SuperMicro gets some more courage

From selling their motherboards under Monarch brand, to publishing their own press release on their Opteron boards for Google, now, SuperMicro puts Opteron solutions on their front page.

In some sense, Intel and AMD are treated equally on SuperMicro home page. Neither Intel nor AMD appears there, only product names such as Xeon and Opteron. But there are more Intel code names (such as Blackford, Glenwood, Mukturd...), as Intel dudes are very productive in making new names -- unfortunately, naming seems to be the only thing they really excel at.

Merom (Core 2 Duo) Notebooks consumes 80 watts under load

PC Magzine tests some Core 2 Duo notebooks and finds Merom literally hot. Now, your laptop will have a more violent explosion.

Some people say 80 watts was measured when it's plugged into the AC outlet. They say, if using batteries, the Core Duo will clock down and save power. I find this argument unavailing. You can't argue that Core 2 Duo is fast at 80 watts, then argue it's low power at half the speed.

In a Tomshardware test, Turion 64 X2 frags Core Duo in most of the tests. The 2GHZ Turion TL60 frags the Core Duo 2GHZ T2500, 7 to 4. Toms also benched the two thin & light notebooks, the Turion TL60 of course frags the 1.86GHZ Core Duo T2400 (7 to 2). Moreover, Turion 64 X2 is 64 bit.

You should note that these reviewers are not very bright dudes. They can install and run benchmark programs, but whenever they try to figure out something themselves, they make themselves look like fools. Remember Anand's so called negative Opteron scaling? This is the case here. Toms' so called multitasking benchmark is simply stupid. He runs multiple different programs simultaneously and measures the final completion time. Obviously, the result of this test will be determined by the slowest single threaded program and OS scheduling algorithm is a major factor. The better way to do multitasking test is lauching multiple instances of the same program.

Some people keep saying Intel CPUs were not the cause of laptop explosions. They are wrong. Batteries won't explode by itself, therefore the explosion was caused by something else. If you take the batteries out of the notebook, the batteries won't explode--even Qanta airline folks understand this. If you take out the CPU, the batteries won't explode. Thus, the Intel CPU is a key factor in the explosion of notebooks. I have analysed this and my findings were conclusive. It was obvious that the Intel CPUs could not sustain a fire of several minutes, but it was the initial cause of the explosions. Some IBM folks pointed out that CPU heat is causing laptop to explode.

I have previously conjectured that the Core Duo CPUs melted and/or exploded, creating a short circuit and leading to subsequent explosion of the capacitors and batteries. So far, there is nothing to disprove this analysis.

Some readers have taken my last senetence above out of context. Let me summarize my analysis again:

1) Batteries won't explode by themselves.
2) Therefore, explosion was caused by something else.
3) If you take out the Intel CPU, the batteries in Intel laptops won't explode.
4) Therefore, Intel CPU is a neccessary factor for the batteries in Intel laptops to explode.
5) Furthermore, I have shown Intel laptops' max power consumption is large and its spike may be off normal operating parameters.
6) All other elements of the notebook are less likely candicates of the cause of laptop explosion due to their lesser current reqiurements and lesser power densities.
7) With others ruled out, Intel CPU is the only direct cause of explosion of Intel laptops.

Some readers brought out problems with G4 laptop fires. That only proves that the G4 is also bad. It doesn't proven Intel is good.

Get it?

As for the sequence of events, my analysis was that Core Duo first overheats, resulting a short circuit, then the large current and the failure of Sony batteries' cutoff mechanism result in secondary explosions.

How did a short circuit develop?

There was an INQ article showing a picture where the pins of an Intel CPU melted. Now, consider the situation where two melted metal pins, one ground and one positive make contact with each other -- that's a short circuit. The power on that short circuit is V^2/R, since R is near zero, the power is extremely large. A battery stores about 50WH, now imagine that much of heat releases in a fraction of a second, most of the heat is inside the battery. The result is catastrophic as evidenced by those explosions.

"Dual core Sempron" frags Pentium XE 955

The Chinese tested Athlon 64 X2 3600+ (2GHZ, 2x256KB) against a Pentium D 915 (2.8GHZ, 2x2MB). The X2 3600+ frags Pentium D 915 by 30% in most tests (39% in HL2, 29.8% in Doom3, 31% in FarCry). For Pentium D to compete against the "dual core Sempron", it needs a clockspeed of at least 2.8GHZ*1.3 = 3.64GHZ.

Pentium XE 955 is only 3.46GHZ, so it will be fragged by X2 3600+ with a good margin.

95% of Intel's CPUs are below Pentium XE 955, which is priced at $999. Therefore, the dual core Sempron frags 95% of Intel's production, at a price of $125.

A lot of readers have asked me to ban those pure insulting comments. I think it's a good measure to make this place more enjoyable. So, from now on, unless a comment makes some prima facie showing of IQ, it will be disallowed.

People question IBM and Intel's "quadcore" CPUs

Both Intel and IBM's "quadcore" are actually double dual core: a MCM (multi chip module) that glues two dual core CPUs together. Only Advanced Micro Devices and Sun Microsystems have the ability to design sophisticated true multi-core CPUs.

The above are people's opinion.

My comments on DELL at Rahul's blog

Rahul, the CTO of VooDooPC wrote about his exchange with Michael Dell. It is nice to see that Rahul's wisdom made some positive effects on the industry and Michael Dell was open to suggestions after all. Hopefull, he read some of my emails.

I made the following comment:

/*********************************************************************

Michael Dell is quite young, but his thinking is getting old and outdated. He is stuck with a PC mentality, while the focus of computing has shifted from clients to servers. It's the network computing age. DELL should have focused on enterprise server business and converted its commercial desktop customers into server customers. That requires DELL to have some real understanding of computing and make some serious investment. But DELL is stuck with selling cheap PCes at high volumes and razor thin margins.

All DELL people, Michael Dell, Kevin Rolins and Jim Shneider are very arrogant dudes. They only toned down when their profits were down and stock crashing. They failed to realize that DELL is just a screw driver company. I can give a screw driver to any Joe Bloe and he can learn to make a PC in minutes... Zero barrier of entry.

I gave numerous unsolicited advices to DELL based on my accurate projections:

INTEL and DELL have become each other's liability, Intel's predicament, DELL will spit at INTEL and kneel to AMD soon, Intel should establish a uniform pricing scheme,
on AMD-DELL alliance.... and many many more.

I predicted that unless DELL goes AMD64, it will face a Enron scenario due to lack of growth and subsequent crashing of its stock. You have to realize that DELL has almost zero assets, all its value is based on hyper growth. Once growth is gone, all you see is a company with $3 billion shareholder equity, 6% of DELL's market cap.

I suggested to DELL that he should switch to AMD before Intel makes the inevitable move of establishing a uniform pricing scheme. That way, if Intel stopped giving DELL discounts, it would be viewed as a punishment for DELL going AMD, which would be bad for the lawsuit. By doing that, DELL can enjoy some Intel discounts at the initial stage of transitioning to AMD. But DELL waited too long, Intel made the first strike...

Going forward, DELL's future is quite bleak. Unless it can show double digit growth, its stock will continue to fall, and its cash will continue to burn. AMD64 market presents to DELL a 30% growth oppurtunity. DELL has one last chance, but it needs to act quicker.

I suggest DELL folks go to AMD HQs and take a week of crash course on basics of computer architecture and the future of network computing. AMD camp has all the grand masters in server computing: lead architects of Alpha, PA-RISC, Sun UltraSparc, PA-RISC, Itanium, Power4, PowerPC, RS6000...and Opteron are all in AMD camp. Intel will be totally crushed by AMD. There is an imbalance of brain power there.

*/


As I wrote previously, the fact that DELL acquired Alienware is a sign of its PC mentality. DELL should have acquired some server startup, such as Fabric 7. A move like that would make SUN very nervous.

CM1 is all you boys and girls need

$140 price tag, 400MHZ AMD CPU, 1200x900 LCD, Audio, Wif-Fi, built-in web cam, SD slot.

I think CM1 is a case of reverse discrimination. Now 3rd world kids get the cutting edge technology with automatic peer-to-peer WiFi grid network, while American businessmen get bulky and hot Core Duo notebooks that explode during conferences.

I'd like to have an AM1 (Adult's Machine 1) with the same specs.

I ran into this workstation virtualization solution product page, which is quite interesting, as it supports AMD-V. If you know any similar products, please comment.

I spent hours trying the virtualization workstation product linked above. It's a wonderful piece of software. I run a Windows XP installation inside a Windows 2000 server host. I see no slow down. I am installing a bunch of software into the hosted Windows XP. Now, imagine a hacker who hacked into the Windows XP installation --It's like Neo entered Zion -- he thinks Zion is real, but it is just another virtualized world.

K8L in sample production since June

We now know the next generation quad-core (K8L) described in the June analyst meeting is Rev G. There is a confusion about the names, but Rev G is the one with native quad-core, HT 3.0 and L3 cache (check the link to a report at IDF)*. The chip is code named deerhound, and we were calling it K8L, and we thought it's Rev H, but it is not. Rev G core will be 60% faster, Rev H (2008) will be 150% faster. Hector Ruiz said 2008 will see a real killer, he was talking about Rev H. If your memory doesn't fail you, you should recall that Rev G had been sampled and production ready since June. See Daryl Ostrander's presensation.

Someone analysed the bugs list of Intel's Core 2 Duo (Woodcrest, Conroe). One word, unreliable.

DELL Korea decided to stop selling Intel Xeon servers and switch to AMD Opteron.

* See http://www.itjungle.com/breaking/bn031506-story01.html

"To get to quad-core processors, AMD will be moving to a 65 nanometer process in 2007, which will include a totally revamped Operton core, code-named "Deerhound" apparently and presumably also known as Rev G. These Rev G chips make jump from HyperTransport 1.0 to the HyperTransport 3.0 interconnect (hey, what happened to HyperTransport 2.0?), and a new architecture that incorporates L3 cache. "

Core Duo users are second class citizens

Their laptops explode and burn tables, trucks and even houses. They face humiliation at airports. Now, Microsoft has decided that Core Duo won't be able to play HD videos.

Can they switch to MacOS? Yes. But, then, their Core Duo MacBooks may randomly shut down.

It must suck to be a Core Duo user. Their computing life has no future.

AMD needs to do a spy hunt

Reading the book "Inside Intel", one story struck me. There was an Intel FAB technician who stole a lot of gold -- in the form of tiny grains which he made into gold bars. Following a lead, Intel determined to nail him -- orders were given from the very top. An Intel manager and his girl friend spent several months stealing the trash from this guy's home, to be sorted on his kitchen table. Finally, they found evidence -- a correspondence from a gold shop. Police raided the guy's home and found unfinished gold bars and a machine to make them.

Strangely, Intel did not press charges against the guy. Instead, he was given very good references at Intel and found a job in an AMD FAB.

I wondered what the guy did in exchange for years in jail. I bet there were many similar cases.

In 2004, AMD was readying a dual core, somehow, Intel heard of it and rushed to deliver a double core.

According to this INQ story, the K8L name first came out from Intel, even before some AMD guys knew of it.

In mid 2007, AMD will introduce true quad-core described by Phil Hester in the June tech analyst meeting. It's unclear whether that's the K8L. In any case, Intel is in a helluva rush to push double dual core(Colvertown). They knew something.

SUN: dot dot dot

SUN folks are true engineers. They are not good at writing application software that requires certain level of business logic. But, when they wake up, they are good at one thing: enterprise mission critical systems with good price/performance.

The most recent report shows that SUN's 2Q06 server revenue market share had a 19% gain over 1Q06, while HP, IBM and DELL showed decline. Am I surprised? No. This is just the beginning for SUN.

Solaris 10 + Opteron + Niagra is the future of enterprise computing, folks.

The SUN Fire x4500, x4600 servers are amazing stuff. I talked about them previously, showing benchmarks where ths x4600 smashes 16P Itanium 2 Superdome. With a x4500 and a x4600, you can start your Web 2.0 shop and support 1,000,000 users, enough to get you to pre-IPO. The total cost of two machines is about $100,000. If SUN can sell 10,000 of them, that's $1 billion revenue. The numbers add up.

SUN folks now say they are the dot in Web 2.0. AMD64 makes that possible: it allows SUN to build high performance, highly scalable and reliable systems at affordable price points. Coupled with Solaris 10, SUN creates an unbeatable proposition for enterprises, large and small.

Going forward, as Solaris 10 getting easier to use and more popular, I think Linux will face serious challenges. HP and IBM can only hope that they can pump enough effort into beefing up Linux, otherwise, they will be left without an OS.

In Opteron server design, SUN is clearly ahead. HP and IBM are at 4P 8 core only. The Sun x4600 with four Opteron 880, 4 GbE ports and 16GB ram costs $25K. I went to hp.com, configured an HP Dl585 with four Opteron 880s, 2GbE ports and 16GB ram, the price is $22K.

If I were a customer and only needed 4P, I would rather buy a Sun Fire x4600 with 4 processors, reserving the right to upgrade to 8P (16 core) later. Furthermore, I can upgrade the x4600 to quad core, 32 way monster.

K8L does asymmetric operation

We all know that the 4 cores of the K8L can operate at different clockspeeds. This makes asymmetric operation possible. For example, if the load is currently single threaded, K8L can shut down 3 cores and overclock the remaining core by 50% to 4.5GHZ, all done within the same power envelope. There are some thermal issues to solve, but those should be trivial, compared to designing a CPU.

Pentium D can independently clock the two cores, because it's a double die super glued solution. I am not sure if Conroe can do it, considering that L2 is shared.

INQ reported that SuperMicro finally finds the courage and takes the liberty to put its AMD solutions on its home page. Before, you can only find the link from amd.com . The story was that SuperMicro had to design AMD boards in a basement, due to fear of agents.

Woodcrest seems to have issue with large amount of RAM

With 2-4GB ram, Woodcrest was faster than Rev F Opteron by 2%. With 16GB ram, Rev F Opteron was ahead by almost 11% in doom3, quake4, prey and Painkiller BOH.

Core2 is inherently a 32 bit chip.

Intel has already found the same excuse for its market share loss. Again, it's having chipset shortages. It seems that Intel needs to copy-exact three more FABs.

Meanwhile, morale is extremely low, as thousands of workers wait for their pink slips.

AMD 65nm in production

See FAB Tech report here, 65nm AMD procs are coming sooner.

This site reported a dual core duel between Athlon 64 X2 4200 (2.2GHZ, 2x512KB) and Conroe 6300 (1.86GHZ, 2MB), X2 4200 wins except the SuperPi bench. The X2 4200 had 18% clockspeed advantage, that's enough to defeat Conroe. In particular, in the HDBENCH floating point test, the X2 4200 won by a 35% margin. From SpecFP_rate2000 scores published by SUN, a Socket F Opteron 2.6GHZ beats a Woodcrest 3GHZ by 38%. Intel's Core 2 architecture is not well suited for HPC applications.

With 65nm parts, I expect AMD to achieve a clockspeed increase of 25%, sufficient to decisively defeat Conroe XE 6800 in all benchmarks by brute force. Coupled with architectural enhancements with K8L, AMD may achieve a whopping 60% performance/core advantage over Core2 on integer performance. On floating point, the K8L core will be at least 2x faster than Core2.

Tyan launched six socket F 1207 boards.

This CNET page has some nice DELL quotes over the years. From saying AMD was fragile to embracing AMD64 across the board, Michael Dell finally learns a tough lesson as his company's future looks increasingly bleak.

Athlon 64 X2 4200+ at 25C under load

Interesting test at LegitReviews on Athlon 64 X2 4200+ (2.2GHZ). They lowered the voltage from 1.35v to 1.125v while keeping the same clockspeed. As a result, power consumption dropped by 28 watts (from 218 to 190) under load.

A number of Intelers asked me to tell them what this result means (which was obvious to me), so let me elaborate to them, using high school physics and elementary school math. Correctly applying these physics and math knowledge requires IQ. So the following can prove at least I have a high school diploma and sufficient IQ to interpret this LegitReview result. It will also prove that Intelers do not have sufficient IQ to do the same. But it can't prove my Ph. D -- that's actually very hard to do. However, I don't have the need to prove that to Intelers any way. It's like a five star general doesn't have to prove his rank to a foot soldier.

At a fixed frequency, both the CMOS switching power and resistive power are proportional to V^2. We have (1.125/1.35)^2 = 0.694

28/(1-0.694) = 91.6

Using an AC-DC conversion efficiency of 83%, we found AMD AM2 (89 watt)'s power dissipation under load at default voltage to be

91.6 *0.83 = 76 watts

Thus, when under volted to 1.125 v, the CPU power is 53 watts. One of our AMDers asserted that lowering vcore made the chips an EE version. He was absolutely right, demonstrating the inherent intelligence advantage of an AMDer over an Inteler.

Intel's Conroe is rated at 65 watts, but that was typical power, not max. Furthermore, Intel's Conroe chipset consumes about 20 watts TDP.

DELL win vastly boosts AMD lawsuit

I have long predicted that Conroe launch time is DELL's AMD time, you can go back to read the numerous articles I wrote on that, including this one. Some idiots always ask the obvious question, why? Why does Dell have to go AMD when Intel releases a faster CPU?

For that, go back read my articles and think.

Conroe is the last straw that pushed DELL to AMD.

You will see Dell AMD desktops next month, and Dell AMD laptops in October.

I expect AMD64 to be 50% of DELL's business by 3Q07. Intel will lose market share and its ASP has crashed. Intel is expected to BK in 1Q08 to 3Q08 time frame. By then, DELL will be near 100% AMD.

Since AMD's average cost for a finished CPU is only $40, well below Intel's cost (about $75), AMD can help DELL maintain profitability and growth by offering some signup bonus.

The DELL win is a major boost for AMD's anti-trust lawsuit against Intel. AMD launched Opteron in 2003, gained little ground in two years. Immediately after the lawsuit, AMD quickly gained market share and acquired customers. This proves that with Intel's illegal activities curtailed, AMD can grow quickly in a freer market. Thus, AMD's failure to grow was not due to its own problems, as Intel claimed, but due to Intel's illegal monopolistic behaviour. This will be crucial in the damage phase of the lawsuit, once Intel is proven guilty.

Consider the alternative. If, after the lawsuit, AMD's market share stayed the same, Intel would have a very strong defense. Intel would say, "See, I stopped doing those tricks AMD complained, but nobody wanted AMD any way, so I was not the cause of AMD's past failure and I am not liable for AMD's past failure. The damage I caused was zero." Now, AMD says, "Jury, look, we sued, Intel stopped its illegal behaviour, now everyone is using us. Let's compute the damages AMD suffered in 2000-2005. "

Experts are now repeating my analysis on HP

See my October 2005 article titled "Carly Fiorina saved HP", and today's news analysis (Ex-CEO Gets Some Credit for HP Surge).

That's what you get from this Journal: insight and vision into the future.

DELL's Enron time

I predicted that DELL will go down like Enron unless it goes AMD by the end of 2Q06. Now, it's 3Q06, DELL is not shipping AMD yet. There are reports saying DELL bought two million AMD CPUs, but you have to wait a few more days until September 2006 for the back to school season. DELL is very slow in saving its own life. This AMD move would mean DELL's Enron time may be slightly delayed. However, there is a 75% probability that an Enron scenario can't be avoided.

Look at Dell's results, shareholder equity is now only $3.1 billion, and the market cap is still $52 billion. A price/book of 17. Folks, this is a house of cards, it can crash to the ground any moment. The fair price of DELL stock should be less than $3, at a level close to Gateway's.

I told Michael Dell that he needed to act quick to save his company. I warned him that once a company like DELL crashes, Feds will knock on doors trying to find scapegoats to send to jail. Had Enron survived, Jeff Skilling would be enjoying his half billion. But Enron crashed, and Skilling is in jail. SEC is now probing DELL. You can bet when DELL crashes, a penny of mismatch in books will be evidence of criminal activity.... You sign the paper submitted to SEC with materially untrue stuff? That's securities fraud..

Crooks laughing with their loot

2Q06 AMD institutional holding numbers are out. Guess what, those crooks who downgraded AMD loaded up their boat. AMD's institutional ownership reached 82.66%. These criminals.

Goldman Sachs (James Covello ): increased its AMD holdings by 53.88%
Merrill Lynch ( Joe Osha): increased its AMD holdings by 16.77%
Bank of America (Sumit Dhanda): increased its AMD holdings by 7%
UBS: increased AMD holdings by 33.36%

You have to understand that most of the small funds have low IQ folks that follow big firm analysts, such as Joe Osha and James Covello... Joe, James and Dhanda said sell, a lot of the smaller fishes dumped their AMD shares.

On August 15, the new institutional holdings are published. The low IQ dudes found the smart crooks have loaded up the boat. The low IQ dudes now thinks, hmm, why is DELL going AMD? Why did Joe's firm load up? ...

K8L taped out

Native quad-core, all those super frags ready to push Intel back by 3 to 5 years. AMD taped out dual core Opteron in June 2004, and launched it in April 2005. Expect AMD to release K8L in about 7 months as AMD has greatly reduced fab cycle time.

DELL looks increasingly silly. Everyone is there pumping Socket F, but DELL is missing. The only news about DELL are Core Duo laptops exploding all over the place, burning tables and trucks. Apparently, Dell's non-existent R&D is a limiting factor.

AMD updated its pricing list. The Opteron 2210 1.8GHZ dual core CPU is at $255, and the X2 5000+ (2.6GHZ) is at $301. You should notice that AMD is now selling Opteron 8216 HE 2.4GHZ. You also notice that there are some empty spots in AMD's lineup, which indicates that AMD's CPUs are so uniform, it can simply rate them at different speeds without the intermediate grades.

Gone are the days Intel charges $3000 for its Xeon MP CPUs. INQ reported that the 16MB Tulsa is priced at $1980. But only total retards will buy such a chip, you spend more money cooling the sucker and get half the performance of an Opteron. Eight Tulsa Xeon cores hanging on a 667MHZ FSB is retarded, each core gets a 80MHZ share, less than the bandwidth of a 80486.

SUN released some benchmarks showing socket F Opteron smashes Woodcrest Xeon in FP performance. In particular, a 2P Opteron 2218 (2.6GHZ) SUN Fire X2200 M2 outperforms 2P 3GHZ Xeon 5160 Woodcrest HP DL360 by 38% in SPECfp_rate2000 (117/84.7).

One thing I notice is that most of the designs use Nvidia's chipsets. All SUN designs use Nvidia chipsets and all SuperMicro socket F boards are based on Nvidia. I recall during Broadcom's 1Q06 CC, they claimed they would take most of the next generation Opteron chipset business. I guess Nvidia fought back..

Athlon 64 X2 3600+ for $130

The 2GHZ, 2x256KB Athlon 64 X2 3600 PIB is being sold at $130 in some markets, $20 below the X2 3800+. Assuming the retailer gets 30% profit, AMD sold the chip for $100. There is no way Intel can compete against these CPUs at these prices.

Based on this projected AMD production schedule, we can estimate AMD's 3Q06 desktop revenue. Assume the ASP of X2 is $135, Athlon 64 is $75, Sempron is $55. We have
(2210+950)*135+(2250+1750+20)*75 + 4620*55 = 0.4266b + 0.3015b + 0.254b = 0.98 b

The overall AMD desktop ASP is $83. The total desktop market is 35 million units per quarter. Intel will take the remaining 23 million market. However, if Intel sells at the same ASP as AMD, then Intel's desktop revenue for 3Q06 will be $83 * 23 million = $1.9b, which will be a $1.4 billion drop from 2Q06. You can see there is no way for Intel not to BK as long as there is an oversupply. AMD has lower cost, so it can flood the market at will. I told Intel folks last year that their only way to survive AMD capacity expansion is to cut production and cause a price hike. But, these morons choose the exact opposite approach.

The $0.98 b revenue is enough for AMD to break even. Now, add the Turion and Opteron sales.

Matrix Interpretation

I was reading this Wikipedia page about Matrix Revolutions, I was surprised to see what was obvious to me from watching the movie was not even discussed.

The following is my interpretation of Matrix

0) Neo was a program created to introduce a balance in the system (Everything you see Neo doing/seeing is thus 100% virtual).
1) Zion itself was a simulated world, it's just another level of control.
2) At the end, the Architect of the system simply relaxed the thought control rules in the Matrix.

When the Zion council talked to Neo about those mysterious machines that kept Zion alive, it was a strong hint that Zion itself was just a simulation. Neo stopped those sentinels with his hand, because he realized that he was just a super porgram and what happened in Zion was also a simulation. Neo entered a state of limbo, because of this cruel realisation that even Zion was a simulation and a mechanism of control, thus there was no more purpose to fight for it. The humans were simply imagining Zion in their vats.

The final outcome of Matrix was simple: humans accepted that they would never leave their vats, and the Architect allowed more humans to enter Zion or do whatever they wanted. At the beginning, Neo was not allowed to hack in the Matrix, agents (thought control) were everywhere). Basically, the end was just a freer Matrix -- which is better for both the system and humans.

Now, you realize that you, who is watching the movie, is also in the Matrix.

Matrix is a reflection of the western society. The people think they are free, but they are not. Freedom is just another level of control. You can say whatever you want, but eventually, the big money and media will filter the information and control the minds of the majority of people. Those who know the truth can speak out, but their views won't become main stream, but serve to strengthen the standard view. Originally, as in the beginning of Matrix, there were agents trying to shut up those dissenting opinions, at the end, the few people designing the system realized that actually allowing those free expression will make the system even more robust. As the dissidents will stop trying violence as long as they can express their views.

Now, who is controlling the minds of people? The big money. Who controls big money? A few very smart individuals behind closed doors.

Look at the world events, you see this happening every day. Some individuals arrange to have some video shown on TV, you watch it, as expected, your mind and body react to it, cause and effect. The initial command to air the video may come from somewhere in middle east, the signals got sent via EM waves, finally reach your brain... Simple as that.

Our only hope is that technology becomes so pervasive and advanced, that we will know the truth of every thing as it happens. The 50x15 and pervasive 64 bit computing vision will be a small step toward achieving this goal. As we progress, no more CNN, no more Fox, no more Forbes, all these crap will become history. The question is then, whether something like Google, which is supposed to be based on algorithms, will become another level of control.

Intel yard sale continues

The Dialogic unit, which Intel bought for $777 million was sold for $50 million. 600 workers were dumped. You wonder why Intel needs to sell this business for such a tiny amount and book a $0.7 billion loss. $50 million is just the price of 50,000 Conroe XE CPUs.

Intel has to sell 10 Con XE 6800 CPUs per month to make enough profit to feed one worker, unfortunately, Intel's FABs are unable to produce so many Conroe CPUs. It can make 2000 Netburst CPUs per worker per month, but the profit is not enough to feed one man and his family. Now, the 600 men division had to find a better owner.

The fire sale of these business divisions further proves that Intel is doomed. Intel is trying to improve efficiency. However, Intel's problem is not losing small money on these tiny side busiensses. Intel's problem is its main businesses are making less and less. These 600 workers may have incurred some loss, but it should be very small. The biggest loser in Intel is the flash business, and Intel isn't selling that to Spansion. In fact, Intel is pouring more money in the NAND ventures to compete against Spansion's NORAND. To solve Intel's problem, it must improve efficiency in its micro processor business. I can tell you that can't be done within 6 months. AMD spent three years on that and the process is continuing. Intel can't just lay off a bunch of workers and become more profitable. For instance, suppose there are 5000 Netburst design engineers at Intel, which of them can Intel lay off? That's a tough question. Each of these engineers may hold a tiny piece only he knows. And 65% of Intel's CPUs are Netburst. You can imagine that the Netburst camp still owns Intel today. You wanna fire a Netburst crew? You lose 65% of revenue and die. You bet on Woodcrest? Woodcrest is 2P only, and AMD will own 100% of 4P. So Netburst people are all in a life-death fight against CORE2 folks for their jobs and existence.

So, there is no way for Intel to improve profitability by cutting costs as its CPU revenue keeps declining at a fast rate. AMD will exit 2006 with 40% market share (run rate), Intel will be about 55%. Intel will see both units and unit price crashing all over the place. The result will be massive operating losses from 3Q06 onward. BK time is 5 to 7 quarters.

In other news, the AMD Athlon 64 x2 3600+ PIB is released. Expect massive flood of DELL and Lenovo AMD64 desktops and Turion X2 notebooks soon. AMD is heading to 40% market share and Intel heading to 50%.

AMD heading to 40% market share

I projected that AMD will exit 2006 with 40% of the x64 market share (run rate). Now it seems this target is very realistic. It's simple math, starting with 22%, to end with 32% average, assuming a linear acceleration, the end rate is 42%. AMD plans to ship 26 million desktop CPUs in 2H06. Lenovo and DELL will both expand their AMD64 offerings.

I found this interesting piece of old news.

This report from HKEPC is very very revealing. As you can see, as of Q2, only 10% of AMD's desktop CPUs are x2 dual core (1 million), about 50% are Semprons (4.95 million), there were only 10,000 FX CPUs, or 0.1% (I guess AMD sold five FX CPUs in the whole 3rd world). Now, you understand why AMD's ASP was less than $100. Moving to Q4, AMD will produce 6.58 million x2 CPUs, 2.810 million single core Athlon 64, and 4.985 million Semprons. Now, you can also understand why AMD will see revenue growth. With 600% increase in x2 volume, AMD may even see ASP increase.

Now, go back and read my analysis on AMD's chip output on July 19, 2006. I estimated that AMD's Q3 chip unit to be 118.5% of its Q1 level. Now, look at this table, which shows AMD's Q3 desktop CPU unit (11.815 million) exactly 118.5% of its 1Q06 level (9.965 million). This was an exact match. Did I have AMD inside information? No. It's just a matter of correctly using the available data and construct the right formulation of the problem to be solved.

Opteron rules them all

After SUN's Galaxy x4100, x4200, x4500, x4600 (8P 16 core), Blade 8400, IBM's x3455, x3655, x3755, LS21, LS41 mass frags, it now appears that SUN has even more frags ready for Intel - the Sun Fire x2200 m2. It seems the same 4x4 technology will be used in low end servers to frag 2P Woodcrest. This is exactly what I predicted when 4x4 was announced, using 2 AM2 CPUs in a cheap 2P server.

So the theme will be this, at 2P level, two socket AM2 CPUs will frag Woodcrest. Above 2P, Intel doesn't have an effective x64 solution. Opteron frags them all (Tulsa, Itanium).

Centrino notebook burnt down a truck

See this report. Be afraid when you board a plane with people carrying such notebooks.

AMD has 44.66% of US retail notebook market. No single fire incident was reported with AMD CPUs. On the other hand, numerous Intel notebooks exploded. No wonder Dell is going to launch AMD notebooks soon.

Centrino seems to have a wireless backdoor

One can break into someone else's Centrino notebooks in WiFi range without much effort. I think hackers must love that. Imagine one hacker was on board the same plane as, say, some dude with a lot of secrets. Assuming that dude's Centrino notebook didn't explode into flames, the hacker would have plenty of time getting into the notebook and retrieve data at relatively highspeed.

The vulnerability in the Intel Centrino wireless device driver could allow attackers within range of Wi-Fi signal to execute arbitrary code on the Centrino notebook. It is apparently a buffer overflow problem with Intel's code. All the hacker needs to do is send Centrino an oversized peice of packet with malicious instructions attached at the end. The Intel device driver code will then happily run the malicious code.

AMD's Turion 64 is immune from such bad behaviour because it's armed with NX-bit technology that can prevent execution of data. That's why you don't see similar reports for AMD64/Turion 64.

Intel's Centrino is a legacy 32 bit technology that doesn't have AMD NX. Intel has copied AMD NX bit in its newer EM64T implementaion on Netburst Pentium 4. It's unclear or unproven whether Intel's Conroe has correctly copied AMD NX.

No demand for Core 2 Duo

Opteron 265: 1021 backordered. Con E6400: 231 backordered. Con E6600: 708 backordered.

Athlon 64 X2 AM2 stock level: 21781. Note, this is the stock level in national warehouses. You can check other drop shippers, they also show the same stock level.

Conclusion:

1) Demand for Conroe is less than the demand for 2P Opteron.
2) Intel yield on Conroe must be extremely low.

AMD fully executing the kill strategy I proposed

AMD changed the rule of the game. It's called x2. On server, the game is 4P, on desktop, the game is 2P. In 3rd world, it's affordable 64 bit (32 bit x2).

1) 4x4 will start well below $1000 and become a main stream platform, which will permanently pin Intel at half the performance. We should be able to get two x2 3800+ or X2 4600+s that will frag Con XE 6800+ by up to 80%, at lower cost. Once we get to K8L, we will have 8 cores cranking. 4x4 will open a new era in personal computing: from 1P to 2P. Just like AMD64 did to server computing: moving the main stream from 2P to 4P, and pushing Woodcrest down to the ultra low end.

2) Major price slash on Socket 754 Semprons (starting at $38) to halt the Celeron flood in the 3rd wolrd

3) we already talked previously about the low cost dual core priced below Intel's ASP.

Well, AMD and Intel execs receive my regular briefings on strategic analysis. Whether they arrived at similar conclusions independently before or after my proposal is another question.

DELL to ship 1.2 million AMD64 boxes in Q3

As expected, DELL goes AMD64 around Conroe launch time. Nvidia is still the chipset of choice for Athlon 64. Intel will deliver fewer than 1.5 million Conroes for Q3.

IBM and Novell folks talk about AMD64

See AMD enterprise event #6 (video). Three IBM execs talk about the new Opteron servers. Novell's CEO and CTO are there pumping SuSe Linux 10. IBM+Novell+AMD join forces. I bet they have seen K8L in action.

This report claims that socket F Opterons have been shipping for weeks. Socket F Opteron has built-in PCI-E support.

The specs of IBM x3655 is here, and the specs of IBM x3455 is here. From the x system spec, you can see Rev F opteron's clockspeed is 2.8GHZ. The specs for IBM blades is here. Since a 3GHZ Woodcrest is only 2% faster than 2.6GHZ Opteron in Apache benchmark, a 2.8GHZ Rev F should beat 3GHZ Woodcrest in that benchmark. In FP performance, expect Rev F Opteron to smash Woodcrest.

More Intel channel partners will be stuffed to death

We re-reported that Intel's forced channel stuffiing has BKed a few Intel distributors. Now, there is a growing concern that more will be stuffed to death.

I wonder how long Intel can keep doing this to make their quarterly numbers.

IBM: Socket F Opteron 21% faster

While details are scarce, IBM claims that the socket F opterons performs 21% better than socket 940 in integer and floating point performance. We should see some nice benchmark numbers soon. IBM is very good at getting the highest scores.

As expected, IBM has unleashed five Opteron models, LS41 4way blade, LS21 2way blade, and x3755, x3655, x3455 rackmount servers. I think the x3755 is 4P, and x3655 is 2U 2P.

Some reporters asked the dumb question of why IBM didn't give more details. The answer is so obvious: AMD hasn't officially launched Socket F opteron yet. IBM was just too anxious seeing those big SUN boxes taking market share. There was an early report that AMD was going to launch Socket F on August 1st, but delayed to August 15th. I wonder who AMD was waiting for.

SUN is still quite ahead with its x4600 (8P 16 core) that smashes 16P HP Itanium 2 Superdome.

Larry Singer, Sun's senior vice president and strategic insights officer, commented on IBM's move: "We wish them well. In a couple years, they'll get competitive, and of course by then we'll have our next generation of servers out."

Indeed, SUN's x4500, x4600, and its 8400 blade are engineering wonders designed by Andy Bechtolsheim.