is...a Libertarian
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.
wbmw, we are not talking about centrino notebooks. I am not sure what you believe to be the "greater platform integration" that Intel will offer in the next 3 years -- while changing the basic design of Xeon -- but I think Sun will want to keep its future in its own hands (i.e. Sun will design its servers based on the components it deems best). If Xeon is a superior product, it will be offered, otherwise including Intel branded chipsets or motherboards is just so much fluff.
I agree with your conclusion wrt Sparc (though I think it applies equally to all server chips other than perhaps Power). However, I disagree with your prognostication on the readoption of Xeon, unless there is a significant disparity in performance between Xeon and Opteron going forward -- there is no real need. Additionally, I don't think the Xeon, whatever it may be in 3 years, will be anything based on P4.
Sun is trying to sell a car badged Sun and telling its customers not to worry who makes the drivetrain. If their customer base buys into this message, they will be successful.
I understand your question now. I think we won't really know the answer until some indication of the performance of the chipset is available.
Smallpops, perhaps you care to elaborate on your question - because as it is written it has been asked and answered.
I've been looking at both stocks myself. They seem like cheap options plays that don't have time decay. I think Cray is the more likely to pay off as they have a history with the government which allows them to get more access to grants and orders (I know this is not how it is suppose to be, but that's the real world).
If I buy in, I would really just be looking for a couple of bucks, before selling out. Longer term I think both companies are poor investments, because they are vulnerable to competition and the vagaries of government funding.
Given all the recent announcements around the XD1 architecture, it looks like the Octigabay acquisition was the right move for Cray.
Now if they could only turn a profit.
OT: but of interest to mozilla guys
I got quite a laugh from this MS guy who appears to be quite serious.
http://www.danamania.com/gl/article.php?story=2005020905570523
Microsoft may use the Pro/Home differential to get revenue from the upgrade. They have said previously that they are not looking at the 64-bit migration as a revenue opportunity, but by not offering a Home variant, I think they are backdooring the customer. How many celeron customers buy Pro? Even if it is on a new computer, Microsoft makes more from Pro than from Home.
Keith if you look at B2L's IDC numbers and the Sun slides, I think you can interpret their comments to mean x86 server sales grew by 160%. Now last year that was all Xeon32, but given the reported trends over the last couple of quarters (also from IDC), you can assume that most of the sales were Opterons.
I would estimate that they sold around 10K Opteron servers in the last quarter (see analysis on SI).
http://www.siliconinvestor.com/readreplies.aspx?msgid=20949660
wbmw, Suns 4th quarter x86 servers increased by 160% YoY.
http://www.sun.com/aboutsun/investor/earnings_releases/Q205_SLD.pdf
@chipguy Here's what Sun, the great white hope of droids everywhere, did last quarter in x86 sales:
I think you better read that article more carefully.
The company shipped 13,002 servers with Intel's Xeon processor in the first nine months of 2004, of which 58 percent used Linux. For Opteron servers, Sun shipped 10,566, of which 73 percent used Linux. For x86 servers overall, 65 percent shipped with Linux."
Nothing about last quarter. But if you are interested they did increase x86 server sales by 160% last quarter on YoY basis.
http://www.sun.com/aboutsun/investor/earnings_releases/Q205_SLD.pdf
I agree. Most people I know leave their computer on continuously and the only time a reboot is necessary is when the software crashes by way of a BSOD. The builders of the test systems should have taken more care.
fnf,
I agree that this stock has a lot more near term downside risk than upside potential -- especially if Intel says something the market does not like. I too am looking to buy back in, but I will wait until after the AMD conference call to assess what my entry price might be.
IB
I think the generally held view is that Intel was tired of having cloners compete with them and decided to resolve that problem by creating a proprietary architecture. Their intention was to move the world over during a 32-64 bit conversion (servers first). IPF was the vehicle. They recruited MS as a chauffeur.
Now like all smart planners, Intel management had a backup plan which they would not reveal, lest it put a bullet in the head of IPF (for general purpose computing). Intel kept denying or at least refusing to acknowledge the existence of x86-64 (both the discarded variety and the AMD clone) until it became apparent that the market preferred this architecture and that by not offering a competitor they risked losing their dominance to AMD. Only then did they acknowledge they had an x86-64 chip.
Intel would have had legal liability if they made statements regarding Yamhill -- so they didn't. It was left to the journalist, enthusiasts and reviewers to decipher Intel's message.
Revisionist history? Why don't you put "intel yamhill backup" into google and see what you come up with. To get you started here you go.
A Backup Plan for Intel
Itanium alternative reportedly under development.
While Intel Corp. has spent the past eight months touting its Itanium processor, it reportedly also has recently been developing an alternative 64-bit chip it can turn to should the Itanium flop.
Analysts have long suspected that Intel was developing this alternative, but recent reports of the work surprised some Itanium users and could undermine current sales efforts and indirectly help rival Advanced Micro Devices Inc.
Details about Intel's alleged covert chip development surfaced Jan. 25 in a report published by the San Jose Mercury News. Citing unnamed sources, the newspaper said several senior Intel engineers critical of the Itanium's architecture, developed over seven years at a cost of more than $1 billion, were given the go-ahead by top executives to design an alternative 64-bit processor based on the X86 design used in the company's Pentium line.
Known internally by the code name Yamhill Technology, the alternative 64-bit chip designed for high-performance computing reportedly is being developed at Intel's facilities in Hillsboro, Ore. While Intel spokesman Howard High, in Santa Clara, Calif., declined to comment on the report, industry analysts said the story confirmed what they had long suspected.
"It's really just insurance," said Kevin Krewell, a chip analyst with Cahners In-Stat/MDR, in Sunnyvale, Calif. "But, of course, it's something they don't want people to know about."
Word that Intel is developing an alternative 64-bit processor based on its popular Pentium design could undermine its efforts to sell computer makers and corporate customers on the Itanium, which features an entirely different architecture and requires users to adopt new software to run.
So far, the Itanium has attracted few buyers. Since its release in May, the chip has appeared in less than 1 percent of all 64-bit servers sold, according to Gartner Dataquest, in San Jose, Calif. Intel said it's confident sales will pick up when it releases a faster version of the chip, code-named McKinley, by midyear.
But for many companies, the rumored alternative design based on the Pentium X86 architecture would be more attractive. "A lot of customers evaluating Itanium would say, 'Well, if we could have a compatible chip that does a lot of what Itanium does but is also more compatible with our existing 32-bit software, that might be a better deal for us,'" said Nathan Brookwood, an analyst with Insight64, in Saratoga, Calif. "That could take the wind out of the Itanium sails."
Word of Intel's reported backup plan caught Itanium users off guard.
"I first heard about it just a few minutes ago from a colleague," Mahlon Stacy, system administrator for Mayo Foundation Medical Science, in Rochester, Minn., said when contacted soon after the initial reports were published. But while his staff has already worked to port some of its applications to its Itanium systems, Stacy said the Mayo Clinic still could readily adapt should Intel switch architectures.
"As a research department for a medical institution, we're fairly flexible and can roll with the punches," he said.
But software vendors that have invested millions in recompiling their programs to work with the Itanium could be left holding the bag.
"I can't imagine them doing something like that this late in the game," said an executive at one such company, who asked not to be named.
However, Intel made a similar move in the early 1980s, and the move paid off. At the time, Intel touted the IAPX432 as the basis for its transition from 16-bit to 32-bit processors. "It was going to be a revolutionary processor," Cahners' Krewell said. "But it was about 10 years ahead of its time in terms of process technology and chip integration point of view—and totally flopped."
But Intel had a backup plan, as a team of engineers designed an alternative 32-bit chip based on the company's 16-bit 286 design. That processor, known as the 386, served as the foundation for Intel's highly successful Pentium line.
While Intel's reported secret development efforts may once again prove a wise move, they may also inadvertently benefit the company's longtime rival.
AMD, of Sunnyvale, is planning to introduce a 64-bit processor based on the X86 architecture, similar to Intel's reported alternative design. But for the chip to be successful, AMD needs Microsoft Corp.'s support.
"AMD's X86 64-bit processor has an uphill battle with one really key vendor, Microsoft. If AMD doesn't get support, its chips are only going to be of interest to a relatively small group of Linux hackers," Krewell said. "But if Intel has plans for such a chip, then Microsoft has to do it."
AMD said word of uncertainty about the Itanium within Intel's ranks should give customers pause. "It's not just Intel making a bet here on Itanium, it's CIOs who have taken a chance and bought Itanium machines, or the software developers that have spent $100 million porting software to Itanium, that are at risks," said Pat Moorehead, AMD's vice president of customer advocacy. "After hearing about how Intel's wavering, they are going to be asking the question, 'Hey, what's going on here?'"
http://www.eweek.com/article2/0,1759,1502959,00.asp
Tenchu, does this have some cost benefit that makes it compelling? Are the specs required for high-end / commodity hardware essentially the same? In other words, by creating a common interface will you have to make the commodity infrastructure more expensive to accomodate the tighter tolerances required in a high-end system?
@wbmw I also don't see how it changes Intel's previous commitment to converging IPF and x86 infrastructure through a common system interface.
I guess my question more precisely is what's the point? Is there some benefit to making high-end server hardware common with low-end and desktop hardware? Is the plan to have Integrity servers and Xeon workstations use the same motherboards and chipsets? Is this even possible?
wbmw, I presume with no more windows and no more workstations, Intel's plans for making IPF and x86 infrastructure interchangeable have been binned.
Given its new limited role, what advantage does IPF offer that makes it a more compelling choice than any other RISC chip? In other words, what exactly does it offer that is meaningful enough to overcome customer inertia, HPC aside?
pfosse, kpf, thanks for your opinions. I guess I agree that they could cherry pick the best engineering talent and license the tech (if they haven't already - I seem to recall some sort of licensing deal in the past).
The only real reason I could see to buy TMTA, which may not justify the price, is ownership of the tech. If the tech is good, they could horde it for themselves (i.e. keep Intel away from it -- could Intel be interested in the code morphing tech for IPF?) or use it as a bargaining chip to get other tech.
In the end, if they are interested they probably can get it a whole lot cheaper in bankruptcy.
With a market cap of around $185 million, does it make sense for AMD to buy TMTA? Is TMTA's IP worth that much? AMD does have a deficiency in the low power mobile segment -- perhaps an acquisiton might accelerate a solution?
I prefer the IBM x series. Thin, light, good battery life and great support (if you need it).
http://www-132.ibm.com/webapp/wcs/stores/servlet/CategoryDisplay?categoryId=2072542&storeId=1&am...
AMD confirms raid on illegal CPU labeling in Taiwan
San Jose, Calif.—Taiwanese law enforcement officials have raided multiple businesses that were allegedly "remarking" AMD processors, an official from AMD confirmed to EE Times Tuesday (Jan. 4).
Remarking is the act of purposefully mis-labeling or reclocking low cost, slower speed grade processors as higher performance, higher priced processors and selling them at higher performance processor price points. It is a form of fraud.
AMD spokeswoman Cathy Abbinati said the legal investigation is still ongoing and thus AMD is not releasing details of the raid.
She did confirm however that local Taiwanese law enforcement raided "multiple businesses" simultaneously over the weekend.
She would not comment on the value and volume of processors that had been discovered in the raids but said that previous reports from other press services on value and volume of processors seized "are completely unsubstantiated and without merit."
The raids in Taiwan follow raids conducted by law enforcement in Thailand in November in which several Thailand-based businesses were allegedly remarking AMD processors and reselling them.
Abbinati said AMD investigates all reported incidents of remarking and recommended that consumers purchase their processors only from authorized AMD distributors to ensure they are not purchasing remarked processors.
AMD does not plan to issue a formal release regarding the raids in Taiwan, Abbinati said.
http://www.eetimes.com/semi/news/showArticle.jhtml?articleID=56900438
Obviously nothing today. But what about Unreal Tournament 2005 or 2006? If developers are just starting to exploit the abilities of 64-bit, how long until someone uses them fully?
If you office, email and surf, you will never know the difference. If you rip home movies, play games or do other sundry multimedia, you may come to miss the capabilities.
@Keith Information I´ve received about Gateway´s / eMachines new retail line-ups ("Winter")
Is this information available to the general public, or just to you?
Keith, manufacturers typically build models from specific retailers (i.e. W2828) -- which I might add are not necessarily listed on their website.
Are you saying that emachines no longer intends to build any AMD based desktops? What are you basing this prognostication on?
Keith, I think we will have to differ on our interpretation of this. I appears that the analyst was confused as well and asked for greater clarification, using the term "processor production". Hector answered by saying that 50% of wafer starts would be 90nm K8.
Keith, please see my edited post with the quote (clarification was provided in the Q&A).
http://investorshub.com/boards/read_msg.asp?message_id=4901900
Keith, Hector said in the Q304 conference call that 90nm wafer starts would cross over by the end of Q4.
Question – Krishna Shankar: Yes. In the processor business and the AMD64, did you have good growth both in desktops and notebooks Q2 to Q3? And also going into the fourth quarter, will – did you say more than 50% of your processor production will be on 90-nanometer?
Answer – Henri Richard: I will answer the first part of the question. We had a double-digit growth in both mobile and desktop Athlon 64 processor.
Answer – Hector de J. Ruiz: And in terms of the 90-nanometer, we expect going out of the year to have over half of our wafer starts to be 90-nanometer AMD64.
http://epscontest.com/transcripts/04q3_amd.htm
I know, I was just moving the discussion along -- I really do want to know why it makes sense and I thought someone on the board could shed light on the topic. Your post was just a convenient segue to my question.
I find it hard to believe that the VIA processor + ecosystem is a cheaper alternative than using a low end Sempron, Celeron or for that matter Geode.
Why does this make sense?
Keith, Sun offers RH and Novell on all their x86 systems, preinstalled. They also certify, but do not offer preinstalled Windows. As the Register spin on the ML article points out, Sun doesn't need to buy a Red Hat to sell Linux -- it is in fact one of the most foolish things they could do at this stage.
Sabre, which owns travelocity (if your interested), is the largest GDS in the world. It was previously owned by AMR corp and was the in-house CRS for American Airlines.
Historically, like to other GDSs, Sabre ran on IBM mainframes.
Built on the Transaction Processing Facility (TPF) mainframe databases first developed by IBM for the airlines in the 1960s, these reservation systems, as they exist now, have layers of archaic business rules built one on top of another in 360 Assembler language. Each change has to take into account the previous 30 or 40 years' worth of rules. With a shortage of TPF programmers, those in the workforce are expensive to hire, so any programming changes to the business rules in a fare-pricing structure take a long time to make and cost a lot of money.
http://www.computerworld.com/industrytopics/travel/story/0,10801,68735,00.html
I recommend 3rd party cartridges. They print equivalently, and are a whole lot cheaper. I would recommend ABCInk.com, as I have had personal experience with them and they are reliable. There are, however, many, many vendors to choose from.
I believe what I originally said was let's not forget that HP, co-inventor and largest supplier of IPF by a magnitude, also lost.
I was commenting on the Bull victory -- that is that they are a French company and so they won. No commentary on IPF. You were trying to put forth an alternate explanation.
It's interesting that you are attributing contradictory behavior to two companies, to fit the facts as you would like to see them.
For IBM, their bid included Power based systems, because they thought their own chip was superior.
For HP, their bid included Opteron based systems, because they thought their own chip was inferior.
Now I understand -- Bull didn't win because they were French.
@semi, The point is, if they "thought" to use Opteron, because they "thought" it was the best, they were wrong, and they lost.
What does it say about IPF's future if the co-inventor, and largest vendor, is "thinking" that their best proposal includes Opteron? I can't believe you actually suggested such a concept.
Yes, that would be quite telling, if HP thought there best proposal included an Opteron system. That would be quite a message to the marketplace.
I like your way of thinking.
Semi, let's not forget that HP, co-inventor and largest supplier of IPF by a magnitude, also lost. Makes you wonder.
Then what do you make of Elmer's and SmallPops' analysis?
http://investorshub.com/boards/read_msg.asp?message_id=4805392
http://investorshub.com/boards/read_msg.asp?message_id=4806693
Hasn't IBM historically been the second of third largest supplier of IPF systems -- since the launch of IPF?
In the end, IMO the French did what they are wont to do and that is go with a French supplier.
Elmer, you know we can discuss it on SI. I have started a thread.