The processor architects always love to add features and they hope when circuitry is done that they will come with the silicon area budget. However when the circuitry to implement the processor takes form, reality hits. The architects start removing features and negotiating for a bigger silicon budget.

Generally, they get very little increase in the silicon budget. If the process defect density has improved more than expected, the yield numbers go up for a given die size. So the architects may get an increase the silicon area budget and can drop fewer features.

The result: processors always go on a feature shrink before the design is released for manufacture.

Moores's Law gives the architects the ability to add features on the next process generation.

Use of the Pentium 5 name would also allow P4 processors to compete with the Opteron ( a 32 machine in fact, with possible 40 bit memory addressing in the future)and the Athlon, all flavors XP to A64, could compete with Celeron processors.

With an R&D budget of $2 Billion per quarter

The actual facts have never been important to you but Intel is currently spending $1 billion per quarter for R&D. A quick check of Intel's financials shows this. It's an easy piece of work.

http://www.intel.com/pressroom/archive/releases/20030415corp.htm

The conclusions from your dream world are, as always, not worthy of comment.

somehow work the 'g' into the name - thereby making Centrino look like old technology

AMD can easily do that by branding it the Guano technology. There are however some negative aspects to that solution.

The Inquirer reports that AMD finds Opteron sales are lower than expected.

Stocks building, Opterons gathering dust bunnies

http://www.theinquirer.net/?article=10308

Page listings of boxed Opterons on Pricewatch do not count as sales. Tier 1 server OEMs do NOT buy proccessors from Pricewatch sources. IBM, HP, Dell and the rest of the Tier 1 OEMs buy buy directly from the manufacturer... and they don't buy processors "in a box".

Of course, the fanboys DIYers and white box makers do buy from Pricewatch sources. Some wave of acceptance that group is.

Opteron: the server processor that makes a good kiddy game-machine.

Opteron has great infrastructure support with new systems being shipped almost daily

Ship 2 today none tomorrow. Profits are just around the corner.

Not a good way to run a successful business.

The 2- and 4-way servers must be selling as hot cakes

Is that why AMD announced a $100 million revenue miss for Q2?

Perhaps the reason that there are no 32 bit benchmarks is that the IPF is not used to run Quake or Doom, unlike the Opteron which is a nice fit for that market as a mid level performer. Maybe, someday if a 64 bit game gets released, the Opteron will be able to offer leading edge performance in the game arena.

You see, the IPF is targeted as an enterprise server and uses its new architecture to achieve performance for the targeted function. That is the arena where the newest member of the IPF outperforms all other processors.

What, by the way, are the Quake Benchmarks for Power 4? We know what they are for the Opteron.

AMD Expands Options for 4-Way and 8-Way Enterprise Computing

Well now we know the identity of feeble announcement by AMD to silence the thunder of the Madison launch will be.

Oh Dan3, what are the names of the 64 bit ops that will run on the Floperon?

Is this CEO Jerry Sanders ??

http://www.dilbert.com/comics/dilbert/archive/dilbert-20030627.html

A company can go broke before it has zero cash on hand. An example is United Airlines.

United reported in Q3 2002 that it had $1 Million in cash and $344 Million in restricted cash. In December 2003, it filed for bankruptcy.

When the cash available to pay bills gets too small, a company has to starts using cash to pay its important bills and lets its less important bills slide on the hope that things will get better. In the case of United, critical suppliers like jet fuel vendors were about to require cash on delivery terms for fuel. Once United declared bankruptcy, it in effect started over without debt payments and without paying its past due bills. It could then afford to pay for its fuel.

The fuel vendors were happy because there new fuel bills would get paid as part of the bankruptcy. The old bills would get a place in line along with the other creditors.

If United can get its act, together then it can reorganize and emerge from bankruptcy with a new start. If not, then United will be sold and the money distributed to creditors. United has been having trouble developing a viable business and there are reports that they may be liquidated.

Many of AMD's debt obligations have financial conditions that must be met. At some level of low cash -- $500 million perhaps -- the conditions will be activated and the creditors can then accelerate the repayment terms. If this happens, it will not be possible for AMD to pay its bills and they are in deep trouble.

Another possibility is that at about $500 million cash, AMD will have trouble making payments for its operating expenses and debt payments.

The condition for bankruptcy of a company is an inability to pay debts as they come due. At that point, the company can declare bankptcy like United Airlines did or its creditors may force AMD into bankruptcy. In my view it's more likely for the creditors to take this action than AMD so that the large creditors can protect their security interests.

Speaking of a diaster, do you think that AMD will make it through Q3 without moving into bankruptcy while losing about losing about $200 million a quarter with the current burn rate? This sad excuse of a company has only $800 million in cash and a debt load of $1.79 billion.

It is is quite likely that creditors will push AMD into involuntary bankruptcy. At about the $400 million cash level AMD will have trouble meeting its payroll.

In fact, I'm here to tell you that I don't see a single consumer-/small-business-level binary architecture that will have a longer longevity than AMD64.

Does the IA32 binary architecture that has been around since 1985 count as "at least a decade" ?

The life of AMD and its AMD64 may be very short lived.

AMD has reported negative cash flows totaling $1.1 billion in the past four quarters. It will add to that dismal performance in Q2. AMD can not continue for long to loose cash at that rate before it will enter bankruptcy.

AMD's Q1 2003 earnings report shows about $800,000,000 in cash and $1,790,000,000 in debt. Now, AMD has claimed they were on their way to cutting their costs so that they could break even at $800,000,000 in quarterly revenue.

If AMD only has $615,000,000 in revenue this quarter, and 800 million is break even - then AMD will have a net cash loss of about $185,000,000 !!

That should bring AMD's CASH position down to $615,000,000 - and their debt will still be $1.79 Billion. That's a 3 to 1 debt to cash ratio, and the cash is dropping at about $150 million to $200 million per quarter.

My guess is that AMD's creditors are going to panic when AMD's cash drops below $500,000,000 - and that should occur in Q3 of this year - just a few weeks/months away.

Like AMD, Apple is going to milk the 64-bit marketing bullet to its fullest. Unlike AMD, Apple might actually do a good job at it.

Also, unlike AMD Apple actually has some 64 bit apps to run on its system.

Are you suggesting that it is inappropriate to counter false information presented to this board? If so, please state as one of the rules of conduct so it is not the subject of sarcastic attack by the moderator.

What is missing is any sense of certainty of earnings or even survival in the minds of investors

Here is a certainty. AMD will lose money in Q2. And its cash will decrease.

Continued losses and continued cash decrease will result in AMD's inability to pay its bills when they come due. That is the recipe for bankruptcy, or if you prefer BANKRUPTCY.

Because whenever Microsoft advertises 64-bit retail WindowsXP</I>

Where did you get the Windows XP will support support AMD64? Microsoft has indicated that it will introduce AMD64 support in in Windows 2003 Server sometime in 2004. There is no indication that Windows 2003, the retail version will support AMD64.

At the retail level, a Server edition of Windows will not be installed on the machine. Retail does not including custom spec white boxes.

I think you may be off the mark here Semi. After a quick look, it seems like the Trade Act provides supplemental unemployment benefits if there is a downturn and there is an effect on domestic production due to imports. Another clue is that Intel joined in the petition to get the benefits, they were not the target of the petition.

Definitely could be wrong but that it is what it looks like after a quick look. If someone is curious enough to find an authoritative answer, it is probably available here:

http://www.findlaw.com/

A needlessly scatalogical troll.

Read the source message that generated the reponse. If you have a complaint with that message too then you have a point. If not you are a hypocrite.

However the source message did not generate any critical response from you the first time so the evidence is that you are practicing hypocrisy.

The spec on the HP system sure makes use of the wonderful 64 bit memory addressing of the Hamster. The max memory for the system is 1 GB and the standard memory is 256 MB. It just goes to show that the Hamster is a 386 clone.

I bet there are some diapers going on the AMD fan boys when they see that.

What makes X86 servers so appealing is their complete compatibility with X86 desktops and their moderate costs.

Xeon has 80% plus of the total market because of that reason and the extensive infrastructure Intel provides. AMD provides a junk processor with no infrastructure and only the Tier 4 outfits no one has heard about are using a few -- very few -- Floppers.

Note that Sun chose Xeon processors for their new Linux server model -- NOT Floperon.

It has been explaiined several times and you just don't want to listen.

The Pricewatch quotes INCLUDE GREYMARKET PARTS, excess inventory from OEMs and occasionally distis. The presence of Itanium reflects closing excess inventory as the new generation Itanium II is entering the market.

It is that simple.

Another big hint: Tier 1 OEMs don't buy processors for server systems off Pricewatch. That is the reason that Floperon appears so much on Pricewatch -- OEMs aren't buying and DIYers and fan boys are buying.

I´m sure that if that is the case, and you behave accordingly

Oh wonderful, now we get a one-man behavior code to stifle all dissent.

Dan3 comes on the air with his usual deception composed of half truths and outright falsehoods and anyone who points out the problem gets stomped. No attempt to control the behavior of deliberate deception is made.

The problem with one-man codes is that when one is wrong they are often very wrong. The only solution is to vote with your customer $$. And that is why I have declined to send my $$ to iHub.

The only shortcoming of the Linux box is with large simulations where it runs into the memory limit.

Do you use the 36 bit extended memory capability of the dual Xeon machine? From an inexpierienced view, that would seem to solve your memory issues to a large extent even though there is some extra overhead when using the extended memory.

Carl's beef is that he bought on heavy marign. This was great leverag when the stock was going up. When the stock fell and margin calls started, he learned that the leverage works on the way down too.

So now it is the fault of Intel management that he played the margin game and lost.

Another way to "save" profits for shareholders is to not only eliminate stock options for employess but to eliminate salaries. Of course the employees might not stay very long and the profits might disappear pretty fast.

Options are one element of compensation package and when granted to a wide range of employees can provide a powerful incentive to deliver.

A good discusion of the Pentrium-M and the Centrino technology

Unfortunately, Intel has confused the situation by introducing two brands, Pentium M for the processor itself and Centrino,

The muddled marketing message should not obscure the fact that the new processor is a winner.

On Mar. 12, Intel finally introduced a processor designed from the get-go for mobile use. It turns out this makes a big difference.


http://www.businessweek.com/magazine/content/03_12/b3825030.htm

MARCH 24, 2003

TECHNOLOGY & YOU

Intel's Laptop Chip Is Going Places

The Pentium M will lead to smaller, speedy notebooks that boast longer battery life

Although Intel (INTC ) chips have been running portable PCs for nearly 20 years, the company has never designed a processor specifically for mobile use. Every laptop chip, from the 8088 of the mid-1980s to today's Pentium 4-M, has been a modified version of a processor originally developed for desktop use. On Mar. 12, Intel finally introduced a processor designed from the get-go for mobile use. It turns out this makes a big difference.

Unfortunately, Intel has confused the situation by introducing two brands, Pentium M for the processor itself and Centrino, a name used on laptops that use the Pentium M along with an Intel chipset and Intel Wi-Fi wireless network adapter. Moreover, Intel must overcome years of marketing in which it promoted the clock speed of its chips. At 1.6 GHz, the fastest Pentium M sounds slower than the 2.4 GHz Pentium 4-M it supercedes. But clock speed only measures how fast a processor can execute a single instruction, and many other factors determine overall performance.

The muddled marketing message should not obscure the fact that the new processor is a winner. Although it will run all software written for the rest of the Pentium family, the Pentium M is not derived from any earlier product. It was developed under the code name "Banias" by a research team in Israel with the goals of reducing power consumption and heat output while optimizing the processor for wireless use.

The immediate payoff for using the Pentium M is some combination of smaller and lighter notebooks and longer battery life, without sacrificing computing power. Laptop makers report battery life gains of from 10% to 40% in designs where a Pentium 4-M or III-M is replaced by a Pentium M. For example, Gateway has seen its 450 laptop go from a bit over three hours to five with a standard battery. At the same time, says Jay Parker, senior marketing manager for Dell Computer, "we are finding that a 1.6 GHz performs the same or better than a 2.4 GHz Pentium 4-M."

Next week, I will take a closer look at some of the first Pentium M products. Some laptop makers' first models will use existing designs, so the benefits are limited mainly to longer battery life. The big payoff will come with new mechanical designs that take advantage of the processors' lower power and cooling needs. For example, the new Sony Vaio Z1A, the IBM ThinkPad T40, and Dell Latitude D600, all announced at the Mar. 12 Intel launch, will be significantly thinner and lighter and deliver more battery life than the products they replace.

As time goes on, the capabilities of Pentium M products will expand. One likely area is power management options. For example, Mark Hanson, general manager of Sony PC marketing, expects options that let a laptop user trade a bit less processing power for assurance that the cooling fan won't turn on noisily during a presentation. With any new processor, battery life typically improves as manufacturers learn to tweak settings.

Though computer makers love Pentium M, they're not thrilled with how Intel is going to put most of its marketing effort behind the Centrino brand, leaving it largely up to the laptop makers to establish the Pentium M. Many manufacturers believe this branding strategy is designed to coerce them into using Intel's Wi-Fi network adapters, which creates two problems. One is that some corporate buyers insist on specific alternatives, especially Cisco's Aironet adapter. Another is that Intel offers only one flavor of Wi-Fi, 802.11b, at a time when two faster alternatives are hitting the market. Some corporations want adapters that support both 802.11b and the higher bandwidth 802.11a, something Intel won't offer until midyear. And for consumer products, laptop makers are getting interested in the high-speed 802.11g, to which Intel is not yet committed.

This will create confusion. For example, Dell's Latitude D600 and D800 will carry the Centrino brand. But buyers can choose a Dell 802.11g wireless adapter (which also works with existing 802.11b base stations) at no extra cost. Or they can opt to pay extra for an adapter compatible with a, b, and g networks. IBM will offer ThinkPads both as Centrino products with the Intel adapter and as Pentium Ms with Cisco's version.

<bThis muddle is unfortunate and unnecessary. But there is a positive side. As part of its marketing push, Intel is spending heavily to expand the availability and awareness of public wireless networks. That should be a big benefit to everyone, whether they use Intel's Wi-Fi adapters or not.


By Stephen H. Wildstrom

Cash used to purchase an investment or pay for plant or equipment is not an expense against earnings. If fact, large companies are required to use accrual accounting instead of cash accounting. The rant that options MUST count as an expense against earnings totally misses the real issue.

Financial reports should present a true an accurate set of information to allow investors to evaluate the performance of a company. If the issuance of options is disclosed, as it should be, the information is then available ofr the investing audience.

Well now that you mention AMD's failures...

The Performance Rating fraud has stretched to the breaking point just as it did for Cyrix.

http://www.extremetech.com/print_article/0,3998,a=36734,00.asp

Analysis and Conclusion

The PC Mark memory test results on the previous page provide a pretty strong indication that there's a delicate balance between clock rate and cache size.

When Intel introduced the Northwood CPU, with its larger cache size, the slowest Northwood was 2GHz -- the same as the maximum clock rate of the older Willamette core. There was no question as to which Intel CPU would be faster.

Here, AMD is introducing a product that clocks slower than its previous top of the line, but with more cache. The net result is that applications that are sensitive to cache size may run better on Barton, but apps which like higher clock rates (and have tight loops running in L1) will run faster on the 2800+.

This is all somewhat moot, as AMD will not be shipping a 2800+ Thoroughbred at all, opting for the slower-clocked, Barton variant. However, it is interesting to see the tradeoffs here.

Of course, you might look at the Barton 2500+ and note that it clocks more slowly than the Thoroughbred 2400+ -- but that's not apples-to-apples either, as the 2400+ only supports a 266MHz frontside bus.

The real contradiction, though, is that the Athlon XP 2700+ clocks at 2.16GHz, while the 2800+ will clock at 2.083GHz. It's likely the 2700+ will perform better in some CPU-intensive apps than the 2800+. This leaves us scratching our head over the entire product naming scheme.

All of this indicates that AMD is playing a tricky and delicate balancing act with its product numbering scheme. As we've seen, the 3000+ isn't superior to the 2800+ in every case. When compared to the P4 at 3.06GHz, the results are also somewhat mixed -- AMD is faster in some cases, Intel in others.

I suspect the facility would at best be a test and assembly plant if the rumor is true. No way that Intel would build a fab in India.

My recollection is that Intel said it would only go to 0.090 masks on criitcal layers because of high mask costs and would use 0.13 masks for other layers.

When do you expect all the "prelease" versions of OS, drivers and hardware to actually get released for AMD 64?

Even when final release arrrives in maybe 2004 the performance improvement will not likely be adequate.

AMD took a charge (one of many charges) last quarter and a few analysts asked about it - it was a $47 million write-off, I believe - that AMD tossed into a category of R & D write-off.

That amount was almostly certainly what they spent at UMC.

Re: "I cannot for the life of me figure out how you got "benchmarks", plural, out of that article."

You could look at the title of the article* but that would be kind of hard on you wouldn't it.

*AMD Athlon-64 Benchmarks and more

The leave one wall practice is common in California. When you remodel instead of build new the permit requirements and costs are much less. There also are favorable real estate tax consequences for the remodel.

The definition of fraud fits your work very well.

Fraud: A piece of trickery; a trick.

http://dictionary.reference.com/search?q=fraud

You omit an important fact. The test method disabled the case fan.

Concealing this facts and considering a disabled case fan a "typical circumstance" results in a crude and gross distortion of the meaning of "typical". The next time you construct a fraud you are less likely to get caught if you hide the evidence.

Testbed configuration looked as follows:

o Intel Pentium 4 3.06GHz CPU;
o GlacialTech Igloo 4310 Pro cooler;
o ASUS P4PE mainboard;
o 256MB PC2100 DDR SDRAM;
o GeForce4 MX440-8x graphics card;
o IBM DTLA 15GB 7200rpm HDD;
o SoundBlaster Live! Value sound card;
o CD-ROM 24x drive;
o FDD;
o InWin J-536 case (we turned off the system fan);
o Windows XP Professional operation system.

There was a patent lawsuit by Broadcom against Intel after Intel filed its trade secret law suit. That recently had an appeal decision that was reported here. Maybe that is what you were thinking about.

The GPS locating "feature" has a very nasty consequence. When you have a cell phone turned on your every movement can be tracked.

At least for now the GPS tracking feature can be disabled. How long before the government requires the "feature" can not be deactivated? Then they can track everyone's movement in the Total Information Awareness Database the feds are constructing in the interests of National Security.