Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
AMD has stretched the PR scheme to the breaking point as did Cyrix. When that happened Cyrix lost all credibility and disappeared from sight.
If the water option grants to the Intel executive officers become quite valuable in the future, then the shareholders will be very happy. The stock price could climb to just the strike price of the options -- where the executive officers make nothing -- and the shareholders would smile.
Incentive compensation that depends on company performance is certainly better than paying double-digit millions in current compensation without regard to performance as measured by an incresed stock price.
The Centrino platform specification, standard or whatever you want to call it, addresses these problems.
Why Wireless Doesn't Always Work
http://www.zdnet.com/anchordesk/stories/story/0,10738,2909323,00.html
What I was trying to say was that the Centrino platform is a specification for a collection of features, not just a wireless connection. This specifcation will be relied upon by the wireless ISPs to allow an easy connection.
Over time if the specifcation becomes widespread, it will achieve the status of "standard" in the same sense that uses the term Industry Standard server to refer to an Intel platform.
The Centrino platform has features that allow it to make a wireless connection with a lot less effort than is currently the case. A current ZD net article talked about the many difficulties getting a wireless connect to work properly. A major feature of Centrino is to allow use of encryption, LEAP encryption and connection using VPN so you can access a corporate connection with no hassle.
It is not an easy task to get a wirelss connection working and it is certainly not something that the typical user can handle with ease. Wireless will not thrive until a non-technical business or home user can enable a connection at a Hotspot.
You should gather some information before you take a wild shot in the dark on virtually no iformation just a wild what if by the notoriously unreliable Inquirer. See my post, IBM is actively working to get Hotspot vendors to sign on to the Centrino standard.
Quit whining and grow a wee bit hicker layer of skin.
Both Intel and IBM are working hard with a host of Hotspot ISPs to make sure that the infrastructure is available to take advantage of the Centrino standard. The vendors are merrily jumping on the bandwagon because they don't want to be left behind.
That does not sound like IBM is wondering about adopting Centrino, quite the opposite.
Not a single OEM of any note has announced that it will use the Opteron. So AMD has to build its own servers and give them away to get people to use the things.
This is not totally accurate but it is a good approximation of the situation. There is lots of fine print in the actual rules.
In the old days, Incentive Stock Options (ISOs) were a great deal. You could exercise, hold for the long term and only pay long term capital gains tax when the shares were sold. Then AMT hit and if you held the stock through the end of the year you had to declare the difference between market price and cost at the time of exercisefor an AMT calculation. There was no withholding of tax at the time of exercise.
The big surprise for many people was that they owed tax on a "gain" that that occcured on paper. If you held for a long term, you paid at the long term capital gain rate with a credit for your prior AMT payment. This worked if the stock went up.
A disaster was created when the stock fell and a person owed AMT on a gain that was never realized. In fact, the stock was often below the exercise price when the AMT came due after the crash. The strange result was that you owed AMT even though there was a loss on the transaction.
Even in normal times, many people were surprised when AMT generated tax liability. If you had not planned ahead and made estimated tax payments you could also owe penalties for underwitholding.
Non-qualified options were considered ordinary income at the time of exercise and tax was witheld at that time. This helped many folks because there was no AMT surprise.
People that were aware of the AMT problem would often engage in a buy and same day sale to recognize the gain as ordinary income and set aside an estimated tax. Then the person would take the leftover cash and decide where best to invest it. Another apporach was to sell somestock, pay AMT and the tax on gain and keep the rest of the stock.
Because of these and other factors, ISOs became unattractive and many companies, including Intel, started to grant nonqualified options.
Interesting observation. Intel switched from ISO's to nonqualified options in the early 90s, so almost all live options for Intel employees are non-quals.
Re: "With the exception of remaining options, that leaves Otellini with practically no position in Intel - only a few thousand shares"
You are only off by factor off by a factor of 100 for stock owned.
When you look at an investment portfolio's value you don't toss out the value of options to obtain its total value. Are you trying to suggest that a vested, but unexercised, in-the-money stock otion has no value? Guess that is another example of the divide equals multiply math aka more is less.
I think that the "consumption" definition means that AMD had a "Buy One, Get One Free" sale during Q4. The Free one came out of already "sold" inventory.
Note that the graph shows doubling of transistors every 18 months and that is Moore's Law. The author says that performance must double on schedule or Moore's Law breaks down.
The inability to understand the distinction and get the first premise correct is a red flag for the fact that the autor's thoughts are a total misunderstanding of the industry.
For what it is worth, Bill Siegle is a former IBM guy.
When you are making audio chips for greeting cards the latest and greatest process is not necessary.
If you are making state of the art processors, guess what, state of the art process technology is mandatory.
"Strained silicon" to pump up chips
http://zdnet.com.com/2100-1103-976120.html
By John G. Spooner
Special to ZDNet
December 5, 2002, 11:34 AM PT
The strains of creating faster processors will get an airing among chipmakers next week.
Researchers and design engineers from IBM and Intel will present papers at the International Electron Devices Meeting (IEDM) next week that detail their work on "strained silicon," a manufacturing technique that promises to boost processor performance by up to 20 percent.
The San Francisco convention, one of the principal events for semiconductor designers, will feature a keynote speech by Intel Chairman Andy Grove as well as presentations from Sony, Samsung, Pennsylvania State University and the Massachusetts Institute of Technology, among others.
Strained silicon--like double-gate and triple-gate transistors, silicon on insulator (SOI) and other new design ideas coming to the fore--should allow chipmakers to maintain the performance curve established by Moore's Law, the semiconductor industry principle stating that manufacturers will double the number of transistors on chips every two years.
Typically, increasing the number of transistors (a trick accomplished by shrinking them) leads to higher performance and new capabilities for processors. But this approach also leads to escalating levels of energy consumption, among other problems--spurring designers to come up with creative solutions.
"There is a fundamental change in the methodology of progress," said Bernard Meyerson, chief technology officer of IBM's Technology group, which includes IBM Microelectronics. "You are going to work a hell of a lot harder to stay on a predictive curve...You are actually going to have to increase the innovation in materials."
This change in perspective means that deep research in materials and transistor design--especially the kind of research that IBM and Intel researchers specialize in--will become much more important in the future. The concept of strained silicon has been around for 30 years, but it only recently moved to the center of concern for chip designers.
Strained silicon--which will appear in Intel's Prescott chip next year--stretches the distances between the silicon atoms in transistors, the tiny on/off switches that form the basis of a chip. Moving these atoms slightly farther apart reduces the atomic forces that interfere with the movement of electrons through the transistors, leading to better performance and lower energy consumption for the chip.
In strained silicon, the atoms are stretched by inserting germanium atoms into the chip's silicon lattice. Another chipmaking technique involves adding a layer of silicon-germanium into the transistor bed. The two technologies are conceptually related--but different--and can be used in the same chip.
Intel's Grove is expected to touch on some of the hurdles facing chip designers during an IEDM keynote address scheduled for Tuesday. The speech is expected to cover the semiconductor industry's approach to past problems and its future challenges, including the use of new materials and the introduction of new transistor designs, an Intel representative said.
R&D rivalry
The convention will also mark the latest episode in the ongoing research and development battle between IBM and Intel. The two companies are largely the leaders in the field, but certainly not friends. IBM often points out that it is responsible for many semiconductor design breakthroughs, such as using copper rather than aluminum wire to connect circuits. Intel counters that it sells far more chips than IBM does.
As previously reported, IBM researchers are slated to present three papers on their efforts to build a double-gate transistor using the Fin Field Effect Transistor (Fin-Fet) approach. These transistors could be used to control power consumption, among other things, in future chips.
IBM Microelectronics also plans to unveil a paper on its new 350GHz communications transistor, announced last week. Meanwhile, IBM Research has lined up a paper on 3D circuit designs, which use two or more layers of transistors.
For its part, Intel discussed its take on multiple-gate transistors at its Intel Developer Forum this fall. The chip powerhouse looks likely this time to concentrate on discussing its forthcoming 90-nanometer manufacturing process, which pairs strained silicon and other new materials to boost performance. It will present separate papers on the process for PC chips and for communications processors.
But beyond that, the company is particularly proud of its ability to implement strained silicon for a low cost. The technology is a "key part of our strategy now and one of the elements that we use to extract more performance out of our transistors," said Mark Bohr, an Intel research fellow specializing in manufacturing technology. "It's going to go into manufacturing next year and in a cost-effective manner. That's a significant accomplishment."
The use of strained silicon raises chip manufacturing costs by 1 or 2 percent for a performance gain of 10 percent to 20 percent, Bohr said.
IBM looks likely to bring strained silicon to market later than Intel. Initially, the company said it would introduce strained silicon in 2003. However, it has since decided it can reach the performance levels it needs at the next process step--the 90-nanometer level--using existing technologies such as silicon on insulator.
IBM now plans to introduce strained silicon in its 65-nanometer manufacturing technology, which lies two chip generations and about three years in the future, said Jeff Welser, a project manager at IBM Microelectronics.
News.com's Michael Kanellos contributed to this report.
Tech titans launch Wi-Fi company
http://zdnet.com.com/2100-1105-976225.html
By Richard Shim
Special to ZDNet News
December 5, 2002, 12:37 PM PT
Intel, IBM and AT&T have officially thrown their combined weight behind the effort to create a nationwide network of public "hot spots" that would give people wireless broadband Internet access from just about anywhere.
As expected, Intel Capital, along with Big Blue, AT&T and investors Apax Partners and 3i, announced the creation of Cometa Networks--formerly known as Project Rainbow--a new company focused on deploying hot spots throughout the United States. Hot spots are public areas where people can access the Internet using products based on 802.11b, or Wi-Fi, a wireless networking standard with a range of about 300 feet from a network's access points, or radio transmitters.
"Wi-Fi will bring (wireless) computing into the general public environment, leading to increased business productivity and new consumer applications," Lawrence B. Brilliant, chief executive officer of Cometa Networks said in a release Thursday.
The Cometa effort is the latest in a series of moves to create a nationwide Wi-Fi network. A handful of companies already exist that sell monthly subscriptions that provide access to hundreds of different hot spots. There are also a number of free public networks created by those who don't mind sharing their Wi-Fi connections, but cable companies have recently begun an effort to crack down on such setups.
"There's a landgrab mentality in the market right now, as companies look to attract carriers who want to avoid undertaking the capital expense of building out a national Wi-Fi network," said Navin Sabharwal, an analyst with research company Allied Business Intelligence. "Cometa will act as a wholesaler to wireless carriers looking to add Wi-Fi to their services."
Sabharwal expects others to attempt similar nationwide services.
Cometa plans to deploy 20,000 Wi-Fi access points in the coming years. Since the company will act as a wholesaler, selling access to carriers and large companies, it will not set pricing for services and will not market itself as a brand name.
Cometa's network and service will be designed to work with the same sign-on procedures, e-mail addresses, IDs and passwords people use on their desktop PCs, so they can move out into the field with a notebook computer or handheld device and continue Web surfing and e-mailing with few hassles.
Cometa will sell the service, based on technologies from Intel, IBM and AT&T, to telecommunications companies, Internet service providers, cable operators and wireless carriers, which will in turn sell it to corporate or consumer subscribers. The new company will begin setting up hot spots throughout the top 50 cities in the United States next year, with AT&T providing network infrastructure and management, and IBM providing wireless site installations and back-office systems. The deal's financial terms were not disclosed.
Cometa will have offices in San Francisco and New York.
0.18 Micron Itanium 2 SPANKS the 0.13 micron SOI Clawhamster !!
Compliments of Paul Engel
http://www.aceshardware.com/
ANSYS 7.0 ClawHammer Benchmark Results (CPU)
By Brian Neal
Monday, December 2, 2002 1:54 PM EST
Thanks to Armand for pointing me towards some interesting ANSYS ClawHammer benchmarks. The CPU tested is an 800 MHz ClawHammer, presumably with a 256 KB L2 cache. The system was configured with 512 MB of memory (speed unknown, PC2100 or better) and a 7200 RPM 60 GB IBM Deskstar 60GXP (ATA-100). This configuration is somewhat minimal given that most of the other systems listed are configured with 1 GB or more of memory. As a result, it's difficult to make a clean CPU-to-CPU comparison without knowing what role the memory size and also disk speed play in determining overall performance.
Nevertheless, ClawHammer/SledgeHammer benchmarks are few and far between and this is one of the more interesting sets of results encountered so far. In the first table, you will find the system configurations, with links to the results for each system leading to the second table.
System Configurations
System CPU CPUs Used Memory Disk
AMD ClawHammer 797 MHz ClawHammer 1 512 MB IC35L060AVER07-0
Evo W6000 Wkstn 2.2 GHz Pentium 4 Xeon 1 1 GB COMPAQPC MAN3184MP
Dell Precision Wkstn 530 2.8 GHz Pentium 4 Xeon 1 1 GB FUJITSU MAN3184MP SCSI
Dell Inspiron 8200 1.7 GHz Pentium 4M 1 1 GB HITACHI_DK23DA-40
HP rx5670 1 GHz Itanium 2 1 24 GB NA
HP rx5670 1 GHz Itanium 2 4 24 GB NA
IBM Z-Pro 733 MHz Pentium III Xeon 1 512 MB IBM-PSG Tornado 18
ANSYS 7.0 Benchmark Results
CPU Seconds bmslsbpl bmslsbpm bmtnsbpl bmttnbpm bmslsbpx
797 MHz ClawHammer (1) CPU Time 276.219 147.609 999.062 455.078 347.500
Elapsed Time 302 203 1039 501 639
2.2 GHz Pentium 4 Xeon (1) CPU Time 211.781 102.906 634.156 290.984 227.312
Elapsed Time 251 174 663 343 288
2.8 GHz Pentium 4 Xeon (1) CPU Time 185.875 86.484 482.297 241.469 191.125
Elapsed Time 225 150 511 291 255
1.7 GHz Pentium 4M (1) CPU Time 278.751 132.250 803.435 399.835 300.993
Elapsed Time 331 247 859 486 395
1 GHz Itanium 2 (1) CPU Time 170.840 102.130 822.810 406.850 199.880
Elapsed Time 175 104 825 410 204
1 GHz Itanium 2 (4) CPU Time 280.570 147.280 890.980 603.940 304.770
Elapsed Time 93 65 570 254 156
733 MHz Pentium III Xeon (1) CPU Time 685.219 315.953 1969.328 744.297 635.859
Elapsed Time 765 443 2002 842 934
The results are presented both in terms of CPU time and elapsed time. The CPU time is the time spent on the CPU and the elapsed time is, of course, the "real" time. As you can see in the table above, there are situations where the CPU time will be higher than the elapsed time, as is the case with multiprocessor systems where the benchmark is running on multiple CPUs (and accumulating time on multiple CPUs as a result). With the ClawHammer's low clockspeed and limited system memory taken into account, it is performing rather well in many of these benchmarks. It significantly outperforms the 733 MHz Pentium III Xeon (the only other system with 512 MB of RAM) in all tests, though that system is only configured with SDRAM. It also outperforms the 1.7 GHz Pentium 4M in two out of five benchmarks and comes relatively close to the 2.2 GHz Pentium 4 Xeon in those same two tests. The 1 GHz Itanium 2 wins almost every benchmark except for one, while the 4-way Itanium 2 system improves on the uniprocessor results considerably.
For further analysis, take a look at this post by manfred.
Re:"how do you account for all the time it has taken to bring this new core to fruition?"
The answer is easy. Hammer is a poor design and AMD's SOI process doesn't work. The two factors combine to produce Hammer slow clock speed and poor performance, so poor that the Hammer can not be released.
AMD Has Fallen and It Can't Get Up
http://www.thestreet.com/_yahoo/tech/kcswanson/10054856.html
Re:"AMD's inferior CPUS aren't selling - and 2000 more AMD's folks are being throttled."
Note that it is the troops that get throttled. Where's Mine Jerry still rides off in his limo and flies the company jet to his Beverly Hills office.
Re: "So, basically, to keep the P4 from throttling, you should not use power intensive applications, or if you do, only do so on cool days."
You make a huge leap to assume that the opposite of "a high ambient temperature environment" is a "cool day". All indications are that clock throttling ocurrs only in very unusual conditions provided that proper thermal design was used for the system.
But when the thermal limits are encountered clock throttling operates as a fail safe to protect the processor. The practical s that cause throttling may well be a failed fan, disconnected heat sink or a debris clogged air intake.
Your rant is nothing more than "the sky is falling" chatter and you provide no example of a condition when clock throttling ocurrs.
Re:"Just how in the name of common sense can AMD bold faced announce stepped up attacks on Intel and Intel's markets while massively scaling back their (AMD's) expenses and headcount while they are already losing money and market share with their current expenses - before massive cutbacks?"
Perhaps AMD found some magic mushrooms in the rain forest. That is the only logical explanation.
Elmer -
Re:"Better than tax free! Doesn't it actually result in a tax credit for Dr. Moore?"
My understanding is that there is a tax deduction, a bit different than a tax credit, based on the value of the stock transferred to the Foundation, a charity. But the value of the stock is so great that Gordon will never get to use up the deduction on his tax bill.
But a tax dedution is not the reason that Gordon Moore established the Foundation.
Re:"I remember Gordon selling a bunch of his stock in order to fund this foundation."
Not to be picky but Gordon actually transferred the intc stock to the Foundation. The difference is significant because it is a tax free transfer of the gross value of the stock.
Over time, the Foundation sells the intc stock to fund its activities.
Re:"So what is this charge going to do to AMD's already weakened cash position?"
Probably, the charge will have little current effect on the cash position. Much of the charge will constitute write offs and will not affect cash. Some of the charges will constitute reserves against future cash expenses such as severance packages and that won't effect the cash position until the money is actualy paid.
Re:"2600s and 2700s are shipping and in boxes (alienware)"
That's strange. The fastest AMD box at Alienware on their product listing is a Model 2400.
Do you have a source for your statementor are you just guessing that Alienware is actuall shipping the product?
Googlegear has 3 GHz available in Retail boxes.
Only AMD uses Silicon on Paper (SOP) technology
http://www.googlegear.com/jsp/ProductList.jsp?ThirdCategoryCode=010408
Are the new processors a P4 based design?
If we assume that a fab is depreciated over 3 years, and assume 7500 wafer starts per week, that comes out to over $1700 per wafer for depreciation alone. So $800 covers the cost of the raw wafer and labor as well?
You have the concept right although I don't have sufficient knowledge to say whether the actual numbers are correct. But as a first approximation they are good.
Is 3 years too short a time for depreciation?
I think that 3 years is the depeciation schedule for fab equipment. It used to be longer nut the Semic mfgs pushed for the 3 year period on the basis that fab equipment is obselete as a practical matter after 3 years.
The depreciation of equipment is not a cash issue as Andy Bryant pointed out in the Webcast. The money was spent when the fab was built. Bryant went on to say that in times of excess capacity depreciation costs will impact earnings but not cash. On the other hand the lack of capacity in a recovery costs a lot of cash so it is best to have some excess capacity in a downturn. Intel manages business to maximize cash and does a good job per Bryant.
Bryant's comments were in response to an analyst question dinging Intel about capitol spending that was too high. The thrust of Bryant's answer was that the length of the downturn could not have been forecast and the capacity will be available when the upturn occurs.
The wafer cost does include depreciation for plant and equipment and is a large component of the ~$2500 for a 300mm wafer. The raw materials and labor that go into wafer cost is a small share of the total.
R&D is taken as an expense when it occurs. The Banias R&D for the most part is taken before the processor goes into production although some fine tuning of the processor will be ongoing.
Re:"My approach is to look at the financial statements of INTC,"
You have totally the wrong approach for determing the contribution of the Banias processor.
Look at Revenue less Cost of Goods to determine the the CASH contribution of the Banias processor. The G&A change is trivial and the R&D cost was expensed yesterday. It is a cloud nine exercise to determine what R&D costs to attribute to the Banias processor anyway.
The cost of a 300 mm wafer is about $2500. The die size will produce about 300 plus good die per wafer. So you end up with about $8 per die. Test-assembly and packaging is about another $15 to maybe $20 for a toal COG of about $30 on the high end. Actual COG is certainly much lower. As Elmer said, probably less than $25.
The CASH contribution of the Banias processor will be very, very good. And Intel manages their business to preserve CASH per Andy Byant in the recent Webcast. And they do that very well.
Re:"FWIW here are some die cost estimates from the MPF."
There is a hint in the data that the estimates are not actually die costs but rather finished product costs (die, test and assembly plus package) The notation on Athlon w or w/out module suggests this.
Re:"The 2400+ and 2600+ product only appeared today and were not available when I looked on Tuesday. Very few units are currently available..."
I called Upgrade Source and none are in stock. The earliest available date for stock is the 24th. IIRC that is the available date stated by AMD but the reality may be different
I suspect that ASPs for AMD are close to the $50 range.
I was looking for some components and found some startling price information for AMD and Intel processors.
Thhe 2400+ and 2600+ product only appeared today and were not available when I looked on Tuesday. Very few units are currently available and they would have had virtually no impact on Q3 sales. Probably the supply is so small that they will have negligible impact in Q4.
Second, the Atlon XP 2200+ is the fastest Athlon that is available as a practical matter. On Tuesday, the 2200+ was the fasted available period.
The Athlon 2200+ is selling at about the same retail price as the Pentium 4 1.8 GHz ($166 vs. $162) !! AMD's prices are of course lower than retail prices.
So the fastest AMD processor, actually available, sells at about the same price as the 6th fastest Intel processor.
Athlon XP 2600+ - $308 ... Pentium 4 2.8GHz - $519
Athlon XP 2400+ - $199 ... Pentium 4 2.66GHz - $412
Athlon XP 2200+ - $166 ... Pentium 4 2.53GHz - $257
Athlon XP 2100+ - $138 ... Pentium 4 2.4GHz - $209
Athlon XP 2000+ - $112 ... Pentium 4 2.0GHz - $179
Athlon XP 1900+ - $ 99 ... Pentium 4 1.8GHz - $162
http://www.upgradesource.com/
Re:"For the first time, I'm thinking that AMD may actually be bankrupt next year."
AMD will hit a real trouble zone even before they reach bankruptcy when they start to violate the terms of their loan agreements. Watch for announcements that they have renegotiated the terms of their credit agreements.
Re:"I think that other charge was for different processors (P4?), and that this portion of the case is just Itanium. Anyone with a better memory feel welcome to correct me."
Correct. The Texas case concerns the Itanium and the Alabama case was about the P4. Note that the Intergraph case has a strange twist. Intel already has a license based on the prior settlement. The only thing to fight about in TX is the size of the payment. The Reuters article covers this and so did the original PR on the settlement. $150 million was at stake in the trial and Intel lost.
If Intel appeals and wins, no $$ are due. If Intel appeals and loses it will owe Intergraph another $100 million for $2 a total of $250 million. If Intel wins it owes nothing. Therefore the appeal decision is almost a double or nothing bet.
Re:"It was then that news came in that Hector was buying shares. I couldn't believe he'd put his own money on the line..."
The Purchase by Ruiz was little more than a calculated attempt to wave the AMD flag and obtain some positive PR. His purchase amounted to about $150 K and that is little more than a Las Vegas weekend for Ruiz. He received $6.5 million in Pay -- no exercised options -- in 2001.
http://biz.yahoo.com/p/a/amd.html
AMD can easily pay back any losses to Ruiz with a little extra compensation. Another perspective is that the limo and driver for Jerry "Where's Mine" Sander's added more than that amount to his compensation per SEC reports. And no numbers were split out for the Beverly Hills office established for Jerry's convenience. That was just rolled up as a "necessary" business expense for AMD.
Re:"To me, "a standard" means it must be integrated."
Intel produces a reference design for the laptop board that virtually all makers use. In this respect, if Intel incoprpates 802.11 it becomes an industry standard.
Of course an OEM can develop their own but few will try.