Thursday, December 14, 2006

Patty sounds desperate

You have to be nuts, he said. The cruel reality is more and more Intel exclusive OEMs are switching to AMD and Intel's market is shrinking fast.

Intel will exit 2006 with less than 25% of its products being Core 2, leaving 75% of its products total junk. That's why you don't see AMD in a hurry to bump up clockspeed or release K8L. K8 is more than enough to frag Intel. The $140 Athlon 64 x2 3800+ is faster than 75% of Intel CPUs. The $50 Sempron is faster than 50% of Intel's CPUs.

AMD has started designing a 8 core chip. However, the solution for heavy computing needs will be solved using Torrenza and specialized co-processors. I can forsee an Azul Vega2 chip with Java capability plugging into a Torrenza socket, or a PERL6 chip that is 100x faster in doing RegEx. As AMD64 becomes more and more powerful, and Torrenza able to meet any specialized need. The RISC architectures such as Power are pretty much goners.

By 2Q08, Patty will be jobless.

34 Comments:

Anonymous Anonymous said...

Patty? Yum!

4:54 PM, December 14, 2006  
Anonymous Anonymous said...

Sharikou sounds worried. ;)

5:52 PM, December 14, 2006  
Anonymous Anonymous said...

Yes... er um.. ok. Intel's market share is "shrinking fast" even though they regained market share last quater. That makes a lot of sense eh?

AMD's processors are a joke. It takes a 5000+ 2.6Ghz Athlon X2 to just to beat a 1.86Ghz Core 2 E6300. That's pathetic.

AMD will be bankrupt as soon as my pizza arrives. That should be any minute now.

8:41 PM, December 14, 2006  
Blogger Sharikou, Ph. D. said...

Intel's market share is "shrinking fast" even though they regained market share last quater.

No. Intel lost market share in 3Q06 -- it was just AMD not gaining as fast. Go back and read the news.

But AMD was capacity contrained. Now, AMD has added a lot more capacity. Intel's unit volume will keep shriking, and with ASP keep falling, BK is inevitable for Intel.

9:29 PM, December 14, 2006  
Anonymous Anonymous said...

Sharikou,

I never thought your arguments can get lamer.. I stand corrected

9:29 PM, December 14, 2006  
Anonymous Anonymous said...

No. Intel lost market share in 3Q06 -- it was just AMD not gaining as fast. Go back and read the news.

It seems to me that Intel are bottoming out, or starting to bottom out, because of their transfer to a better product that is selling well and AMD are therefore are topping out, or starting to top out, because of this.

By the end of 1Q07 Intel will have more % of C2D+C2Q and AMD will still have the same range whilst waiting for K8L.

AMD's figures should look good for this period though because of the boost from ATi, though even ATi seems to be losing ground to nvidia at the moment.

AMD has new chips coming out soon 5400+ 5600+ 6000+ it seems but they are still on 90nm so do not make as much money as the 65nm ones even if they are fast and cheap. It will be interesting to see the price of the 5600+ and 6000+ against E6600/6700.

10:18 PM, December 14, 2006  
Blogger Scientia from AMDZone said...

This comment has been removed by the author.

12:11 AM, December 15, 2006  
Anonymous Anonymous said...

If Intel had not counted VIAs cpu sales as there own they would have already dropped out of the number 1 chip sales slot.
Intel did not regain market share, they just slowed down the rate of decline.
While AMD had a 90+% increase in sales intel had a 22% drop off in chip sales.
Intel now has the quad core xeon prices lower than the dual core conroes.
I expect intels prices to drop even further as there sales decline even more rapidly due to the platform problems exposed by there quad core offerings.
AMD stock price jumped nearly $3.00 today due to intel future promises and announcements.
If intel does not correct there platform problems soon only intel fanboys will be buying there chips.
Intels fromt side bus system is just too big of a handy cap when talking quad core processors.
AMD is doing very well thanks to intels continual mis steps.
The conroe seems to be killing intel.

12:43 AM, December 15, 2006  
Anonymous Anonymous said...

Sharikou... you do understand that the 75% P4 based chips are not junks. Lots of people are willing to pay for them, seriously. They do not perform (as) well, but they are not junks.

And to some clueless anonymous...

"AMD's processors are a joke. It takes a 5000+ 2.6Ghz Athlon X2 to just to beat a 1.86Ghz Core 2 E6300. That's pathetic."

You should really say this to Sun or Oracle or all those big and small companies who rely on AMD processors 24/7. Just because your E6300 do SuperPi faster or transcode mp3 and wmv back and forth more quickly doesn't make it a bit more useful or productive.

1:16 AM, December 15, 2006  
Anonymous Anonymous said...

At first I was going to leave a comment regarding something in the post but I'm going to have to add comments from your comments above.

Post Related quote:
"Intel will exit 2006 with less than 25% of its products being Core 2, leaving 75% of its products total junk."
Is this to say that you're finally acknowledging that the Core 2 isn't junk? Hooray! The first step in recovery is admitting you were in denial.

Comment related comment:
"But AMD was capacity contrained. Now, AMD has added a lot more capacity. Intel's unit volume will keep shriking, and with ASP keep falling, BK is inevitable for Intel."
There is still 3-6 months would of loadings to get started before AMD's new fab is running at full capacity, but even then, their still making their money off the 90nm fab and not the 'new to AMD, old to Intel' 65nm factory.

As far as unit volume, ASP falling and the BK, I guess we only have to wait a couple more quarters for you to back peddle.

Tick-Tock-Tick-Tock-Tick-Tock...

1:49 AM, December 15, 2006  
Anonymous Anonymous said...

From the yahoo article:
"market share shifts due to less than optimal product mix from the company in the coming quarters."

Exactly. I think the OEM must be nuts with the current Intel CPU line. If Intel enter the E4xxx and E2xxx before exiting the P4/PD/Celeron lines it will be a complete mess.

2:25 AM, December 15, 2006  
Anonymous Anonymous said...

The cruel reality is more and more Intel exclusive OEMs are switching to AMD and Intel's market is shrinking fast.

Name the OEMs otherwise you are just spouting hot air.

The cruel reality is that AMD were caught flat footed by Intel and are now playing catch up. Their response to core2 has been very poor and they only seem able to compete at the low end and the server environment.


Intel will exit 2006 with less than 25% of its products being Core 2,

Considering that Core2 was officially released 27th July and was a major design change, thats a pretty good roll out.


That's why you don't see AMD in a hurry to bump up clockspeed or release K8L. K8 is more than enough to frag Intel

Then reason that AMD do not bump up clockspeed is that K8 is reaching it's performance ceiling. The FX74 barely overclocks to 3.2 from all the reports I've seen and even the first 65nm are not exactly great overclockers.

K8 is outperformed by an similar priced core2 cpu and is left far behind when overclocking is brought into the equation.


The $140 Athlon 64 x2 3800+ is faster than 75% of Intel CPUs. The $50 Sempron is faster than 50% of Intel's CPUs.

The x2 3800+ is faster than non-core cpus but it's a little depressing that the only place AMD can compete on the desktop is at the very low end of the market. It will also be interesting to see how it fares against the E4xxx line once it's released.

2:58 AM, December 15, 2006  
Blogger Unknown said...

"The x2 3800+ is faster than non-core cpus but it's a little depressing that the only place AMD can compete on the desktop is at the very low end of the market."


So, you agree that the x2 3800+ is faster than 75% of Intel's desktop CPU's. I'd say it's rather depressing that 75% of Intel's desktop chips are at the "very low end of the market". And it's also depressing that AMD's lowest-end cpu can beat 75% of intel's cpu's. Depressing for intel, that is.

Since performance in the core series is very cache based, the new low end core chips with less cache that Intel is planning to release are probably not going to perform very well.

4:18 AM, December 15, 2006  
Blogger Tom said...

Sharikou>>your news about intel only having 25% conroe cales by end December is based on a 3 month old article it is dated sep 18. so I would not really say it is anything new right.

“If Intel had not counted VIAs cpu sales as there own they would have already dropped out of the number 1 chip sales slot.”

As far as I remember VIA had belov 5% of the CPU market and the chips are all priced quite low. The numbers you seems to refer to is revenue so if you sell 1 500$ chip or 10 50$ chips still creates the same revenue.

Intel did not regain market share, they just slowed down the rate of decline.
While AMD had a 90+% increase in sales intel had a 22% drop off in chip sales.


No wrong again they had a increase / drop in revenue that has nothing to do with the actual sales numbers. And please do remember that if AMD had not purchased ATI their increase would only have been 25% (they rose from 4 billion to 7.5 billion but 2.5 billion is ATI revenue so the CPU related increase is 1 billion)

6:50 AM, December 15, 2006  
Anonymous Anonymous said...

Hmmm seems like some old Intel share holder are still in a denial. Reminds me of the Titanic sinking ..where some people believed that the ship is unsinkable..and did not leave it.

BTW this December issue of Fortune Magazine has an unbelivebale Full 4 Pages of logo ad for the Intel Xeon.I cant believe the desperation , or might be an attempt to convince the crookst..those marketing idiots ..what a waste of money and adspace , not to mention the paper in which they are printed on ..

7:27 AM, December 15, 2006  
Anonymous Anonymous said...

"Intel will exit 2006 with less than 25% of its products being Core 2"

The article said "these products will represent less than 25 percent of shipments exiting 2006" and was dated September 06.

Core launch was in july so Intel has only 5 months to sell Core Duo to meed those projections.

That said, I bought some AMD last week.

7:53 AM, December 15, 2006  
Anonymous Anonymous said...

“AMD Live! outselling Intel Viiv by 40%!!!”

This most account for some of the success, Don’t ask for proof, just google it your self.

Did you notice yesterday when AMD price was soaring all the annalists were putting AMD as a hold and today the same folks are pumping Intel, especially BofA.

7:58 AM, December 15, 2006  
Blogger Roborat, Ph.D said...

25% Core2 duo targetted at the the high end high ASP market is more than enough for this segment.
The key is to drive out the competition in this lucrative segment and demand premium!
AMD tried to get back in the party with the 4x4 but we all know how that turned out, LMFAO!

8:57 AM, December 15, 2006  
Blogger PENIX said...

AMD is wise in their forward thinking. The multicore game is an endless game of tug of war with Intel. Specialized cores will allow them to push past Intel yet again. Intel's stone cold grip on the industry is over. Now Intel is just waiting for AMD to make a move so they can quietly copy it and catch up. Pathetic.

10:27 AM, December 15, 2006  
Anonymous Anonymous said...

The Sheepshagger dijo...

At first I was going to leave a comment regarding something in the post but I'm going to have to add comments from your comments above.

Post Related quote:
"Intel will exit 2006 with less than 25% of its products being Core 2, leaving 75% of its products total junk."
Is this to say that you're finally acknowledging that the Core 2 isn't junk? Hooray! The first step in recovery is admitting you were in denial.

Comment related comment:
"But AMD was capacity contrained. Now, AMD has added a lot more capacity. Intel's unit volume will keep shriking, and with ASP keep falling, BK is inevitable for Intel."
There is still 3-6 months would of loadings to get started before AMD's new fab is running at full capacity, but even then, their still making their money off the 90nm fab and not the 'new to AMD, old to Intel' 65nm factory.

As far as unit volume, ASP falling and the BK, I guess we only have to wait a couple more quarters for you to back peddle.

Tick-Tock-Tick-Tock-Tick-Tock...

1:49 AM, December 15, 2006


http://www.theinquirer.net/default.aspx?article=36389

AMD 65 nm parts are already on big OEM companies, and big distributors like DELL and HP.

11:03 AM, December 15, 2006  
Blogger Scientia from AMDZone said...

Well, neither comment is accurate.

Intel's share is shrinking fast: FALSE

Intel regained share in Q3: FALSE

AMD actually took about 0.6% volume share from Intel in Q3. Intel's share of the market compared to AMD's:

Q4 05 - 77.94%
Q1 06 - 77.88%
Q2 06 - 77.14%
Q3 06 - 76.56%

If you want some laughs you can add in VIA's portion which inflates the total volume. Basically, VIA had some x86 compatible chips that were End Of Life and they were dumping these during the 1st and 2nd quarters. Think of this as a clearance sale by VIA. So, if you count the VIA sales as part of the X86 total this artifically create a larger drop for Intel in Q2 which then creates an equally artificial increase in Q3 (when VIA stopped selling these chips). This larger drop and then larger increase can then create the illusion that Intel is rebounding and gaining momentum and this is where the erroneous comments have come from.

However, ignoring VIA shows a pretty steady but slow trend. Intel is slowly losing share to AMD. It is always possible though that this trend could change in Q1 07.

Keep in mind though that with about 75% of the volume share and a 0.6% loss each quarter it would take Intel 6 years to drop down to a mere 60% of the volume share.

11:31 AM, December 15, 2006  
Blogger Scientia from AMDZone said...

My Blogger login ID had stopped working here and I was unable to login. This seems to be related to the Beta version of Blogger and the switch to Google ID's. I was going to explain this in detail but the post started getting long. So, just check my blog if you need more information. I have to do a post there anyway since I've upgraded my blog to the beta version.

11:42 AM, December 15, 2006  
Anonymous Anonymous said...

True, it is a bit depressing that AMD is king of the low end. That is where the money is, though, and it pays nicely for ccHT, Torrenza and other developments that seem to have a real future.

Meanwhile, kudos to Intel for making the core architecture perform the way it does, but how much is it doing for them?
Anytime any company spends this much time discussing the competition, you know they are in serious trouble.
Marketing 101

Lucky for us, the race is back on.
Speaking of which, has anybody seen a test of a NUMA-aware linux OS on Quad FX?

11:56 AM, December 15, 2006  
Anonymous Anonymous said...

Just for perspective...

This is just a primative way in which to spin the numbers, and has no real bearing, this also shows the Dr is just pulling numbers out of the air.

Assume the microprocessor market makes 100 processors.

AMD has 25.

Intel has 75.

Of that 75, one quarter are Core based.

This would equate to 18.75 of the 75 being Core based.

18.75 * 100 / 25 = 75%

Using Sharikou logic, Intel with its 25% of production being Core, it is better than 75% of AMD's total production capacity.

In another light, of the 75 not being core based would leave 56.25, or the other 75%.

6.25 * 100 / 56.25 = 11%

Which would mean AMD could FRAG 11% of Intels production.

Which in turn leaves (56.25 - 6.25) 50, or about 37% of Intel products without competition from AMD.
I am not trying to say anything negative about AMD, just Sharikou's numbers.

12:49 PM, December 15, 2006  
Anonymous Anonymous said...

"No. Intel lost market share in 3Q06 -- it was just AMD not gaining as fast. Go back and read the news"

Actually, you financial-challenged, biased individual - both gained market share last quarter (at the expense of via). But why let the facts get in the way...

5:17 PM, December 15, 2006  
Anonymous Anonymous said...

Only two things matter right now

1) Benchmarks? INTEL wins them all

2) Capability to ship CPUs with those benchmarks. LOL INTEL owns that too!

6 months ago that wasn't true. AMD owned the benchmarks but not the ability to ship volume. So without it they were stuck in the 20%... Too bad Hector and his merry men didn't have the balls to mortgage their future in 2001 and build the big 90nm factories that would have gotten them 50% MS by 2006 if they had balls. NOW they mortgage the future on the disaster strategy of platforms..

What a band of morons

5:30 PM, December 15, 2006  
Anonymous Anonymous said...

LOL the PhD spins and spins and only makes himself dizzy..

The bottom line, INTEL rules in profits, market share, money to roll back into new designs and new process technology.

Their silicion technology is a year ahead, they have 3x more factorys. That scale means they can invest more to be ahead.

Tick Tock Tick Tock AMD is finished no more glory, no more benchmarks, no more.. Poor Hector

I'm all eye's to read how AMD can win. If both AMD and INTEL execute to their capability AMD will lose MS and profits will evaporate. They will keep their 20-25% but thats it.. end of the golden two years.. Poor Hector, poor Sharikou, poor amd fanboys.. LOL

Convince me otherwise!!!

5:43 PM, December 15, 2006  
Anonymous Anonymous said...

"Intel's share is shrinking fast: FALSE

Intel regained share in Q3: FALSE

AMD actually took about 0.6% volume share from Intel in Q3. Intel's share of the market compared to AMD's:"

You're point is taken but in terms of MARKET SHARE it is the TOTAL MARKET, not only that market composed of AMD/Intel. Technically (and actually) Intel GAINED market share in Q3'06 (the numbers are the numbers). Did this come from AMD or VIA?

Did AMD eat VIA's market share and then Intel took AMD's? I don't know. Did Intel directly take the share from Via? I also don't know - in fact noone can tell.

Scientia you're numbers are interesting but you're inherent assumption is Via's segment is split between AMD/Intel like the rest of the entire X86 market. Do you know this to be fact? Does anybody? What if all of Via's sales were in AMD's market or Intel's market?

Add in the fact that the last 4 quarters are less than 1% (absolute change). Just out of curiosity what is the margin of error in the market share #'s? If you graph those numbers on a scale from 0-100% it would look pretty flat to me...

6:09 PM, December 15, 2006  
Anonymous Anonymous said...

The one (of many) problem with all of tour BK and market share analysis - is by the time AMD addresses their capcity issues, Intel will be able to ship much higher %'s of Core 2?

Most of the P4's will be EOL'd by Q2'07... when again was F30 going to be fully converted to 300mm/65nm again. Heck isn't this just about the time when F36 gets fully converted to 65nm.

You just don;t get it - by the time AMD gets their capacity in order, they won't be competing with 75% P4 product mix...

By the way any comments on 65nm X2 die size vs 90nm die size? The early #'s I saw showed only ~33% reduction, not the 50% you were projecting...

6:21 PM, December 15, 2006  
Anonymous Anonymous said...

Smoking what??

"True, it is a bit depressing that AMD is king of the low end. That is where the money is, though, and it pays nicely for ccHT, Torrenza and other developments that seem to have a real future."

Low end is the loss maker - and here is where AMD plays

2:34 PM, December 16, 2006  
Anonymous Anonymous said...

Tick-Tock-Tick-Tock Patty has just learned to read the clock!
Tick-Tock-Tick-Tock Patty realises he can't deliver a new chip round every block!
Tick-Tock-Tick-Tock everyone knows Intel's still a flop!

7:53 AM, December 17, 2006  
Anonymous Anonymous said...

You should really say this to Sun or Oracle or all those big and small companies who rely on AMD processors 24/7. Just because your E6300 do SuperPi faster or transcode mp3 and wmv back and forth more quickly doesn't make it a bit more useful or productive.

Errr... Oh yes. A faster processor doesn't make any difference! In fact I've decided to get rid of this PC and go back to a 486 since a faster processor doesn't make my PC more useful or more productive according to you!

If newer, faster, processors did not make a computer more useful or productive then why wouldn't we all still be using Intel 4004 processors!?

11:42 PM, December 17, 2006  
Anonymous Anonymous said...

New 65nm AMD cpu have(air cooled)a 3.1 GHz ceiling, it's time to change something more than a simple die shrink.

2:56 AM, December 18, 2006  
Blogger george said...

what about a ruby co-processor for yarv when it comes out. or a GCC coprocessor for compileing c++ at lightning speeds.

1:30 PM, December 21, 2006  

Post a Comment

<< Home