Tuesday, October 31, 2006

People rather wait for AMD than go Intel

Why don't they just go Intel? There seem to be plenty of Core 2 Duos around on newegg.com.

It's interesting to note that people would rather wait for AMD.

BTW, the new DELL AMD Notebooks look good, much safe than the exploding Core Duo notebooks.

Monday, October 30, 2006

People prefer Opteron over Woodcrest

Woodcrest is for ultra low end uncertified use only, Rackable report indicates. Customers want proven Opteron solutions, instead of flaky Woodcrest.

Rackable always offered both Xeon and Opteron servers.Before the launch of Woodcrest, Rackable bet a lot of people will switch, and Rackable CEO pumped Woodcrest at IDF. Guess what? Companies don't buy the crap, and Rackable posted a loss.

"There have been more barriers to change than we had initially anticipated." Rackable CEO said.

What he doesn't understand is, software compatibility is the deciding factor. People who have standardized on Opteron will not switch to Xeon with ancient FSB. The only thing good about Woodcrest is that it's faster than Netburst. But that's not good enough for enterprises which need dependable computing power.

On performance, Intel's Woodcrest is only good for workstation applications. For server applications which need a lot of bandwidth, Intel's FSB is a major bottleneck. On floating point, Woodcrest is about 30% slower than Opteron.

For Rackable to make a profit, it should stop doing Xeons. It's simply stupid to spend 50% of development and marketing effort on about 10% of its sales.

People are ignorant of copyright laws

Some reader at INQ complains about the jailing of BT admins, which shows a total lack of understanding of the copyright law. P2P operators are liable for copyright infringement because they contribute to or induce direct infringement. In this case, the torrent admins are liable for all the infringement done by others. For more information, refer to the Grokster decision by the US Supreme Court. Once liability for infringement is established, it doesn't make a difference whether it is direct or indirect.

A company caught in copyright infringement is pretty much in a deep hole. The copyright owner can sue for actual damages plus infringer's direct and indirect profits. Worst of all, the copyright owner only needs to show gross revenue of the infringer, the infringer has the burden to show deductible expenses. That's why we see Google deleting copyrighted videos.

A few copyright cases may be enlightening:

McRoberts Software, Inc. v. Media. 100, Inc., 329 F.3d 557, 566 (7th Cir. 2003)

Wall Data vs. Los Angles Sheriff's Department (May 17, 2006, 9th Cir.)

Andreas v. Volkswagen of. America, Inc., 336 F.3d 789, 56 Fed. R. Serv. 3d 6 (8th Cir. 2003)

AMD launches Radeon X1650 XT

AMD gracefully invites Intelers to the wonderland of high end graphics.

None of Intel's crappy graphics engine can play modern games.

Sunday, October 29, 2006

Intel's ASP from selling T-Shirts to partners

$20 to $50, higher than the price of a Celeron.

Intel is surely frugal these days.

Saturday, October 28, 2006

Institutional ownership of AMD reached 117%

Look at the official data at nasdaq.com. Total AMD shares outstanding is 486 million, total shares held by institutions is 569.8 million. You ask how this can be possible?

The answer is simple. It's called naked shorting. Basically, you sell millions of shares that don't exist, it's like printing fake money, you sell it, and get real cash. The naked shorting creates fantom shares called FTDs (Fail To Deliver).

Imagine a smarter Bin Laden, instead of bombing buildings, he just short the hell of every stock with FTDs, rake in massive amount of cash, and say bye bye to Americans that are shafted. This is what is happening and SEC is doing nothing.

PS: ATI acquisition only issued 58 million shares.

Friday, October 27, 2006

DELL's sub $500 slashes Intel's throat

It's amazing that someone would reach a conclusion that DELL's AMD products hurt AMD. They all forget a simple fact: AMD is ramping up like crazy. Temporary shortages will be relieved by FAB36 doubling capacity. Basically, AMD is selling all it can make, and Intel keeps piling up inventory.

If one looks at Intel's 3Q06 result, mobile is the only thing that kept Intel in business. Intel's 2Q06 desktop+server revenue dropped near 30% from 3Q05, and almost flat from 2Q06, even though Intel was shipping Core2 and Woodcrest. Mobile revenue dropped 4% from 3Q05, but up from 2Q06. Intel is still getting good money and ASP from its mobile business.

Now, imagine DELL flood the market with massive number of Sempron notebooks. Every AMD notebook is at least $200 loss of revenue for Intel (CPU+chipset). 1 million DELL AMD notebooks is a loss of $200 million for Intel. And, we expect DELL to ship about 2 million AMD notebooks.

The whole AMD v. Intel game is a bloody death struggle, whoever has the lower cost will win. AMD can achieve profitability at an ASP of $80. What AMD must do is denying Intel's the oppurtunity to dump its inventory.

The DELL-AMD alliance is a nail on Intel's coffin.

PS: Someone questioned why Intel sticking to Dell is a death trap for Intel, but AMD-DELL alliance is good for AMD. The people asking such questions are typical liberal arts people without the ability to apply logic -- they only know to use analogy. AMD and Intel are at completely different situations, AMD is gaining market share no matter what and Intel is losing market share regardless. It is called the destruction of monopoly. The only option Intel had was to die a quick death or to seek a lower level of survival.

PPS: Some Intelers brought up the Anand busting articles, which have already proven true. Intel claimed 40% performance lead at IDF, I estimated the lead to be around 10%. Intel's have been busted and its fraudulent comparisons of Woodcrest and Opteron have been exposed. There was nothing to disapprove my earlier conclusions. The industry has widely recognized that Intel has become more competitive with its Core2, but it has also recognized that Intel lead is likely to be very short lived. While Intelers cheer on the reported need for AMD to do A1 silicon for K8L, such revisions are the norm. Any delay in K8L merely gives Intel some additional time to breath. Once K8L is out, the game is over for Intel.

Thursday, October 26, 2006

AMD should create a quasi quad core

AMD should be able to do this with a blink of eye. Just connect the ccHT links on two Opteron dies and arrange to deliver power to the second die. This should be much easier to do than Intel's shared bus double die solution, where two dies generate two loads.

BTW, after I posted a criticism of Rahul's rather primitive way of demonstrating Kentsfield performance (running multiple SuperPIs), I found all my comments are gone.

While I am not saying there was a causal nexus between the two events, I do think, sometimes, you need to be a Doctor of Philosophy to be able to face criticisms and adhere to fundamental principles.

Wednesday, October 25, 2006

AMD and Microsoft will rule

Fusion of the forces is complete. Vista will be the final nail on Intel's coffin.

Nvidia, HP, DELL, Lenovo, Asus, Broadcom praise AMD's leadership. Expect every AMD CPU to be DOOM4 capable, while Intel users won't be able to use Vista unless they get a AMD or Nvidia graphics solution.

The iNazis tried to block people from buying ATI products.

As I have shown here, Intel is hopelessly kicking below AMD's waist.

Free Image Hosting at www.ImageShack.us
For those who analyse AMD+ATI merger, they all fail to see the real hand behind it -- Microsoft. Microsoft has worked closely with AMD on Vista's Aero engine and Direct X 10. My cheap $40 MB with AMD Radeon IGP can handle Vista beautifully(machine spec). It seems to me that MS and AMD have teamed up to frag Intel. The Intel + Apple axis has enraged the emperor at Redmond. Patty knew Bill Gates is a "ruthless, aggressive, demanding" guy, he knew Intel's love affair with Steve Jobs will be punished. When Paul O said he could't reveal the secret things he is doing with Steve or he would be killed, I bet that Gates was infuriated. Microsoft specifically recommends AMD graphics for Vista. Intel graphics? Microsoft didn't even mention them.

Intel's situation will be extremely bad in the mobile market. Every Turion will be paired with a ATI or Nvdia IGP, while Intel notebooks must use Nvidia discrete graphics or it will simply crap out on Vista's demanding Aero GUI. Even if Intel managed to get Aero to work marginally with its crappy graphics, it will leave no room for playing the simplest 3D games.

Now, you understand why Michael Dell chooses AMD. Dell is a smart guy, smarter than Steve Jobs for sure.

Soon, all Intel graphics solutions will be declared unsuitable for Vista and Intel's BK will be accelerated as a result.

Tuesday, October 24, 2006

AMD maintains massive performance lead

Integer performance, AMD leads by 22% to 100%.

AMD score: 175, or 279.

Intel score: 143

Floating point performance, AMD leads by 61%

AMD: 174, or 182

Intel: 108

DELL claims that AMD server consumes 20% less power than one with Intel's top end Xeon 7140 (16MB L3).

Intel only plays in the ultra low end with Woodcrest.

AMD market share at 23%

According to Mercury Research, AMD share grew to 23%. Intel also grew its share by 3%, displacing VIA. It seems that Intel used $15 Celerons to battle VIA's $30 C3s.

I told you here already. In relative terms, AMD grew faster than Intel.

Some of you will say what happens to this analysis. As of the reader pointed out, the IDC number was for year/year growth, instead of sequential.

Read this for some education.

XO makes Intel inventory worthless

AMD powered XO laptop is entering mass production.

Core Duo and Celerons will have to be junked. That's part of the reason you can't count Intel's inventory as current assets, their shelf life is shorter than that of a banana.

Monday, October 23, 2006

Hector Ruiz and Michael Dell on stage

Ruiz welcomed three guests to the stage: Juergen Rottler, executive vice president of Oracle Support and Oracle on Demand; Michael Dell, Chairman of Dell; and Dr. David Brailer, the United States' first National Coordinator for Health Information Technology.

Rottler discussed how AMD Opteron(TM) has helped Oracle.

Dell addressed how AMD processors will provide a compelling choice for Dell's enterprise customers. Dell Inc released two Opteron servers which beat the crap out of Intel's Tulsa solutions.

FAB36! crank, crank.

Patty cried and cursed seeing the picture above.

Sunday, October 22, 2006

Charlie not good at arithmetic (AMD share estimate)

Charlie at INQ claimed that AMD had 27% desktop market share in 3Q06, which is simply wrong. The source he quoted, BusinessWeek, was using 2Q06 data. 3Q06 data hasn't come out yet.

For 2Q06, we know AMD had 21.6% market share, Intel was at 72.9%. VIA took 5.5%, but it was said VIA number was inflated.

For 3Q06, we know the following:

AMD's overall units grew 18% q/q, with over 50% unit growth in mobile and single digit growth in server.

INTEL had double digit growth in server CPUs, but apparently its mobile and desktop had single digit unit growth. We conclude that Intel's unit growth in 2Q06 was less but close to 10%.

Using their 2Q06 numbers as basis, we have

AMD units = 21.6 * 1.18 = 25.488 units

INTEL units = 72.9 * 1.1 = 80.19 units

Therefore, AMD's market share in 3Q06 is

AMD_share = 25.488/(25.488+ 80.19+VIA),

assuming VIA's units did not change in 3Q06, we have

AMD_share = 25.488/(25.488+80.19+5.5) = 23%
INTEL_share = 80.19 /(25.488+80.19+5.5) = 72%

Of course, if VIA's units went down, then AMD's market share will be higher. If one takes out VIA, AMD's share is 24%.

On ASP, in terms of 2Q06 number, AMD's ASP was 1.33/1.18/1.22 = 92.38% of its 2Q06 level, a near 8% drop.

Saturday, October 21, 2006

Hexus misses the main point

Hexus, using a bunch of handwaving arguments, called my Intel BK prediction a result of fanboyism. I beg to differ.

First, all empires fall and all companies BK, thus the only question is when Intel will BK. Second, a company BKs when it is unable to pay debt. Thus the Intel BK inquiry is a financial projection which may not be directly correlated to technology. DEC had Alpha, but it pretty much BKed.

Since 85% of Intel's stuff are legacy chips, and since AMD is quickly converting to 100% 65nm, we know any temporary advantage Intel had will be short lived. Even today, AMD's X2 3800 is faster than 90% of Intel's chips. Therefore, what AMD can take is solely contrained by its capacity. AMD knows this and it's in no hurry to introduce K8L desktop chips (but 4x4 will get K8L boost soon). People saw Intel's seasonal 3Q06 as a sign of recovery, but paying attention to the details, you noticed that Intel's server+desktop CPU sales stayed almost flat, Intel's revenue growth came mostly from mobile and chipsets. Thus the switching to mobile was the key factor. AMD was unprepared for such rapid market transition, but nevertheless managed to achieve 50% sequential unit growth in mobile. With the soon to be announced DELL AMD notebooks, expect AMD to grab a big chunk of that market.

Essentially, going forward, AMD will be able to sell all it produces, while Intel will fill the remaining void and sit on a higher and higher pile of legacy chips. Intel has lost the control of its own fate and its financial results will be a function of AMD's capacity. As AMD's production scale rises, its workforce and cost stay pretty much the same, but Intel will have shrinking market share and 10x the workforce and cost. Once AMD cross the 33% market share line, it will quickly approach 40% or even 50%. With ASP sitting around $100, Intel will suffer $1 billion to $2 billion quarterly losses.

Based on Intel's level of liquidity, I re-iterate my projection that it will BK by 2Q08.

Friday, October 20, 2006

AMD FAB36 ramped to 13000wspm

See FAB Tech analysis here, which concluded that FAB36 has ramped to 13,000 wspm. Dirk Meyer said current output is 10,000 wafers per month. The difference is wafer starts and wafer outs.

Wednesday, October 18, 2006

AMD market share approaching 25%

AMD result is out. Units grow 18% sequentially. It's interesting to compare my earlier projection on AMD's unit output. It seems it gained some more market share., especially considering the fact that AMD's 2Q06 had an extra week. Revenue grew 32% year/year.

Gartner and IDC reported that PC shipment grew 7% in 3Q06. AMD's market share was 22% in 2Q06, therefore, AMD's CPU market share in 3Q06 is 22*1.18/1.07 = 24.3%. One should note that Intel's revenue growth was mainly from mobile and chipsets.

Given the massive price cuts, AMD did a very good job.

Unlike Intel, who had to sell investments to book profit, AMD had organic growth, its mobile units are up 50%. AMD's gross margin suffered a few points mainly because of transitioning into DDR2. Now that the transititon is done, its margin should improve.

Going forward, with the launch of 65nm parts, AMD's cost will be down substantially.

AMD also indicated that more OEMs will sign up in 2007, presumbaly those Japanese vendors who were excluded by Intel 5 years ago.

On units, AMD will grow 35% in 2006, and 35% in 2007. In other words, a growth of 80%.

Tuesday, October 17, 2006

Another look at Intel's balance sheet

We do this exercise on Intel's balance sheet again.
Curret assets -current liabilities - inventory = $17.6 - $9.05 - $4.48 = $4.07

$4.07 billion is all liquid assets Intel has in the pocket.

What else? Intel has some long term investments, but it's shrinking. It was $4.1 billion in 2Q06, now it's $3.4 billion. Long term debt stayed at $2.6 billion.

The total amount of money(net) Intel has is thus $4.07 + $3.4 - $2.6 = $4.87 billion. This includes everything, cash, short term, long term and strategic investments.

How soon can Intel BK? It can happen really fast -- two quarters after AMD FAB36 fully ramps up, Intel will BK.

Some readers asked why inventory is excluded in computing Intel's liquidity. The point is, assuming Intel continues its operation, it will always carry 4 billion inventory at any given moment of time. At the time of its BK, these inventory worth 0, because they are unwanted. You can't use that inventory, even if it's finished CPUs, to pay debt. So you have to deduct inventory from the current assets in BK analysis.

INTEL inventory piling up

Intel released its results. Its inventory is piling up even after writing off $100 million of junk.

Originally,Wall Street expected revenue of $9 plus billion, Intel lowered it to $8.3 to $8.9, and made $8.7. Operating income fell 56%.

Revenue for server and desktop CPUs dropped from $4.9 billion of 3Q05 to $3.5 billion of 3Q06. Patty won't get a bonus. Mobile CPU dropped from $2.3 billion in 3Q05 to $2.2 billion in 3Q06. Total revenue dropped from $10 billion in 3Q05 to $8.7 in 3Q06.

It's clear that Conroe and Woodcrest can't save Intel. Intel had a whole quarter of Woodcrest sales, and 2 months of Conroe sales. Yet, their desktop+server CPU revenue dropped from $4.9 billion to $3.5 billion, flat from 2Q06. So, in the back to school season, Intel's desktop+server revenue didn't grow with market. This is clear proof that Conroe is a killer -- an Intel killer. On the other hand, Intel's mobile business, which is mostly old stuff, is only slightly down.

Another thing you learnt from Intel CC is that most of their stuff are still 90nm. They only shipped 40 million 65nm CPUs total.

Mysteriously, instead of booking losses of $1.6 billion for dumping businesses it bought at high prices, Intel booked $200 million gain and no reduction in goodwill.

Though Intel made a profit, $400 million was gains from selling investments. The fact its unit actually grew indicates that AMD is capacity contrained. Basically, Intel is only selling what AMD can't fill. We will see this once we have AMD's result and inventory levels out tomorrow.

I re-iterate my projection for Intel's 2Q08 BK.

AMD is expected to post 50% year/year revenue growth tommorrow.

75% of Intel's 4Q06 supply are crap

Read this on Intel's 4Q06 product supply plan:

Pentium 4-- 40%
Pentium D 8xx (90nm) -- 25%
Pentium D 9xx (65nm, Presler) -- 10%
Conroe -- 25%

I think Intel will have to write off all Celerons, no one wants that crap, when they can buy Sempron 64 at lower price.

DELL has plenty of AMD chips.

SUN Micro frags Intel with X4600-M2

Opteron 40% faster than Woodcrest. One Opteron server replaces 50 Intel Xeon server. Direct Connect Architecture more efficient than Intel's FSB.... read it yourself.

Oh, boy, I missed this:

"Additionally, under full load, the 2-socket Sun Fire servers demonstrate better scalability and higher performance than similarly sized Intel "Woodcrest"-based servers, with the ability to consolidate up to 22 legacy servers using VMware."

Monday, October 16, 2006

INQ nicely summarized what I said long time ago

AMD and Intel History. AMD and Intel Architecture . AMD's Architecture Lead Persists.
ccHT is becoming another industry standard.

Intel copycat: Intel is following in AMD's footsteps.

A judge with a sense of humor

Read here. "There's a lot of violence," he said, "a whole lot. Less than we see on television every night."

Yankees have constitutionally protected fundamental rights to express themselves violently.

Saturday, October 14, 2006

AMD retail presence remains strong

Just went to circuit city. There are 13 AMD desktops, 5 Intel. There are 7 AMD notebooks (mostly Turion X2s) and 7 Intel ones.

Also, this DELL deal is unbeatable: X2 3800+, 512MB ram, 19 inch LCD, $569. The advantage of AMD is that the integrated Nvidia graphics is DOOM3 capable, zero problems playing 3D kids games. With Intel, you either have to use crap Intel graphics which can't play anything mildly 3D or buy an expensive discrete graphics card.

There are concerns about AMD's margin because of selling to DELL. I think Intel has deeper concerns. In 2Q06, AMD's gross margin was 57%, that was on an ASP of about $90. Intel had a gross margin of 50%, with ASP at about $125. Now, DELL seems to be draining AMD's capacity, one analyst expected the ASP AMD is selling to DELL is $67. At that price, AMD should be doing OK. The rationale is every AMD chip DELL sells is one less Intel chip sold. For that alone, it's good business. The economics of scale play here. The FABs and people cost the same, the more wafers you push through, the more chips you sell, the lower the average cost. As AMD cross the 33% market share line, it will head straight to 40%, leaving Intel at about 55%. With a 55% market share, Intel will be selling about 30 million chips per quarter, brining in a quarterly revenue of about $3.6 billion, adding some chipset and flash revenue, Intel's revenue will be about $4.8 billion per quarter, and that's $2.2 billion quarterly loss. How soon will Intel BK? Just check its liquidity.

So for AMD, the most important thing right now is to take advantage of its lower cost and drive Intel out of business. It's a war. You are either quicker or dead. Once Intel is dislodged from its stronghold on the enterprise client market, it will be relatively easy to finish the job to wipe it off -- just sell at $20 below Intel's cost.

Of course, Intel can try salvage the situation by raising prices, so its revenue will stay relatively constant even it loses units. But it's too late to do that, and that will resulting in AMD gaining more market share, the vicious cycle continues, until AMD is fully ramped up and be able to supply 95% of the market. At that point, Intel will become a marginal player, just like SGI today. I expect a chapter 11 restructuring, in which Intel will auction off substantial portion of its fixed assets. It won't be able to get much money from its 0.25 micron and 0.50 micron FABs, it will have to auction off some of its 300mm FABs.

Do expect the above scenario to happen fairly soon. Mid 2007 will be a key juncture for the industry, as AMD fully converts FAB36 to 65nm and Chartered FAB7 starts 65nm production for AMD.

Friday, October 13, 2006

AMD K8L production capacity estimate

Look at this picture, which shows Rev H die size to be 283 mm^2.

Using a 17x18 geometry with the wafer program, we found a 300mm wafer can produce 202 dies. To take 100% of the server market, AMD needs to produce about 4 million K8Ls per quarter. In other words, it needs 20000 wafers per quarter, or 7000wspm. This leave 11000wspm to produce dual core desktop chips. We should expect Chartered to add at least 5000wspm.

Thursday, October 12, 2006

Charlie showed some wit

He wrote: "Corporate Marketing is looking awfully redundant, how many people do you need to sell to a single account?"

I think Intel need to hire more marketing droids to sell to a lot of small vendors, 5 chips here and there. Patty will have a lot of fun messing around.

In other news, DELL's AMDT notebook is at the shipping dock. Nvidia is in a tough situation, sleeping with Intel would mean certain death, the only choice is to rely on an expanding AMD market.

How soon will Intel BK? That's an interesting question.

I did a capacity estimate long time ago -- note I was basing my calculations on dual core. My number was > 50%. However, depending on the ramp and yield, I later adjusted it down to 40%, which is very conservative.

Wednesday, October 11, 2006

CNET copies journal of pervasive 64 bit computing

Read this at CNET. Sounds familiar?

You read it here almost a month ago.

XBit has more details on how DELL dumped VIIV in favor of AMD Live! I guess Core 2 Duo's FSB can't really handle the intensive bandwidth requirements for modern media applications.

BTW, INQ reports DELL's new AMD Inside PC. Patty wailed seeing it.

Like a jealous woman, Patty demands exclusivity. Between Intel and DELL, there is a love/hate affair. Intel now hates DELL more than anyone else, even though DELL still buys 25% of Intel's production.

Intel accused of stealing intellectual property

Israeli engineers accused of stealing Transmeta's technologies in Core 2 Duo.

I guess Transmeta wants $50 per Intel CPU, treble that, you get $150 per chip.

AMD having serious problem meeting demand

I was checking the stock levels at this site to monitor the supply of Athlon 64 X2 3800+ AM2, I noticed that since it went out of stock about one month ago, the dates for replenishment got pushed back again and again and again. Today, I read INQ's report on AMD shortages. So I did some more exercise to check other AMD CPUs. The result is shocking. Not only AMD is short of X2 3800+, it's short on even single core Athlon64s.

The AM2 3200 has near 3000 on back order. The 3500 has 2000 on back order. The Sempron 3000 stock level is 1. Go check the x2 3800,4200, 4600, 5000, they are all on back order.

AMD seems to have underestimated demand. I hear Hector yelling at FAB36 and Chartered FAB7 folks to push more wafers.

Tuesday, October 10, 2006

If this is true, Inte's BK will be sooner

AMD will reportedly slashes Opteron prices. Dual core Opteron 265 at $209?? I guess Chartered FAB7 is cranking like crazy.

This also indicates that AMD is preparing the launch of 65nm parts.

Once K8L is out, Intel's server market share will be reduced to near 0. I suggest AMD to release a low end K8L at $300 -- that will totally kill off Intel's server line.

As I said, AMD must sacrifice profit margin to frag Intel, just keep taking market share, and Intel will BK.

BTW, DELL won't pay $50 for a VIIV sticker, but went AMD Live!

Photo of a true quadcore CPU

See this. Patty must hate it. Intel engineers will have more sleepless nights digesting AMD PDFs and pulling their hair -- to copy AMD's features, such as 48 bit physical address.

The Chinese did a test on the low end X2 3600+ and Intel's bread and butter P4D 820 (2.8GHZ), the X2 3600+ is up to 20% faster. Intel needs a 2.8*1.2 = 3.36GHZ Pentium D to compete against the lowest Athlon64 X2.

Conroe, you say? DELL has a quota of 12 chips per day.

Monday, October 09, 2006

Intel will have to raise prices

The screw is tightening for Intel. As AMD take more and more market share, Intel will have to raise prices to delay its BK time. The math is simple. AMD is expected to exit 2006 with 40% market share(run rate), leaving Intel at 57%, and VIA at 3%. Right now, low end CPUs are priced below 512MB DRAM.

Under such circumstances, if the CPU price keeps steady, Intel will BK in five quarters.

Therefore, I expect Intel to raise prices to delay it bankruptcy by two quarters to 2Q08.

4x4 is the game in town

Alienware shows 4x4 power, Conroe does not worth mentioning. It's highly probable that a K8L chip will be used in 4x4 demo.

Conroe is a tragedy. It pushed DELL to AMD, it is not good enough for high end, and it's too expensive for the mainstream.

I expect DELL to go 100% AMD by the end of 2008.

Saturday, October 07, 2006

Core 2 Duo is slow as hell

I was at Frys today and saw a Core 2 Duo 6400 PC. So I played with it, such as launching Internet Explorer and playing some video. The crap is definitely very slow in responding to my clicks. When I launch IE on the X2 4200+ PC next to the Core 2, it's up instantly. On the Core 2 Duo, it takes about 2 seconds. Viewing video is the same. On X2 4200+, response time is near zero. On Core 2 Duo, it's noticeable.

This brings up a question: response time and throughput. With AMD64's Direct Architecture, low latency leads to low response time. With Core 2 Duo's FSB+cache architecture, latency is large. I guess when I click the IE button, the Core 2 Duo has to do a lot of pre-fetching and caching on the FSB bottleneck, resulting in a non-responsive machine. AMD64 does not have this problem, and feel snappier.

Benchmarks mostly measure throughput, not response time.

When you play a game, it doesn't matter how large the frame rate is, as long as the rate is enough for smooth viewing. What matters in a game is the response time between you pulling the trigger and the firing of the shot. A Core 2 Duo may generate 10% higher frame rate, but it may lag by 0.5 second in responsiveness. If that is the case, the dude with Core 2 Duo will be the one that is fragged.

I just downloaded Windows Vista for AMD64 RC2 (64 bit) and installed it on my Athlon 64 X2 3800+ (which I paid $159 for both the CPU and MB with ATI IGP). I am typing this message on this machine right now. I have to admit that Vista looks great. The nice thing is Vista automatically load Cool& Quiet and the PC is running at half the speed. The CPU usage was about 10%. But you definitely need more than 1GB RAM.

Installation was quite fast, except at the beginning it seemed to be stuck without any indication of progress. One nice thing about Athlon 64 is you get native true 64 bit power, unlike those emulated 64 bit technologies.

Carly had a degree in philosophy

Sharp and very quotable words from Fiorina:

"A woman lays people off, she's heartless. A man lays people off, he's decisive."

"The agents of change are always resisted. ... Some people fear the unknown, whatever it may be, even more than they fear the troublesome but known present."

Reading this piece of the news, no men quoted in it had a fraction of the wisdom of Fiorina.

The timing of the book is perfect. HP's board members are exposed--some are leakers, some are leakers allies, others are hunters ... Only Carly Fiorina can look down at this mess with philosophical insight.

Mark Hurd inherited the legacy of Carly Fiorina.

Thursday, October 05, 2006

AMD must block possible INTC-NVIDIA merger

AMD must file opposition with DOJ against INTEL-NVIDIA merger, which will create another monopoly in the graphics chip market. The waiting period for HSR Act is only 30 days, AMD must be prepared.

Nvidia controls 90% of the AMD64 chipset business. If Intel plays nasty, such a merger may cause disruption for AMD's business. By filing an objection and invite DOJ to perform an investigation, AMD can at least buy several months of time to ramp up ATI chipset production.