13, 2007  
Blogger Sharikou, Ph. D said...

Can you explain how AMD will make less money this quarter than last year when they didn't have ATI?


There is nothing wrong. AMD was just pulling back before making a devestating punch on Intel. AMD did a huge transition. Now the transition is complete. Going forward, Intel is simply screwed on all fronts.

Those AMD GPUs, chipsets and CPUs plus the huge capacity increase will push Intel into BK by 2Q08.

11:46 AM, April 13, 2007  
Blogger enumae said...

Sharikou
There is nothing wrong.

Thats your answer?

A drop of about 30% YOY when combining AMD and ATI, and you claim there is nothing wrong?

Maybe it was better when you didn't post. :)

12:07 PM, April 13, 2007  
Blogger Sharikou, Ph. D said...

A drop of about 30% YOY when combining AMD and ATI, and you claim there is nothing wrong?


Strategically, you need to look beyond just one quarter. You usually pull back a bit when preparing to launch a major attack. What is brewing within the AMD camp is an all out assault on all fronts. Intel will be simultaneously dealt with crushing blows on IGP, chipset, server, desktop and mobile.

12:23 PM, April 13, 2007  
Blogger Jeach! said...

Hey Sharikou, on a personal note, I was wondering why you now simply 'link' out to the obvious but don't actually publish 'original' articles anymore?

I miss the old articles which were thought provoking (although sometimes scandalous to some), and made me say; "I never thought about that", or "I didn't see it that way".

12:56 PM, April 13, 2007  
Blogger Dr. Yield, PhD, MBA said...

I'm starting to think that our host is a NewEgg employee.

Since March 26, 7 posts have referenced NewEgg prices, 6 with direct links to the site. That's 50% of the last 14 postings.

Is this a blog or a NewEgg ad campaign?

1:49 PM, April 13, 2007  
Blogger Sharikou, Ph. D said...

I miss the old articles which were thought provoking

That's because I found I can always refer to an old article that already pointed to the future, because my old articles were forward looking. I forsaw Intel's 2Q08 BK back in 2006, and that prejection still holds.

Intel's 45nm won't be enough to counter K10, and AMD capacity already ramped up. K10+capacity = death of Intel. AMD's IGP advantage + DX10+Vista is also a factor.

2:10 PM, April 13, 2007  
Blogger lex said...

Sharikou "Ph"ony "D"octorate

He lives up to his name again.

YoY drop of combined AMD/ATI.. its to clear why. Current offerings suck can't compete thus price drops of epic amount resulting in a loss as AMD can't get enough revenue to cover production cost and R&D. Response, cut 500M in R&D and ramp tools to further reduce manufacturing capability on 65nm ramp and slow 45nm development. Thus they will be luck to get 15% of MS in server in 2008 which is the only thing they will make significat profit.

The "Ph"ony "D"octorate got it right Core2 will be 100 bucks and below soon competing with 2-core Barcelona derivatives. Penryn will be selling in the mainstream between 200-600 bucks and we'll have Nehalem going for 1000 bucks. In 2008 Barcelona will be selling for < 300 bucks ala Opetron and AMD will be well on the way to BK.

The more you use BK in association with INTEL the more you confirm your "Ph"ony "D"octorate.

A case can be made for AMD BK, but INTEL there is simply no way they are going BK right now. Anyone who brings enron is simply more stupid the the "Ph"ony "D"octorate.

Only two ways INTEL goes BK
1) Penrym doesn't work. Answer, it already works and needs the standard design tweeks and validation and is on track a few weeks after Barcelona
2) 45nm has a latent problem. Answer, they have had working SRAMS for 2 years and functional x86 CPUs already, see above 1).

How AMD can go BK
1) ASP continue to be pressured as Penrym and Nehalem are as they sound and INTEL is willing to continue to absorb GM < 50%. On AMD's current MS and margins they will be red and out of cash in early 2008
2) AMD cuts the 500-700M they need to have a non-red bank account. Intel continues to execute as in 1) and punish to even lower GM say 45% and are willing to keep profits in the 4-6 billion range.


The only way AMD stays profitable
1) Barcelona is beats both Penrym and Nehalem for the duration of 2008 and commands > 1k in the 4core server.

Sorry whats the likely hood that Barcelona is better then Nehalem in 2008? Not likely... AMD will wither in 2008 like they did in the 1990s with marginal rushed designs and marginal technology.

6:19 PM, April 13, 2007  
Blogger lex said...

Hey "Ph"ony "D"octorate

You know your INTEL laptop fires were more believable and credible at the time of your posting their your current posting.

Can't wait to see the INTEL and AMD results next week and how you spin them BK lines.

6:21 PM, April 13, 2007  
Blogger abinstein said...

"Sorry whats the likely hood that Barcelona is better then Nehalem in 2008?"

It's interesting to see how Intel fanboys believes in Intel's future product like a religion.

Penryn is already farther away (in terms of release date) than Barcelona, and Nehalem even farther away than Penryn. How do you compare Barcelona, which has been tested and benched (although not by you), with Nehalem, which is still on design board?

6:44 PM, April 13, 2007  
Blogger Not Penix said...

So then what would make you so religiously sure that K10 is all it is hyped up to be?

7:02 PM, April 13, 2007  
Blogger Only AMD said...

Unsurprisingly, Intel feels threatened by the trend of GPUs from both AMD and Nvidia moving toward general-purpose computing. In the presentation, Intel also states that processor profit margins are decreasing, whereas GPU profit margins are increasing.

7:25 PM, April 13, 2007  
Blogger lex said...

Dude... Barcelona is a few months away. The dresden factory should be full of the final stepping and the final qual samples should be shipping to OEM to complete full validation studies..

Why are their now bechmarks if this thing is gong to be spectacular?

Why are their no leaks but still the same "must" "some" qualifiers coming from EVERY AMD executive months.


What did you hear from the INTEL executives about Core2 and what was leaked months before launch.

Everyone laughed because they had seen the same thing from INTEL wiht Netbust. But not a thing from AMD because they got nothing.

I don't need to see anythign the actions speak loud.

Where is Hector Challenging Paul again to a street benchmark race.

Remember in 2006 when things were umm bad for INTEL and their Netbust sucked across teh board what did hector do.. Benchmarks were everything, Challenging INTEL to childish manhood test. What happened when Core2 rumored performance came out. Everyone called the lies and more lies.. Guess what C2 was real.

You think a CEO who was running around making all this noise and acting childish about challenging his competitor to benchmarks would sit quiet and tight if he had a huge winner. Does a Leopard change its spots. Hector got nothing

If you can't understand that you are a "Ph"ony "D"octorate like this Sharikou dude.

7:29 PM, April 13, 2007  
Blogger Randy Allen said...

Because AMD said that Barcelona will be 40% faster "in a wide variety of workloads"! This is the truth! In reality it's not going to be much faster than a 3Ghz Clovertown if it is even faster at all. Penryn will easily wipe the floor with Barcelona.

AMD will probably delay Barcelona a few more times, as they have done with R600.

Current rumors have a paper launch of R600 this month, with availability next month. R610 and R630 to follow later on.

7:34 PM, April 13, 2007  
Blogger lex said...

It is good to be threatened and react...
Its good to see INTEL react to any and all AMD moves. That insures for everything AMD does INTEL has a counter. INTEL has 8x the fabs, and 10x the cash and 100x the profits. What ever AMD does or NVidia does INTEL can counter and replicate as an insurance.

Their mistake was their BADD attitude

Now that they Paul and the marketing monkeys got kicked in the nuts INTEL has final got rid of as BADD attidude.
"B"lind "A"rrogance "D"evelopment ""D"isorder, credit to Rahul for this.

Its the AMD fanbois that are BADD and don't realize that Opetron wasn't skill, execution. It was blind luck because INTEL was just so BADD with Northwood, Prescott and their netbust self. That is gone. AMD shoud be afraid very afriad as they are history.

7:38 PM, April 13, 2007  
Blogger Tim said...

Funny when I click your link, Sharidouche, it still lists at $65.00. You just pull this stuff out of your ass?

8:25 PM, April 13, 2007  
Blogger Tim said...

"It's interesting to see how Intel fanboys believes in Intel's future product like a religion."

Pot calling the kettle black don't you think? Go back thru the Phony's blog posts and see how long he's been saying K10 (or K8L) will "frag" (every phd I know uses that term, really...) Intel....

8:27 PM, April 13, 2007  
Blogger enumae said...

Sharikou

Here is a PDF I made combining the revenues of AMD and ATI, it is showing Q1 2006 -> Q1 2007, with the projected earnings of Q1 2007.

It seems (if these numbers are correct) Intel may make as much in profits as AMD makes in net sales.

Still think there is nothing wrong?

8:51 PM, April 13, 2007  
Blogger Roborat, Ph. D. said...

Looks like AMD is determined to kill Intel at all costs

what's AMD's plan then? make intel so successful that it gets buried in a big pile of money?
AMD obviously didn't think this through enough. Intel uses banks to keep their money. And believe me, the banks they use can hold $trillions in cash so no amount of AMD giving away business will ever bury Intel under a big pile of money.
AMD needs another plan, like make chips for intel. They'll make loads more money than making those crappy Athlons.

3:40 AM, April 14, 2007  
Blogger R said...

I’m not an expert, but I can read. I’ve noticed the Intel side is in denial. Intel’s recent successes with c2d and Apple was a can of woop-ass on AMD, however the near future of modern computing will undeniably belong to AMD/ATI.

6:17 AM, April 14, 2007  
Blogger Rinu said...

AMD is coming with a bang in 2008. Shanghai will be a complete killer. 8 cores with 6 MB L3 cache is a deadly combination.
http://www.fudzilla.com/index.php?option=com_content&task=view&id=529&Itemid=1

By the time Nehalem comes it will be too late for Intel.

6:54 AM, April 14, 2007  
Blogger abinstein said...

"So then what would make you so religiously sure that K10 is all it is hyped up to be?"

As I said, K10 is already benched. Nehalem is not even out. AMD claims 40% increase of performance in K10, and it's not my job to persuade you. What does Intel claim on Nehalem's performance? Does Intel even know how it would perform?

You see something religiously wrong when the fanboys know more than the manufacturers...

8:18 AM, April 14, 2007  
Blogger Giant said...

So they finally decide to use an MCM? But I thought being a native design was just SO important.

Not that it matters. AMD is losing so much money they'll have to cut down on R&D and spending, they'll have no cash to go down to 45nm.

Think about it. Nvidia said they spent nearly $500m developing the worldest most powerful and advanced GPU, G80. To keep doing designs like that you need money. The way to make money is by booking a profit. AMD is loosing so much money that they simply will not be able to afford to continue R&D in both CPU and GPU.

No one will help AMD. Not even IBM. IBM will wait until AMD's stock is at less than $5 then snatch them up at a bargain price. They will sell the scraps of Ati to Nvidia to recoup some of the costs and then fire all the AMD management and we'll see something decent coming from the AMD engineers.

8:22 AM, April 14, 2007  
Blogger enumae said...

Abinstein
As I said, K10 is already benched.

It is?

Where?

Got a link?

9:06 AM, April 14, 2007  
Blogger abinstein said...

enumae:"It is?

Where?

Got a link?"


Apparently you're not important enough for AMD to let you see K10 benchmarking results. But no need to hurry; you'll see them when the chip's release.

OTOH, Nehalem's performance results do not yet exist in this world.

If you don't understand the difference, you're not worthy of further discussion. Sorry.

9:48 AM, April 14, 2007  
Blogger lex said...

Who is important enough to see the benchmarks?

A 400,000 server + customer? They must not be very important as they have gone INTEL

A itlian VP who then has to qualify benchmarks with "certain benchmarks"

So we know enough...already it is good but not good enough. No Opetron to Netbust killer.

Why is there no Nehalem benchmarks? Because the silicon isn't out yet.

I expect Penrym silicon to be available as soon as OEMs get samples and then we'll hear again how INTEL is king in "MOST" benchmarks..

9:57 AM, April 14, 2007  
Blogger Not Penix said...

So your saying that K10 is to compete with Nehalem, well I was thinking possibly Penyrn would be the direct competition as it will be out at roughly the same time,

10:04 AM, April 14, 2007  
Blogger High-K said...

This comment has been removed by the author.

12:23 PM, April 14, 2007  
Blogger enumae said...

Abinstein
Apparently you're not important enough for AMD to let you see K10 benchmarking results.

So have you seen the benchmarks?

If you don't understand the difference, you're not worthy of further discussion. Sorry.

No, I understand that my question was in regards to Barcelona bench marks, not comparisons to Nehalem or Penryn.

Please try and show a little maturity.

I understand your frustration due to the lack of information pertaining to Barcelona's performance.

12:32 PM, April 14, 2007  
Blogger abinstein said...

"I understand your frustration due to the lack of information pertaining to Barcelona's performance."

my frustration is in talking to a bunch Nehalem/Penryn-religious Intel fanboys, and lunatic comments like quoted above.

12:59 PM, April 14, 2007  
Blogger enumae said...

Abinstein

So...Have you seen the benchmarks?

1:04 PM, April 14, 2007  
Blogger lex said...

Where are the damm bencmarks..

Why does every talk about them but now body will quote a number for any single one.. vague comments like 40% faster then Operton or some other similar vague quote.

Why not something like 40% faster at same power for SPECint2006 or someting.

Why, I'll tell you why the numbers suck. AMD still waiting for a last minute design change on a stepping still not out of the fab and validated to fix a huge design bug they don't understand but pray it is fixed on D-step.

That is why there is no benchmark.

1:40 PM, April 14, 2007  
Blogger Tim said...

He's seen AMD's powerpoint presentations and claims. That's it because that's all that's out there.

The fact that we're this close to release and there are *NO* third party benchmarks, leaked or otherwise, should say a lot to anyone with half a brain.

These guys all go by what AMD says - just like sharidouche. When Intel say's their parts outperform by X percentage - it's blasphemy and lies. When AMD does it it's gospel. /rolleyes

1:45 PM, April 14, 2007  
Blogger ililpfxhlu said...

This comment has been removed by the author.

2:02 PM, April 14, 2007  
Blogger abinstein said...

I can't tell you what I saw, and I can't even tell you whether I saw it or not.

If you don't understand, then I'm sorry but as I said you're not worthy of further discussion.

Out.

12:46 AM, April 15, 2007  
Blogger Dahaka said...

enumae you will see the k10 performance when k10 arrives.And you will see k10 killing intel!
You will be upset.I think there are some problem in your brain!:)

Maybe you are hating AMD but AMD will kill intel!Hahaha

2:45 AM, April 15, 2007  
Blogger Giant said...

Kill Intel with what? More Powerpoint presentations that claim the FSB is a bottleneck and that Clovertown is bad because it's not 'native'?

AMD knows that Clovertown is much faster than Barcelona so AMD is deliberately avoiding disclosing any benchmarks.

2:58 AM, April 15, 2007  
Blogger Dahaka said...

Clovertown is not faster than barcelona.You are wrong.Clovertown is faster than k8 but k10 is faster than clovertown.

http://www.spec.org/cpu2000/results/res2006q4/cpu2000-20061016-07636.html

2+2=4 core opteron score is 119

http://www.spec.org/cpu2000/results/res2006q4/cpu2000-20061030-07835.html

4+4=8 core clovertown score is 104

This benchmark results show K10 will realy kill Clovertown.

You will see!
IDIOT!

9:35 AM, April 15, 2007  
Blogger enumae said...

Abinstein
I can't tell you what I saw, and I can't even tell you whether I saw it or not.

Are you under NDA?

I believe you can answer this.

--------------------------

Dahaka
You will be upset.

Why will I be upset?

I have no hate for AMD, competition is good or consumers, and I hope K10 is faster than C2D, and that Penryn is faster than K10, and that K11 is faster than Penryn, see a pattern?

9:39 AM, April 15, 2007  
Blogger Dahaka said...

Enumae I understood you.Maybe you are right.Penryn may faster than k10, k11 may faster penryn...Maybe...

10:22 AM, April 15, 2007  
Blogger abinstein said...

As a matter of fact, K10 won't literally "kill" Intel, which is IMO too big to be "mortally hurt" by any (fair) competition. So Intel lovers should be able to feel safer. But K10 could make a similar difference as K8 4 years ago.

To be sure, there's no way that K10 can be magically faster than Core 2, which already has a decent core design. However, the larger cache in K10 will help (see how poor Conroe-L is compared to Conroe), plus other comparable improvements and the IMC, it's obvoius K10's going to have some performance lead.

This above is just per-core. For multi-processed or multi-threaded apps that utilize more than 2 cores, K10 will trounce Clovertown even more due to the shared L3 cache across all 4 cores.

The question though is whether Penryn, which will be released about a quarter after K10, can get the performance lead back? To answer this we need to look at what is new in Penryn: A smaller transistor size (more cache) and a better power management (less power).

First, for most applications, cache's impact on performance becomes much less past 4MB, so in order for Intel to "justify" the much larger cache size in Penryn it will need to change the apps people use to benchmark the processors. Basically this is a benchmarketing trick that makes people feel better about the processor they buy, even though their apps do not take advantage of the large cache that well.

Second, the better power management in Penryn could mean better performance/watt, but that is excluding the north bridge. A quad-pumped 1333/1600 MHz NB will take at least 15W or more. The "better" power management in Penryn more-or-less compensate its lack of IMC and its use of large cache (more power).

The really interesting competition should be between Nehalem and K10 both at 45nm mid next year. At the moment we simply cannot know how they compare one way or the other.

11:36 AM, April 15, 2007  
Blogger Evil said...

One has to wonder why when AMD lowers prices its a good thing, but when Intel does is all doom saying...

8:46 AM, April 16, 2007  
Blogger Roborat, Ph. D. said...

Evil said...
One has to wonder why when AMD lowers prices its a good thing, but when Intel does is all doom saying...


ofcourse it's not a good thing for AMD. just sit back an enjoy Sharikou's funny methods of spinning horrific news in a positive light.

Ofcourse we all know that a dual core at $57 suggests a massive inventory that needs to be clear.

11:54 PM, April 17, 2007  
Blogger ElMoIsEviL said...

So what happens when the 2Q08 comes along and Intel doesn't go bankrupt? What then?

It's sad that you're not intelligent, but worse that you actually have followers who are even less intelligent.

1:42 AM, April 18, 2007  
Blogger Tanrack said...

The link shows the price to be $65.00. How much does it cost AMD to build a 90nm dual core CPU?

7:27 AM, April 18, 2007  
Blogger Evil said...

Tanrack said...
The link shows the price to be $65.00. How much does it cost AMD to build a 90nm dual core CPU?


I know 18 months ago a 90um chip on a 300mm wafer it was estimated to be costing them about 10-15 dollars. Mind you thats just the core itself, no die packaging... just the raw sand.

8:09 AM, April 18, 2007  

Post a Comment

Links to this post:

Create a Link

<< Home