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.
"Who in their right mind would buy a "Pentium 4 Extreme Edition"? It's obviously a stopgap measure offering absolutely no upgradeability --"
How about somebody with a P4 and who wants a better PC - without having to buy a new motherboard, a new 64 bit CPU at exorbitant prices that has no 64 bit software?
The P4 EE can be plugged into existing motherbaords - slam dunk upgrade - one that is giving AMD and the fanboys an apoplectic fit.
Great move by Intel, one that seems to have caught AMD off guard and blindsided.
"Intel really deserves that all they´ve done over they years will heavily backfire on them one day. I sincerely wish them all the worst."
Taking your failures rather personally, are you?
Just what has Intel done over the years that should backfire on them?
Intel invented the microprocessor - is that bad?
Intel invented the EPROM- is that bad?
Intel and AMD were founded within a few onths of each other - and Intel has become the largest semiconductor manufacturer in the world - is that bad? - while AMD has perenially lost money and actually has no net profit for its entire 34 year existence - is that good?
Intel took on the entire CPU competition - x86, RISC, etc. - and emerged victorious - is that bad?
Intel maintained profitability for every quarter of the the recent three year semicondcutor depression - is that bad?
I guess Intel has done enormously better than AMD - and that must qualify as bad in your eyes to make you hate Intel so much.
Don't take things personally - figure out how to profit from your invenstments.
And betting on a perennial loser may not be a wise thing to do financially, even though it gives you great pleasure.
-SZ
"Low power Opterons in H1 04 must mean the 90nm process is proceeding on target, so far."
Oh yeah?
Then why isn't AMD selling these 90 nM chips today?
Remember back a year or two ago when certain AMD fanboys were extolling AMD's 130 nM process - which has proven to be a bust.
"Looks like AMD has been a better investment than Intel since the introduction of the Athlon."
But not for AMD.
AMD has lost, cumulatively, $1,219,902,000 since the Athlon was introduced in 1999.
Care to guess how much Intel's profits have been, cumulatively, since the launch of the Athlon?
Remember, even skinny pigs bring some money at the slaughter house.
-SZ
BOGUS process claim by AMD - intentionally lying and misleading:
"the first Spansion product using 130/110-nanometer, second-generation MirrorBit(TM) process technology"
There is no such thing as the 130/110-nanometer process.
Later on in the PR, AMD admits to the bold face lie and fesses up that the process is simply a 130 nM process - and that sometime in the future they will shrink it to a not-yet-ready-110 nM process.
Just another example of AMD releasing a bold face lie - in an attempt to convince the public that AMD has something that AMD really doesn't have.
I mean, how low can this company sink?
And why hasn't AMD pre-announced a bang-up quarter yet? Everybody else is upping their revenue estimates. Maybe AMD's quarter is pish poor and they are in the process of dumping their products at cheaper ASPs to unload their bulging inventory.
Heck - on pricewatch, the unsold opterons are piling up refelcting the dismal adoption rate for that part.
Intel Itanium servers inside NVIDIA headquarters.
They probably had to use these to design their Opteron chip sets.
http://www.guru3d.com/article.php?cat=article&id=72&pagenumber=5
No.
Were you?
"My 3000+ PC and a 1900+ with 19" monitors each use 300w when the monitors are on, 150w when they are not. (I have a wattage meter.) I turn the PCs off."
Of course - that is the common sense solution.
"But if I simply apply that I have 350W 24 hours for 30 days a month, I will get $17 per month if I pay 7c per KWH.
What Prescott offers me is that I'll have to pay probably over $20 per month for my computer. Thanks, Intel, I'm glad I have a choice."
So...you're complaining that an Intel Prescott system would run you an extra $3 per month in electrical fees?
Well...if power consumption is so important to you, why don't you turn your computer off when you're not using it ? ("Almost all of my frinds, and myself, we never turn the home PC off.")
Sort of hypocritical - aren't you?
Bashing Intel for a possible $3/month increase in utilities pales in comparison with a 50% savings by shutting off your PC for 12 hours a day - unless you're married to your PC.
And by the way - did you ever think of the possibility that an Intel Prescott could increase someone's productivity such that a $3/month extra utility cost could be made up every few hours of every day of every week of every month of every year due to the higher CPU performance- which would make the Precott owner more than happy to pay for the extra $3/month utility fees - a walloping $36/year.
And don't tell me about gamers...
-SZ
"the better the performance of 3rd party products is, the easier it will become for an manufacture to go with one Non-Intel solution."
And I'm sure Intel will be more than pleased for Broadcom to sell a few million of these at $30 to accompany a few million Intel sales of a $300 - $600 Pentium M processor.
MEGALOL - all the way to the bank.
"I believe Intel is a bit behind the curve here.-- Royal Philips Electronics (NYSE: PHG, AEX: PHI) today announced its first complete plug-and-play Bluetooth™ semiconductor module--"
You have missed the point.
Intel is deveoping a process technology that integrates both digital and analog capability, which will permit, with the appropriate analog circuitry, the ability to integrate a general purpose CPU as well as the essential RF circuits to implement on board radio communications.
In this regard, they are well ahead of the curve.
If they are successful at this (and that success will not come right away) the integrated processor/communications capability will provide enormous value propositions for these chips.
Intel's foresight, and their willingness to invest the R & D dollars for this long range project, will provide another degree of separation between Intel and its competitors in the future.
-SZ
'Tanglewood' to top Intel chip show
By Stephen Shankland and Michael Kanellos
Staff Writer, CNET News.com
August 29, 2003, 12:38 PM PT
http://news.com.com/2100-1006-5069915.html
Intel plans to describe a new high-end Itanium chip code-named Tanglewood at its Developer Forum conference next month, sources close to the company said. The chip will include as many as 16 processors on a single slice of silicon.
So-called multicore designs that include several processors on the same chip are an increasingly popular way to use the ever-larger amounts of circuitry that advanced manufacturing processes permit, but other companies are slightly ahead of Intel in the multicore business. IBM's Power4 chip uses two cores, as will new high-end chips from Sun Microsystems and Hewlett-Packard, but the Tanglewood design appears to have room for growth well beyond that.
"The multicore approach to building processors is a natural way of taking advantage of Moore's Law," said Insight 64 analyst Nathan Brookwood, referring to the rule of thumb that predicts rapid doubling of how many circuitry elements called transistors can fit onto a processor.
A design that extends as far as 16 cores telegraphs Intel's plans far into the future. "Intel needs to be able to demonstrate there's a road map that's going to keep that up. IBM is out there peddling its road map, Sun is out there peddling its road map," Brookwood said. Itanium got off to a rocky start, but Brookwood said Intel has been able to release new designs more consistently now.
Fitting multiple processors on a single silicon slice, called a die, is difficult with large processors such as today's Itanium 2 models. But Brookwood said it's feasible in future generations if processors share high-speed cache memory that today takes up much of the current Itanium's real estate.
Intel's Itanium line is a 64-bit design geared for high-end servers that need to address vast amounts of memory and be able to keep processing glitches from corrupting data. Its chief competitors are IBM's Power products, Sun's UltraSparc line, and Advanced Micro Devices' Opteron; Hewlett-Packard is gradually phasing out its PA-RISC line in favor of Itanium.
Intel has conquered most of the low-end server market with its Xeon processor family, which, unlike Itanium, is a variant of its Pentium line and can easily run the same software. To bring Itanium into less-expensive machines, Intel plans to release another Itanium processor code-named Deerfield.
Deerfield is scheduled to debut Sept. 8, sources said. According to Intel product plans from July, Deerfield will cost $744, substantially below the price of Intel's other Itanium 2 chips, which range from $1,338 to $4,226.
Improving how many instruction threads a chip can handle simultaneously is another area chip designers are tackling. Intel's Xeon and Pentium line have basic abilities to run two threads, and Intel is looking at ways to add more threads on server chips, sources have said. IBM's Power5, coming in 2004, will permit each core to run two threads, and Sun is aggressively pushing multithreading. Itanium currently lacks multithreading, but Tanglewood will include some features for the technology, sources said.
Intel expects Tanglewood not to consume more electrical power than current Itanium processors, sources said, an important feature that would make it easier to design servers without the overheating that causes data corruption and crashes.
The arrival date of Tanglewood is unclear, though it could come as soon as 2006, the year after the dual-core "Montecito" member of the Itanium family is scheduled for release. Intel has extended the current Itanium 2 designs by adding more cache memory this year and in 2004, but that approach isn't on current plans for Montecito, sources said.
Brookwood said it's possible that Tanglewood would start with a four-core design built with a manufacturing process with 90-nanometer features, then move to eight- and 16-core designs with a later 65-nanometer process.
A 16-core design isn't totally unprecedented for the Santa Clara, Calif.-based chipmaker, Brookwood added: Intel sells a network processor with 16 cores today.
Sun's Niagara processor, due in 2005, will have eight cores when it's released, but those cores are smaller because they lack the complicated circuitry for running a single "thread" of instructions as fast as possible. Sun plans higher-end multicore chip after Niagara that will feature both the multicore design and fast individual thread performance.
Along with the new Itanium, Intel will also discuss "Tiano," a new version of the BIOS (basic input-output system) for servers. The BIOS, which records system settings and acts as an interface between a computer's hardware and software, is one of the oldest, least evolved elements of a computer.
A few years back, a number of companies began to work on the Extensible Firmware Interface (EFI), a standard for replacing the BIOS. "Let's get together and replace 8-bit spaghetti code," is how Pat Gelsinger, Intel's chief technical officer, described the effort. Tiano is Intel's first EFI product.
Other chips on tap
Tanglewood will be one of a number of product announcements at the three-day conference, which runs from Sept. 16 to 18. Intel will also further provide details on "Prescott," the next version of Pentium 4, and "Dothan," a new version of the Pentium M for notebooks.
Both chips will begin to ship this year, said Gelsinger.
"We are going to...have '03 product and revenue shipments," Gelsinger said.
Both chips will be made on the 90-nanometer manufacturing process, which means that the average feature size of the chips will measure 90 nanometers (a nanometer is a billionth of a meter). Overall, this means the chips will be smaller and, over time, less expensive to make.
Many analysts and chip executives have said that the shift to the 90-nanometer process will likely prove difficult for many companies, and even prompt product delays because of the complexities involved.
"There are something like 1,600 or 1,700 process steps involved in making a 90-nanometer wafer," Gelsinger said. "It is stunning, the complexity."
Both Dothan and Prescott, for instance, will also use strained silicon, which involves adding an additional lattice of germanium and silicon atoms into the base layer of a wafer that helps electrons move more freely.
Intel's apparent success with 90-nanometer manufacturing could give it an advantage over competitors like AMD, which won't shift to 90-nanometer for a few months.
At the show, Intel will also discuss:
• Content services and content protection strategies for home networks. Last year at the fall conference, Intel introduced the media adapter, which lets home users hook TVs, stereos and PCs into a network. Now that adapters are hitting shelves, the company is trying to line up deals with content providers so consumers can swap files legally.
Gelsinger indicated that the content deals to be announced at the conference won't revolve around trading protected video or music, but around trading photos or piping Internet radio around the house. Still, the company will make announcements regarding content protection that could lead to the swapping of "high value" content, he said.
Along with this, Intel will also provide updates on reference designs for TV set-top boxes based on its silicon.
• A new chipset for notebooks that cuts down on energy consumption. The Pentium M, found in Centrino notebooks, uses far less energy than the Pentium 4 by better controlling transistors. The new chipset, a group of chips that perform input-output functions, will essentially perform the same tasks.
• Wi-Fi chips for handhelds. In March, the company came out with its first Wi-Fi package for notebooks and is now bringing a similar product to handhelds, which will allow the handhelds to be used as phones. Intel will also disclose technical information on new Xscale microprocessors.
• Progress on integrating radios into processors. Currently, radio transmitters for technologies such as Wi-Fi and Bluetooth come on separate chips. Intel will provide updates on progress to integrate these components into other chips, which will make the radios virtually free. These radios will also be able to automatically reconfigure themselves to adapt to different frequencies.
Intel puts desktop, mobile processors on 90nm fast track
By Jack Robertson
ebnonline
08/29/2003 1:00 PM EST
URL: http://www.ebnews.com/showArticle?articleID=14200046
Intel Corp. is sampling the first processors manufactured on its 90nm technology process -- the Prescott for desktop PCs and the Dothan, an improved version of the Pentium M chip for laptops.
Intel said it is on schedule for "revenue shipments" of both processors next quarter. The Santa Clara, Calif., company defines initial revenue shipments as limited-quantity orders placed by OEMs that use the processors to build enough computers to prepare for a formal market launch.
An Intel spokesman said the 90nm process ramp is proceeding as anticipated. Initial production at the company's D1C 300mm-wafer development fab in Hillsboro, Ore., is already yielding validated processors, he said. Fab 11X in Albuquerque, N.M., is primed to come on line in the fourth quarter as planned, while a plant in Leixlip, Ireland, is expected to start 90nm processing in the first half of 2004.
G. Dan Hutcheson, president of VLSI Research, San Jose, said Intel "has already proven its 90nm process in its Oregon development fab. Now it's just a matter of making the technology work in other fabs."
Analysts said PC makers will wait to provide systems based on the Prescott and Dothan until Intel has ratcheted 90nm production to a level where it can meet volume consumer demand.
"Intel has set itself a very aggressive launch of two new processors using 90nm process technology," said Kevin Krewell, a processor analyst with In-Stat/MDR, San Jose. "They need to build enough units to supply customers ramping up their PC production, and that could take a bit of time."
Still, the company is confident enough of the debut of its 90nm process that special briefings on the technology are scheduled at the Intel Developer Forum (IDF) Sept. 16-18 in San Jose.
At IDF, Intel will also turn the spotlight on the Prescott and Dothan as it tries to drown out archrival Advanced Micro Devices Inc., which is slated to introduce its Athlon 64 desktop processor the following week.
Prescott is a significant upgrade over current Pentium 4 microprocessors, doubling the on-die L2 cache to 1Mbyte with an expected 3.4GHz frequency. Intel has said Prescott will have improved pre-fetch branch prediction for improved performance. The long pipeline of the Pentium 4 architecture can be slowed down when a wrong pre-fetch prediction must be purged and refilled with data.
Prescott will also be the first Intel processor to feature new internal security support, code-named La Grande. The company will provide at IDF the first details of how the new security technique works.
The processor includes 13 Prescott New Instructions (PNI), ranging from storing and loading integer data to new floating-point operations.
Intel is expected to follow its usual pattern of fielding several versions of Prescott to serve various market segments. Next year, the company will add both higher- and lower-frequency chips to the line. Versions of the chips running at 2.8 and 3 GHz are planned as lower-priced entries for the mainstream market. Shane Rau, a processor analyst with IDC in Mountain View, Calif., said the lower-speed Prescotts will be "a natural fallout" from bin testing.
Dothan for laptops will have 2Mbytes of on-die L2 cache, twice that of the current Pentium M. The microprocessor will also have new power management features to extend battery time, Intel said.
The first Intel Xeon server processor to be made in a 90nm process, Nocona, is slated to debut late this year and will also be featured at IDF. The first Xeon MP series made with 90nm processing, Potomac, is scheduled to be introduced in the second half of 2004.
The company will also use IDF to unveil a processor chipset for laptops. Al- though Intel declined to identify it by name, some sources believe the chipset is called Alviso and is designed for the Pentium M line.
Other developments that Intel is expected to discuss at the forum include:
-- A next-generation Xscale processor, code-named Bulverde.
-- New chipsets supporting DDR2 SDRAM in server applications, including the Lindenhurst for one- and two-way Xeon processors and the Twin Castle, which supports the Xeon MP.
-- A new desktop PC form-factor specification.
-- A reference design for set-top boxes using a Celeron processor with a legacy 815 core logic chipset.
-- A new Extensible Firmware Interface for processors, called Tiano.
Intel puts desktop, mobile processors on 90nm fast track
By Jack Robertson
ebnonline
08/29/2003 1:00 PM EST
URL: http://www.ebnews.com/showArticle?articleID=14200046
Intel Corp. is sampling the first processors manufactured on its 90nm technology process -- the Prescott for desktop PCs and the Dothan, an improved version of the Pentium M chip for laptops.
Intel said it is on schedule for "revenue shipments" of both processors next quarter. The Santa Clara, Calif., company defines initial revenue shipments as limited-quantity orders placed by OEMs that use the processors to build enough computers to prepare for a formal market launch.
An Intel spokesman said the 90nm process ramp is proceeding as anticipated. Initial production at the company's D1C 300mm-wafer development fab in Hillsboro, Ore., is already yielding validated processors, he said. Fab 11X in Albuquerque, N.M., is primed to come on line in the fourth quarter as planned, while a plant in Leixlip, Ireland, is expected to start 90nm processing in the first half of 2004.
G. Dan Hutcheson, president of VLSI Research, San Jose, said Intel "has already proven its 90nm process in its Oregon development fab. Now it's just a matter of making the technology work in other fabs."
Analysts said PC makers will wait to provide systems based on the Prescott and Dothan until Intel has ratcheted 90nm production to a level where it can meet volume consumer demand.
"Intel has set itself a very aggressive launch of two new processors using 90nm process technology," said Kevin Krewell, a processor analyst with In-Stat/MDR, San Jose. "They need to build enough units to supply customers ramping up their PC production, and that could take a bit of time."
Still, the company is confident enough of the debut of its 90nm process that special briefings on the technology are scheduled at the Intel Developer Forum (IDF) Sept. 16-18 in San Jose.
At IDF, Intel will also turn the spotlight on the Prescott and Dothan as it tries to drown out archrival Advanced Micro Devices Inc., which is slated to introduce its Athlon 64 desktop processor the following week.
Prescott is a significant upgrade over current Pentium 4 microprocessors, doubling the on-die L2 cache to 1Mbyte with an expected 3.4GHz frequency. Intel has said Prescott will have improved pre-fetch branch prediction for improved performance. The long pipeline of the Pentium 4 architecture can be slowed down when a wrong pre-fetch prediction must be purged and refilled with data.
Prescott will also be the first Intel processor to feature new internal security support, code-named La Grande. The company will provide at IDF the first details of how the new security technique works.
The processor includes 13 Prescott New Instructions (PNI), ranging from storing and loading integer data to new floating-point operations.
Intel is expected to follow its usual pattern of fielding several versions of Prescott to serve various market segments. Next year, the company will add both higher- and lower-frequency chips to the line. Versions of the chips running at 2.8 and 3 GHz are planned as lower-priced entries for the mainstream market. Shane Rau, a processor analyst with IDC in Mountain View, Calif., said the lower-speed Prescotts will be "a natural fallout" from bin testing.
Dothan for laptops will have 2Mbytes of on-die L2 cache, twice that of the current Pentium M. The microprocessor will also have new power management features to extend battery time, Intel said.
The first Intel Xeon server processor to be made in a 90nm process, Nocona, is slated to debut late this year and will also be featured at IDF. The first Xeon MP series made with 90nm processing, Potomac, is scheduled to be introduced in the second half of 2004.
The company will also use IDF to unveil a processor chipset for laptops. Al- though Intel declined to identify it by name, some sources believe the chipset is called Alviso and is designed for the Pentium M line.
Other developments that Intel is expected to discuss at the forum include:
-- A next-generation Xscale processor, code-named Bulverde.
-- New chipsets supporting DDR2 SDRAM in server applications, including the Lindenhurst for one- and two-way Xeon processors and the Twin Castle, which supports the Xeon MP.
-- A new desktop PC form-factor specification.
-- A reference design for set-top boxes using a Celeron processor with a legacy 815 core logic chipset.
-- A new Extensible Firmware Interface for processors, called Tiano.
Ex-AMD exec portrayed as greedy in trial
DISCRIMINATION SUIT OPENS
By Therese Poletti
Mercury News - Posted on Thu, Aug. 28, 2003
Attorneys for Advanced Micro Devices portrayed Walid Maghribi -- a former top executive who is suing the company for discrimination -- as a greedy and power-hungry individual who left AMD because he was not going to make tens of millions of dollars.
``This case is not about discrimination,'' said Lynne Hermle, an attorney for Orrick, Herrington & Sutcliffe, in her opening argument in U.S. District Court in San Jose on Wednesday. ``It's about an executive who quit when he could not make tens of millions of dollars from a deal.''
Maghribi filed his lawsuit against AMD in April 2002, contending that he was forced to resign from the Sunnyvale chip maker because of intolerable working conditions. Maghribi, a Lebanese-American who was formerly president of AMD's memory-chip business, alleges that several weeks after the Sept. 11 terrorist attacks, AMD's top two executives learned that he was a Muslim of Arab heritage and his career went into a tailspin shortly after.
But Hermle characterized Maghribi as focused on his own desires instead of AMD's corporate goals and said that he resigned from the company once he realized that his personal goals would not be achieved in a memory-chip joint venture he was brokering between AMD and Fujitsu.
``He was making millions annually; he wanted to be the CEO of a public company,'' Hermle said. ``His objective was not the objective of his employer, AMD.''
Maghribi's attorney, Allen Ruby of San Jose, laid out for the jury a timeline of the events that led up to Maghribi's resignation from AMD in December 2001, after 15 years with the company. Maghribi contends that Chairman Jerry Sanders and President Hector Ruiz changed their requirements for the AMD-Fujitsu joint venture after learning of his Arab heritage, making it impossible for him to continue at the company.
``Wally Maghribi had a wonderful career at AMD for 15 years,'' Ruby said in his opening. ``On Oct. 25, the head of the company asked him about his religious faith and background. He informed him he believed in Allah. . . . Forty-two days later, Mr. Maghribi's career at AMD was over.''
Maghribi contended that Sanders wanted AMD to get control of all the sales for the joint venture, something Fujitsu would never agree to.
Bertrand Cambou, who replaced Maghribi as the head of AMD's memory business, was called as a witness by Maghribi's attorney. Cambou, who is now the head of the joint venture by AMD and Fujitsu, said Fujitsu would never give up its sales in Japan, because it would be seen as a ``loss of face.''
In a taped deposition played in court Wednesday, Ruiz responded to Maghribi's allegation that he had told Maghribi and his wife a demeaning joke about Arabs at a company function in October 2001. Ruiz characterized the exchange as an anecdote he related to Maghribi's wife about his being mistaken for an Arab during an encounter at a gas station.
--------------------------------------------------------------------------------
Contact Therese Poletti at tpoletti@mercurynews.com or at (415) 477-2510.
"All rumors that we have heard have pointed to less than 80,000 K8 CPUs total. (Q104 is looking up to over 300K."
Looks like the Opteron launch was a bust.
This data confirms the fact that Opterons are piling up unsold on the pricewatch sites.
-SZ
"The first Athlon64 chips are set to hit the street in September, but won't start coming out in massive volume until the debut of 90-nanometer manufacturing, Marty Seyer, vice president and general manager of AMD's Microprocessor Business Unit, said in an interview earlier this month. "
Incredible.
That will make them 3 years late - instead of 2 years late.
I can see the press release now. "Marty Seyer, VP of Baloney at AMD, says the newly revised schedule for the Athlon64 will align it with the "expected" delivery date for Windows 64 from Microsoft".
Can anybody say "Hektor is a hero" ?
"If AMD is shipping in high volume by the end of the year, Athlon64 will be about 2 years later than the very earliest dates. That's not bad when compared to Itanium (which was a more complex project) that is about 4 to 6 years behind the very first plans"
Not relevant, Dan3.
First, Opteron competes with Xeons - right? And Intel has been shipping Xeons since 1997 - 6 years now.
Second, the Athlon64 will be 3 years late when it finally ships in any volume at the end of next year - since AMD promised it for 2001.
That makes the simple to design-simple-to-make-smaller-than-Intel's chips Athlon64 just as late as the start-from-scratch Itanium was late.
I know you are in denial over this - just as you are in denial that AMD is such a financial and manufacturing basket case.
-SZ
"The number of people needing more than single digits of 64-bit applications will be miniscule."
Yes.
And that will parallell the actual number of people buying the 64 bit CPUs from AMD - miniscule.
The piled-up volume of unsold Opterons appearing on pricewatch gives testimony to this dilemma.
-SZ
"people who buy processors in a box are sophisticated enough to know what they want -- the DIY crowd. "
And the DIY crowd want to pay $690 per 244 Opteron?
"While the Opteron doesn't give a definite lead over the Xeon 3.06 GHz/1MB in multi-processing or the Pentium 4 3.2 in single-processor operation, it sure is close on both fronts. "
Whatever happened to AMD's promise (Jerry, Hektor, Dirk) that the Opteron would be the fastest CPU on the planet at launch?
A sample of Intel's 90 nM capability and future product plans.
Intel tricks out 90-nm TCP/IP processor
By by Ron Wilson, EETimes
08/26/2003 8:53 AM EST
URL: http://www.siliconstrategies.com/article/showArticle.jhtml?articleId=13900085
PALO ALTO, California -- Intel Corp. has detailed an experimental 90-nanometer custom processor that handles TCP/IP ingress processing at measured data rates up to 9.64 Gbits/second.
Presented last Tuesday (August 19, 2003) at the Hot Chips Symposium, the 460,000-transistor chip, which combines a fairly conventional processing data path with the unconventional use of content-addressable memories (CAMs) and control flow, illustrates the potential for fully programmable protocol processing in the 90-nm era but shows the lengths to which designers must go to seize the technology's benefits.
The processor, internally known as TIPP, is part of a much larger effort to study TCP/IP offload processing using software-programmable means. Intel's Microprocessor Research Labs in Hillsboro, Ore., is directing the effort. The program is not limited to specialized processors but is also looking at the problems of protecting operating system and application processing from TCP/IP processing on conventional processors.
Approaching 10-Gbit/s wire speed required a purpose-built TCP/IP processor. The Intel device is a tiny, 2.2 x 3.5-mm die fabricated in one of its more advanced 90-nm logic processes. The processor is intended to handle ingress processing, eventually as one core in a two- to three-times larger system-on-chip (SoC) that will handle the full TCP/IP termination problem.
Intel researcher Jianping Xu outlined the key issues in 10-Gbit/s wire-speed processing. The fundamental problem is time: To keep up with 10-Gbit Ethernet, the processor must cope with a packet every 67 nanoseconds. That puts the task beyond the reach of general-purpose computing. But Xu argued that the ability of a software-based solution to respond to changes in protocols made a processor-based approach worth trying.
Speed aside, TCP processing is known to be messy, requiring demultiplexing of protocol information, packet filtering and protocol processing. Not the least of the problems is that TCP packets are not guaranteed to arrive in proper order, so packets must be reordered on the fly.
The Intel team devised a specialized architecture comprising a pipelined execution unit with tightly coupled scratch pad RAM, two CAMs, input and output buffers, and a dedicated block for transmission control. A local-instruction ROM drives the processor, using a wide, 112-bit control word that is apparently stored fully decoded in the ROM. The architecture handles packet reordering by using the CAMs to access packets by sequence number, rather than by sorting.
In a separate paper delivered by Greg Regnier and others from Intel at the Hot Interconnects conference-also held at Stanford last week-researchers described the use of dual-processor Xeon systems in TCP/IP processing and the offloading of protocol processing. The paper compared the efficiency of a symmetric-multiprocessing approach, in which the Linux TCP/IP stack was handled as just another kernel task, with an asymmetric approach in which one Xeon processor was dedicated to TCP/IP and the other served as a pure Linux application processor. The comparison showed an advantage to offloading protocol processing, even onto conventional hardware.
But the implementation of the TIPP processor, rather than its architecture, was particularly instructive. Intel used many of the advanced design techniques it has discussed over the past year or two. And many of those techniques are aimed directly at known problems with 90-nm processes: power and leakage current.
Power consumption is a serious issue with the chip. Operating at the top of its curve, at 1.72 volts, the device must dissipate about 6.4 watts from a die of less than 8 mm2. Looking into the details of power dissipation makes the scale of the problem more apparent: at the voltage center, around 1.2 V, parts of the execution unit and instruction ROM must dissipate more than 250 W/cm2.
A substantial component of power consumption in 90-nm processes is leakage current. Leakage also makes some structures more challenging to design. In its campaign against the problem, the Intel team employed a dual-threshold-voltage process and adaptive body bias-a technique the company has discussed previously. The part also uses a novel register file design that can tolerate relatively high leakage levels.
Much of the design was synthesized, according to Xu, but the key high-speed paths-the execution core and instruction ROM-were custom designs. The approach also makes use of semidynamic flip-flops. Less-critical parts of the design run at lower frequencies.
The ability to use synthesis for much of the chip, focus both logic and circuit design resources on a few critical blocks and still achieve high throughput may be a key new skill for SoC design teams. But it remains to be seen how well the design style can be supported when the tools come from commercial EDA vendors.
"However, when Windows 64 comes out, they can buy the new OS, install it, and get a performance boost without having to shell out more money for new hardware. "
Why don't you show me some evidence that existing 32 bit software will run faster on a 64 bit OS that doesn't yet exist.
Please..because "AMD says so" doesn't qualify as evidence.
-SZ
"They will just have to shell out more money for all new 64 bit software....Of course, we all know that they only have to buy a new operating system, all the old applications will work fine."
And after shelling out for an exorbitant AMD64 machine, then waiting a year or so, and shelling out more for the 64 bit OS, they will have only 32 bit applications to run - and will be in exactly the same position as they are today.
To use that 64 bit OS in any meaningful way will require 64 bit applications, and they will have to buy all new 64 bit applications, assuming they are available, and these will be very expensive because the volume will be low for the ISVs.
Get it yet?
"However, when Windows 64 comes out, they can buy the new OS, install it, and get a performance boost without having to shell out more money for new hardware."
They will just have to shell out more money for all new 64 bit software.
Did you happen to think of that?
"Does anyone wish to comment on the recent sale by Michael Dell of ~ 10 million shares of INTC last week over four days?"
I would like to comment.
Michael Dell sold absolutely ZERO shares of INTC last week.
Where did you ever get the silly notion that Dell even owns any Intel stock, let alone 270 million - dollars or shares - of Intel?
-SZ
What a great system - a 64 bit CPU bundled with a home edition of 32 bit Windows XP.
And just how does this 32 bit Windows XP bring the promise of 64 bit computing to the gaming masses?
• Microsoft® Windows® XP HOME Edition & Microsoft® Works 7.0
http://www.cyberpowersystem.com/highendsystem/Athlon64_sys.htm
"no references to the fact that SGI might bundle SuSe software with their Itanium servers."
Surprise !
Press Release Source: SGI
SGI, SuSE Announce OEM Alliance, Development Agreement
Monday August 25, 9:01 am ET
SGI to Deploy SuSE Linux on SGI Altix 3000 Servers; SuSE Joins SGI in Scaling Linux to 128 Processors
MOUNTAIN VIEW, Calif. and OAKLAND, Calif., Aug. 25 /PRNewswire-FirstCall/ -- SGI (NYSE: SGI - News) and SuSE Linux today announced plans to extend the Linux® OS to new levels of scalability and performance -- offering, for the first time, a fully supported 64-processor system running a fully supported, enterprise-grade Linux operating system.
ADVERTISEMENT
Expected to be available in October, SGI will bundle SuSE Linux Enterprise Server 8 on SGI® Altix(TM) 3000 servers and superclusters. SuSE Linux will support the unique 64-processor single-node scaling capabilities of the SGI Altix 3000 platform. SuSE will also provide third-level customer support to SGI customers in a seamless support agreement with SGI.
The companies also announced that SGI and SuSE will work on co-development efforts, including SGI's program to scale Linux to 128 processors, which the companies anticipate will be available next year.
While other hardware vendors and Linux distributions support only small systems and partitions running Linux, SGI and SuSE can deliver large-scale systems -- with nodes from 4 to 64 and soon to 128 processors -- capable of addressing high-performance computing and technical database requirements. This collaboration opens doors for further expansion of standard Linux in the marketplace.
"Oracle welcomes SGI and SuSE's alliance, because it will enable Oracle customers on that platform to avail themselves of Oracle's support for Linux," said Dave Dargo, vice president of Platform Partners, Linux and Performance, Oracle Corp. "SGI servers, coupled with solutions based on Oracle Database and SuSE Linux Enterprise Server, will give customers increased choice for their enterprise Linux applications."
Not only will the companies continue to expand the capabilities and reach of SuSE Linux, but the new offering also broadens the market opportunities for both SGI and SuSE. Together the companies will develop and deliver an offering for customers whose technical or scientific applications interact with additional large-scale databases and enterprise infrastructure software. Ideal for a broad range of markets -- from bio-informatics to automotive design and production -- the combined offering provides industry-leading computing power, bandwidth and database performance.
"SuSE Linux Enterprise Server 8, certified for the industry's key hardware and software, will enable Altix customers to leverage the inherent security and stability of Linux," said Uwe Heine, chief alliance officer, SuSE Linux. "Together, SuSE and SGI can deliver high-performance Linux systems capable of handling the unique system load demands of scalable integer processing, throughput and memory."
The SGI Altix 3000 family of servers and superclusters combines SGI® supercomputing architecture with Intel® Itanium® 2 processors and the Linux operating system. Since its introduction, the SGI® Altix(TM) supercluster has been recognized as the first Linux cluster that scales up to 64 processors within each node and the first cluster to allow global shared-memory access across nodes. Inspired by the success of the SGI Altix family and the powerful combination of standard Linux and Intel® processors, more than 60 high-performance manufacturing, science, energy and environmental applications have been ported by their commercial developers to the 64-bit Linux environment, more than two-thirds of which have been certified and optimized for the platform.
"SuSE's support of advanced Linux features, such as scalability, the XFS filesystem, security and native 64-bit environments provide SGI and SuSE with a wealth of opportunities for collaboration," said Dave Parry, senior vice president and general manager, Server and Platform Group, SGI. "As a highly respected and experienced provider of open-source solutions, SuSE shares SGI's vision of a broadly scalable and robust Linux environment for technical and scientific computing."
On August 5, SGI announced plans to extend the Altix platform to encompass a record 128 processors within a single instance of Linux. See the press release titled "SGI Leads Effort to Scale Linux to 128 Processors in a Single-System Configuration" on www.sgi.com/newsroom.
Availability
SuSE Linux Enterprise Server 8 is expected to be available on Altix systems from SGI and SuSE in October 2003. SGI Altix 3000 servers and superclusters supporting 64-processor nodes, in configurations up to 128 Intel Itanium 2 processors, are available today from SGI. For customers demanding even larger Altix superclusters, SGI expects to support configurations of 256 processors in September and 512 processors in October 2003. Additional Altix system technical and availability information is posted on www.sgi.com/servers/altix.
This news release contains forward-looking statements regarding SGI technologies and third-party technologies that are subject to risks and uncertainties. These risks and uncertainties could cause actual results to differ materially from those described in such statements. The reader is cautioned not to rely unduly on these forward-looking statements, which are not a guarantee of future or current performance. Such risks and uncertainties include long-term program commitments, the performance of third parties, the sustained performance of current and future products, financing risks, the ability to integrate and support a complex technology solution involving multiple providers and users, and other risks detailed from time to time in the company's most recent SEC reports, including its reports on Form 10-K and Form 10-Q.
About SuSE
SuSE Linux is one of the world's leading providers of products and services based on the Linux operating system. In addition to the operating system and application software for private customers, SuSE offers software solutions and complete systems for the use of Linux in the enterprise. SuSE supports enterprise customers with a comprehensive range of qualified consulting, training and support services. Hosting the world's largest development team for Open Source solutions, SuSE Linux shares its unique project and support information online in the largest existing Linux knowledge database.
About SGI
SGI, also known as Silicon Graphics, Inc., is the world's leader in high-performance computing, visualization and storage. SGI's vision is to provide technology that enables the most significant scientific and creative breakthroughs of the 21st century. Whether it's sharing images to aid in brain surgery, finding oil more efficiently, studying global climate or enabling the transition from analog to digital broadcasting, SGI is dedicated to addressing the next class of challenges for scientific, engineering and creative users. SGI was named on FORTUNE magazine's 2003 list of "Top 100 Companies to Work For." With offices worldwide, the company is headquartered in Mountain View, Calif., and can be found on the Web at www.sgi.com.
NOTE: Silicon Graphics, SGI, XFS and the SGI logo are registered trademarks and Altix is a trademark of Silicon Graphics, Inc., in the United States and/or other countries worldwide. SuSE is a registered trademark of SuSE Linux AG. Linux is a registered trademark of Linus Torvalds in several countries. Intel and Itanium are registered trademarks of Intel Corporation. Oracle is a registered trademark of Oracle Corporation and/or its affiliates. All other trademarks mentioned herein are the property of their respective owners.
( Logo: http://www.newscom.com/cgi-bin/prnh/20010510/SFTH025LOGO )
CONTACT: Ginny Babbitt, +1-650-933-4519, or ginnyb@sgi.com, or SGI PR Hotline, +1-650-933-7777, or SGI PR Fax, +1-650-932-0737, all of SGI; or Joseph Eckert of SuSE, +1-203-270-3711, or eckert@SuSE.com.
--------------------------------------------------------------------------------
Source: SGI
more good itanium news.
SuSE to bundle SGI servers
By Lisa M. Bowman
Staff Writer, CNET News.com
August 24, 2003, 9:00 PM PT
http://news.com.com/2100-1010-5067135.html
In an effort to broaden Linux's appeal, SuSE Linux will bundle its new SuSE Linux Enterprise Server 8 with SGI's Altix 3000 servers and superclusters.
The company plans to start shipping the bundled product in October, it said. The move expands SuSE's 64-bit products and paves the way for scaling Linux to 128 processors, the latest effort to turbo-charge the increasingly popular open-source operating system. SuSE and SGI said they would work together to develop a 128-processer version of Linux.
"SuSE shares SGI's vision of a broadly scalable and robust Linux environment for technical and scientific computing," Dave Parry, general manger of SGI's server and platform group, said.
Mountain View, Calif.-based SGI makes specialized computers for complicated graphics tasks, such as the creation of special effects for movies or the visualization of airflow in car design. Most of its systems use Irix, the company's version of Unix, in combination with MIPS processors. But earlier this year, SGI launched the Altix 3000, which is based on Linux and Intel Itanium processors.
A few weeks ago, SGI said it would ship a 128-bit Altix machine in the spring.
Both companies are aiming to expand their reach through the agreement. SGI gains a standardized version of Linux, which it hopes will attract more applications to its machines.
In a statement announcing the partnership, Oracle said it welcomed the deal because customers running its software on the SGI servers using SuSE's Linux would have access to Oracle's support for the operating system.
SuSE, for its part, gets another high-profile endorsement of its version of Linux and a commitment to further soup up the operating system. In June, Hewlett-Packard expanded its relationship with SuSE, announcing that customers could get SuSE Linux Enterprise Server 8 bundled with its ProLiant line machines, which mainly uses Intel's Xeon chips, or with its servers based on Itanium chips.
"do you have any information on how amd is doing with their transition to .09? I am curious. The more transistors the more leakage.
Only what's been reported in the press (their present schedule calls for them to start shipping late Q1 or early Q2). "
Excellent job at wordsmithing, Dan.
You have studied at the feet of the master, Jerry Sanders, and have learned well.
Yes, you state AMD's "present schedule" for 90nM.
Ah....wordsmithing at its finest.
You cleverly left off the fact that the "current schedule" represents a 6 to 9 month slip from AMD's "prior schedule".
But then again, AMD has slipped every single product and process schedule for almost 2 1/2 years now.
SuSE, AMD's 64 bit Linux partner, is now going to sell 64 bit servers along with their 64 bit Linux software.
Intel Itanium servers, that is.
Sort of makes you wonder why AMD's Opteron/Linux partner wants to sell Intel Itanium servers - no?
Maybe all is not well in Opteron-land.
SuSE to bundle SGI servers
By Lisa M. Bowman
Staff Writer, CNET News.com
August 24, 2003, 9:00 PM PT
http://news.com.com/2100-1010-5067135.html
In an effort to broaden Linux's appeal, SuSE Linux will bundle its new SuSE Linux Enterprise Server 8 with SGI's Altix 3000 servers and superclusters.
The company plans to start shipping the bundled product in October, it said. The move expands SuSE's 64-bit products and paves the way for scaling Linux to 128 processors, the latest effort to turbo-charge the increasingly popular open-source operating system. SuSE and SGI said they would work together to develop a 128-processer version of Linux.
"SuSE shares SGI's vision of a broadly scalable and robust Linux environment for technical and scientific computing," Dave Parry, general manger of SGI's server and platform group, said.
Mountain View, Calif.-based SGI makes specialized computers for complicated graphics tasks, such as the creation of special effects for movies or the visualization of airflow in car design. Most of its systems use Irix, the company's version of Unix, in combination with MIPS processors. But earlier this year, SGI launched the Altix 3000, which is based on Linux and Intel Itanium processors.
A few weeks ago, SGI said it would ship a 128-bit Altix machine in the spring.
Both companies are aiming to expand their reach through the agreement. SGI gains a standardized version of Linux, which it hopes will attract more applications to its machines.
In a statement announcing the partnership, Oracle said it welcomed the deal because customers running its software on the SGI servers using SuSE's Linux would have access to Oracle's support for the operating system.
SuSE, for its part, gets another high-profile endorsement of its version of Linux and a commitment to further soup up the operating system. In June, Hewlett-Packard expanded its relationship with SuSE, announcing that customers could get SuSE Linux Enterprise Server 8 bundled with its ProLiant line machines, which mainly uses Intel's Xeon chips, or with its servers based on Itanium chips.
Why are you listing AMD's and Intel's Book value?
I asked you for AMD's DEBT per share.
That's D E B T - the money AMD owes to its lenders.
Ask your self why AMD has more debt than they have cash.
What would happen to you personally if you had more debt than cash and no income stream to pay off that debt ?
By the way - we all know the trick AMD is using to lower their losses per share - by giving out free shares to Rivet anbd Siegle and Dicky Doran to dilute the shares be xpanding the number of outstanding shares.
If AMD ever makes a profit again, those extra shares will reduce AMD's earnings per share - the flip side of what they are doing now.
"how can AMD legally grant options to their senior staff at a penny per share? Where do they account for these compensations in their balance sheet?"
Cut the crap and don't ask these questions or you will be banned.
Only Intel's accounting procedures for stock options can be criticized on this board.
Whatever AMD dirctors do to steal from the corporate coffers and give unto themselves is perfectly alright and can nevet be called in to question.
How dare you.
"A new fab today costs over 2 billion dollars. These are fixed costs that are the same whether you're AMD or Intel "
This is not correct.
If Intel wants to build a $2 Billion fab, they take the money out of their bank account and spend $2 Billion.
If AMD wants to build a $2 Billion fab, they sell paper bonds, they negotiate loans, etc. and build that fab with other people's money.
The cost to AMD is ZERO.
"The number Intel produced is the number Intel thinks can be absorbed by the market in this quarter at the PREMIUM prices they want to charge. That is....to extract as much money out of the notebook segment of the market as possible"
That is exactly correct.
Intel is in business to make money - isn't it?
Isn't a measure of a company's success the amount of money (profits) it makes?
By the way - why is AMD in business?
Does AMD optimize the number of devices they make to maximize their losses - to sell as many devices at the biggest loss they (AMD) can absorb ?
"Actually, compared to its size, Intel doesn't have much more cash than AMD. Cash per share at Intel is way, way less than AMD."
Why don't you post AMD's DEBT per share and Intel's DEBT per share - would you please?
Those data points ahould be a real eye opener for all AMD investors.
"I may be mistaken, but I don't think directors can get any $0.01 options."
You really love AMD, right Petz?
Bobby Rivet gets 10,000 shares PER YEAR at 1 penny per share.
Billy Siegle gets 6000 shares PER YEAR at 1 penny per share.
Little Jimmy Doran gets 12,500 shares PER YEAR at 1 penny per share.
2002-07-25 RIVET, ROBERT J.
Chief Financial Officer 10,000 Option Exercise at $0.01 per share.
(Cost of $100)
2003-01-20 SIEGLE, WILLIAM T.
Senior Vice President 3,000 Option Exercise at $0.01 per share.
(Cost of $30)
2003-02-11 DORAN, JAMES E.
Vice President 12,500 Option Exercise at $0.01 per share.
(Cost of $125)
2003-07-21 SIEGLE, WILLIAM T.
Senior Vice President 3,000 Option Exercise at $0.01 per share.
(Cost of $30)
2003-07-25 RIVET, ROBERT J.
Chief Financial Officer 10,000 Option Exercise at $0.01 per share.
(Cost of $100)
"AMD recently opened up a significant percentage of Dresden's clean room space by moving development to IBM's facility."
No they didn't.
You just made that canard up out if thin air.
AMD moved engineers and resources from their California development site to IBM's site.
http://news.com.com/2008-1082-995885.html
"Hector also recently talked about how AMD engineers and developers are moving to upstate New York to work at IBM on developing future microprocessor technology? How much of the engineering staff is up at IBM right now?
Our Semiconductor Development Center in Sunnyvale (Calif.) used to house both memory that is flash, and logic technology development. SDC is transitioning this quarter to be devoted to flash. Meanwhile, our logic technology development is going to be done jointly with IBM, and we will have a number of AMD engineers working in East Fishkill (N.Y.). Ultimately, that technology will be deployed in our factories. "
"AMD in Poland !"
Gee that's swell.
AMD loses 2 - 3% overall worldwide market share, dumps another $145 million down the drain in their 8'th consecutive quarterly loss, but they do well in Poland.
I'll let you finish off this joke.
"AMD runs their fabs at low yields - about 30% or 35%
No, they don´t. "
Yes they do.
Do the math.
And don't assume a yield to prove a yield.
The 30 - 35% yield falls out every time.