System builders ready for high-performing, energy-efficient single socket AMD processors—
SUNNYVALE, Calif. -AMD (NYSE: AMD) today announced significant support for the Quad-Core AMD Opteron™ 1300 Series processor from channel partners focused on enterprise and small and medium businesses as well as high performance computing (HPC) markets. The AMD Opteron 1300 Series processor provides the latest in enterprise-class performance and numerous energy efficient features for single processor servers and workstations, a key market focus for many leading system builders.
“Our technology partners depend on AMD to offer solutions that enable them to grow their business and provide value to their customers,” said Gary Bixler, director, North America Channel Marketing, AMD. “The AMD Opteron 1300 Series processor delivers quad-core computing in a stable socket and with a consistent power and thermal envelope that can help both reduce time-to-market for our channel partners and allow them to offer their customers a simple and efficient upgrade path for increased performance.”
Commercial channel customers will have the opportunity to engage with AMD authorized distributors to choose from a variety of motherboards with AMD Opteron 1300 Series processor support.
AMD also offers a robust Validated Server Program for commercial channel partners wishing to leverage barebones server platforms. These tested building blocks allow system builders to focus on the unique configurations their customers request and offer quick integration of new technology into their business. Learn more about AMD’s Validated Server Program at www.amd.com/vsp.
Advanced Clustering
"Advanced Clustering is proud to add the new Quad-Core AMD Opteron 1300 Series processors to our power efficient Pinnacle Server family. We believe that it will offer a great price/performance value for those wanting a Beowulf style cluster on a budget" - Shelly Kelley, Vice President of Sales and Marketing, Advanced Clustering
ASA
“This new offering by AMD deepens our support for offering Quad-Core AMD Opteron processor-based servers and workstations. We are particularly excited to offer these high-performing, energy-efficient processors to customers who rely on this level of hardware to support critical applications.” - Arvind Bhargava, CEO, ASA
Colfax
“The new Quad-Core AMD Opteron 1300 Series processors create a great market opportunity for Colfax. This helps Colfax be well-positioned to provide entry level workstation and server offerings for businesses looking for smart solutions. We plan to offer systems with the new AMD Opteron processor.” - Gautam Shah, President, Colfax International
Microway
“The new Quad-Core AMD Opteron 1300 Series processors will enable our customers to achieve higher levels of performance and energy efficiency. Our WhisperStations-PRO workstation customers will enjoy the performance improvements of AMD’s native quad-core architecture" - Ann Fried, CEO, Microway
PSSC Labs
“PSSC Labs is pleased to introduce Quad-Core AMD Opteron 1300 Series processors into our PowerServe Centro A1400 and PowerStation Centro A1400 models. PSSC Labs continues to be impressed with AMD’s commitment to bringing high-performing, energy-efficient computing processors to market.” - Alex Lesser, vice president, PSSC Labs
Super Micro Computer, Inc.
“Super Micro is a leader in application-optimized, high performance server and workstation solutions and we have announced full support for the Quad-Core AMD Opteron 1300 Series processors on our latest single-processor A+ Servers, including the 1011M-UR and 1011S-MR2, currently on display at Computex Taipei 2008 in booths M619-726, 4th Floor" Alex Hsu, chief sales and marketing officer, Super Micro Computer, Inc.
Verari
“Verari plans to integrate the Quad-Core AMD Opteron 1300 Series processor into our rackmount server offerings. We are enthusiastic about this opportunity to continue to provide industry leading energy-efficiency and performance for our customers.” - Ed Holden, server product manager, Verari Systems
ZT Systems
“ZT Systems is proud to be selected by AMD to display new servers featuring Quad-Core AMD Opteron 1300 series processors at Computex. The ZT Systems 2102Ra Storage Server enables energy efficient performance and low cost of ownership for enterprise customers, and is just one example of ZT’s innovative AMD-powered server solutions for the datacenter.” - Bob Anderson, executive vice president, Business Development and Strategy, ZT Systems
Saturday, July 26, 2008
AMD releases quad-core Opterons for two-, four- and eight-socket systems
On Monday, June 9, Sunnyvale, Calif.-based Advanced Micro Devices Inc. (AMD) introduced four new quad-core AMD Opteron SE processors for two-, four- and eight-socket commodity x86 systems. The new processors are comparable to larger, more costly symmetric multiprocessing (SMP) hardware.
"These Opteron offerings are for users who are most interested in maximum performance and are not concerned as much about wattage," said Steve Demski, AMD's Opteron product manager, "people who may have used high-end Unix systems and mainframes and want that kind of performance at a lower price point."
AMD also released quad-core processors for single-socket systems last week for applications that require a lot of data transfer and communication between processors.
In comparison, these quad-core processors designed for multiple-socket systems are best for applications that require a lot of memory or that perform more computing functions, rather than communications, according to Peter Ungaro, the CEO and president of Seattle-based Cray Inc., which uses various levels of AMD Opteron processors in its systems.
To be specific, workloads typically benefiting from four- and eight-socket server processing (the AMD 8000 series) include large databases, business processing systems (enterprise resource planning, customer relationship management and supply chain management) and business intelligence, as well as other IT infrastructure applications including heavy-duty Web serving and messaging, according to AMD.
AMD Opteron 2000 Series processors for two-socket systems are used for Web, file and print servers, Java application servers, and some email servers. Clustered two-socket systems are also frequently used for applications in the life sciences and oil and gas industries, and in other engineering applications.
Quad-core Opterons for one-socket systems (i.e., the AMD 1000 series) are typically used for Web serving, file and print needs, and other less CPU-intensive applications. Cray's supercomputing systems are unique in their use of single sockets, an AMD spokesperson said.
The new quad-core AMD Opteron SE processors use registered memory and can support up to eight memory sticks per CPU. The chips are available from global OEMs and vendors, including Hewlett-Packard Co., Sun Microsystems Inc., Dell Inc. and IBM.
The new quad-core AMD Opteron processor models and prices are as follows: 2358 SE (with 2.4 GHz) is $873; 2360 SE (with 2.5 GHz) is $1,165; 8358 SE (with 2.4 GHz) is $1, 865, and 8360 SE (with 2.5 GHz) $2,149. The processors are widely available.
In addition, the new processors have set some benchmarks. They achieved the highest SAP-SD two-tier score in eight-socket x86 servers, and quad-core AMD Opteron SE processors also achieved the highest SPECfp_rate2006 scores in both two- and four-socket x86 servers among comparable x86 processors.
"These Opteron offerings are for users who are most interested in maximum performance and are not concerned as much about wattage," said Steve Demski, AMD's Opteron product manager, "people who may have used high-end Unix systems and mainframes and want that kind of performance at a lower price point."
AMD also released quad-core processors for single-socket systems last week for applications that require a lot of data transfer and communication between processors.
In comparison, these quad-core processors designed for multiple-socket systems are best for applications that require a lot of memory or that perform more computing functions, rather than communications, according to Peter Ungaro, the CEO and president of Seattle-based Cray Inc., which uses various levels of AMD Opteron processors in its systems.
To be specific, workloads typically benefiting from four- and eight-socket server processing (the AMD 8000 series) include large databases, business processing systems (enterprise resource planning, customer relationship management and supply chain management) and business intelligence, as well as other IT infrastructure applications including heavy-duty Web serving and messaging, according to AMD.
AMD Opteron 2000 Series processors for two-socket systems are used for Web, file and print servers, Java application servers, and some email servers. Clustered two-socket systems are also frequently used for applications in the life sciences and oil and gas industries, and in other engineering applications.
Quad-core Opterons for one-socket systems (i.e., the AMD 1000 series) are typically used for Web serving, file and print needs, and other less CPU-intensive applications. Cray's supercomputing systems are unique in their use of single sockets, an AMD spokesperson said.
The new quad-core AMD Opteron SE processors use registered memory and can support up to eight memory sticks per CPU. The chips are available from global OEMs and vendors, including Hewlett-Packard Co., Sun Microsystems Inc., Dell Inc. and IBM.
The new quad-core AMD Opteron processor models and prices are as follows: 2358 SE (with 2.4 GHz) is $873; 2360 SE (with 2.5 GHz) is $1,165; 8358 SE (with 2.4 GHz) is $1, 865, and 8360 SE (with 2.5 GHz) $2,149. The processors are widely available.
In addition, the new processors have set some benchmarks. They achieved the highest SAP-SD two-tier score in eight-socket x86 servers, and quad-core AMD Opteron SE processors also achieved the highest SPECfp_rate2006 scores in both two- and four-socket x86 servers among comparable x86 processors.
AMD’s Additions to its Opteron Portfolio
Advanced Micro Devices (AMD) has completed its quad-core Opteron portfolio, as it has started shipping two new processors for the 2300 series as well as two more for the 8300 series. The new chips are designed for high-end four-socket / eight-socket servers.
The newcomers are called Opteron 2358 SE (2.4GHz), Opteron 2360 SE (2.5GHz), Opteron 8358 SE (2.4GHz) and 8360 SE (2.5GHz); these are part of the company’s last set of 65-nanometer server models, as it plans to switch to a 45-nanometer assembly process later on this year. The upcoming chips may bring a speed increase of up to twenty percent.
In the second week of May, AMD launched five quad-core Opteron processors (8347 HE, 8346 HE, 2347 HE, 2346 HE, and 2344 HE), described by company officials as energy-efficient. They run on 55 Watts of AC power and, when compared to the 75-watt Barcelona chips that were released by AMD last year, the Opteron shows a 30 percent efficiency increase. The chips’ speeds, as well as their prices, range from 1.7GHz to 1.9GHz and from $209 to $873, respectively.
Also in May, the company presented its two-year server and workstation product plan. Next year, AMD’s first six-core chip is scheduled for release and for the first half of 2010, a 12-core processor is being prepared.
The newcomers are called Opteron 2358 SE (2.4GHz), Opteron 2360 SE (2.5GHz), Opteron 8358 SE (2.4GHz) and 8360 SE (2.5GHz); these are part of the company’s last set of 65-nanometer server models, as it plans to switch to a 45-nanometer assembly process later on this year. The upcoming chips may bring a speed increase of up to twenty percent.
In the second week of May, AMD launched five quad-core Opteron processors (8347 HE, 8346 HE, 2347 HE, 2346 HE, and 2344 HE), described by company officials as energy-efficient. They run on 55 Watts of AC power and, when compared to the 75-watt Barcelona chips that were released by AMD last year, the Opteron shows a 30 percent efficiency increase. The chips’ speeds, as well as their prices, range from 1.7GHz to 1.9GHz and from $209 to $873, respectively.
Also in May, the company presented its two-year server and workstation product plan. Next year, AMD’s first six-core chip is scheduled for release and for the first half of 2010, a 12-core processor is being prepared.
IBM's Roadrunner Supercomputer Sets World-Record Speed
Beep beep! Sorry coyote, but you're still not wily enough to touch Roadrunner, IBM's supercomputer, billed as the fastest in the world operating at 1 petaflop or 1,000 trillion calculations per second.
To be housed at the Los Alamos National Laboratory in New Mexico, IBM built Roadrunner for the Department of Energy's National Nuclear Security Administration to ensure the safety and reliability of the nation's nuclear weapons stockpile. The supercomputer, named after New Mexico's state bird, will also be used for astronomy, energy, human genome science and climate change research.
So, exactly how fast is the supercomputer? IBM said the speed is roughly equivalent to the combined computing power of 100,000 of today's fastest laptop computers—users would need a stack of laptops 1.5 miles high to match Roadrunner's performance. It would also take the entire population of the earth--about 6 billion people--each working a handheld calculator at the rate of 1 second per calculation more than 46 years to do what Roadrunner can do in one day.
IBM said that in the past 10 years, supercomputer power has increased about 1,000 times. Today, just three of Roadrunner's 3,456 tri-blade units have the same power as the 1998 fastest computer. Now, a complex physics calculation that will take Roadrunner one week to complete would have taken the 1998 machine 20 years to finish.
IBM also said Roadrunner, the world's first hybrid supercomputer, uses a first-of-a-kind design, the Cell Broadband Engine. Originally designed for video game platforms such as the Sony Playstation 3, the engine will work in conjunction with AMD's x86 processors. Other companies that contributed components and technology include Emcore, Flextronics, Mellanox and Voltaire.
In total, the computer connects 6,948 dual-core AMD Opteron chips on IBM Model LS21 blade servers, in addition to 12,960 Cell engines on IBM Model QS22 blade servers.
Standard processing--such as file system I/O--is handled by the Opteron processors. Mathematical and CPU-intensive elements are directed to the cell processors.
The system has 80 terabytes of memory and is housed in 288 refrigerator-size IBM BladeCenter racks taking up 6,000 square feet. Roadrunner's 10,000 connections--both Infiniband and Gigabit Ethernet--require 57 miles of fiber-optic cable and weigh a whopping 500,000 pounds.
The supercomputer's custom configuration uses two IBM QS22 blade servers and one IBM LS21 blade server that are combined into a specialized "tri-blade" configuration. Each tri-blade unit can run at 400 billion operations per second (400 Gigaflops). In total, Roadrunner has 3,456 tri-blades.
Roadrunner was built, tested and benchmarked in IBM's Poughkeepsie, N.Y., plant, which is also the home of the ASCI series of supercomputers the company built for the U.S. government in the late 1990s. IBM's site in Rochester, Minn., contributed to the project by constructing the specialized tri-blade servers. In addition, engineers developed the computers' software in IBM's Austin, Texas, and Yorktown Heights, N.Y., research labs. Roadrunner operates on open-source Linux software from Red Hat.
Despite its massive size, IBM calls Roadrunner an "energy miser." The Armonk, N.Y.-based company said that compared to most traditional supercomputer designs, Roadrunner's hybrid format sips power (3.9 megawatts) and delivers efficiency at 376 million calculations per watt.
As for future plans for Roadrunner, IBM is developing new software to make Cell-powered hybrid computing broadly accessible. Roadrunner's massive software effort targets commercial applications for hybrid supercomputing. Along with corporate and academic partners, IBM is developing an open-source ecosystem that is intended to bring hybrid supercomputing to financial services, energy exploration and medical imaging industries, among others.
Later this summer IBM will load the behemoth supercomputer onto 21 tractor trailer trucks to deliver it to the Los Alamos National Lab in New Mexico.
To be housed at the Los Alamos National Laboratory in New Mexico, IBM built Roadrunner for the Department of Energy's National Nuclear Security Administration to ensure the safety and reliability of the nation's nuclear weapons stockpile. The supercomputer, named after New Mexico's state bird, will also be used for astronomy, energy, human genome science and climate change research.
So, exactly how fast is the supercomputer? IBM said the speed is roughly equivalent to the combined computing power of 100,000 of today's fastest laptop computers—users would need a stack of laptops 1.5 miles high to match Roadrunner's performance. It would also take the entire population of the earth--about 6 billion people--each working a handheld calculator at the rate of 1 second per calculation more than 46 years to do what Roadrunner can do in one day.
IBM said that in the past 10 years, supercomputer power has increased about 1,000 times. Today, just three of Roadrunner's 3,456 tri-blade units have the same power as the 1998 fastest computer. Now, a complex physics calculation that will take Roadrunner one week to complete would have taken the 1998 machine 20 years to finish.
IBM also said Roadrunner, the world's first hybrid supercomputer, uses a first-of-a-kind design, the Cell Broadband Engine. Originally designed for video game platforms such as the Sony Playstation 3, the engine will work in conjunction with AMD's x86 processors. Other companies that contributed components and technology include Emcore, Flextronics, Mellanox and Voltaire.
In total, the computer connects 6,948 dual-core AMD Opteron chips on IBM Model LS21 blade servers, in addition to 12,960 Cell engines on IBM Model QS22 blade servers.
Standard processing--such as file system I/O--is handled by the Opteron processors. Mathematical and CPU-intensive elements are directed to the cell processors.
The system has 80 terabytes of memory and is housed in 288 refrigerator-size IBM BladeCenter racks taking up 6,000 square feet. Roadrunner's 10,000 connections--both Infiniband and Gigabit Ethernet--require 57 miles of fiber-optic cable and weigh a whopping 500,000 pounds.
The supercomputer's custom configuration uses two IBM QS22 blade servers and one IBM LS21 blade server that are combined into a specialized "tri-blade" configuration. Each tri-blade unit can run at 400 billion operations per second (400 Gigaflops). In total, Roadrunner has 3,456 tri-blades.
Roadrunner was built, tested and benchmarked in IBM's Poughkeepsie, N.Y., plant, which is also the home of the ASCI series of supercomputers the company built for the U.S. government in the late 1990s. IBM's site in Rochester, Minn., contributed to the project by constructing the specialized tri-blade servers. In addition, engineers developed the computers' software in IBM's Austin, Texas, and Yorktown Heights, N.Y., research labs. Roadrunner operates on open-source Linux software from Red Hat.
Despite its massive size, IBM calls Roadrunner an "energy miser." The Armonk, N.Y.-based company said that compared to most traditional supercomputer designs, Roadrunner's hybrid format sips power (3.9 megawatts) and delivers efficiency at 376 million calculations per watt.
As for future plans for Roadrunner, IBM is developing new software to make Cell-powered hybrid computing broadly accessible. Roadrunner's massive software effort targets commercial applications for hybrid supercomputing. Along with corporate and academic partners, IBM is developing an open-source ecosystem that is intended to bring hybrid supercomputing to financial services, energy exploration and medical imaging industries, among others.
Later this summer IBM will load the behemoth supercomputer onto 21 tractor trailer trucks to deliver it to the Los Alamos National Lab in New Mexico.
AMD Releases Four New Quad-Core Opterons
Advanced Micro Devices on Monday introduced its fastest quad-core server chips to date, two new 2-socket and two new 8-socket processors dubbed Opteron SE.
Following a highly publicized setback late last year in its initial ramp of quad-core processors caused by a silicon glitch, Sunnyvale, Calif.-based AMD and OEM partners finally hit the market with B3-revision chips in March.
After splashy launch parties around the globe way back in September of last year, the ramp of the product codenamed Barcelona is finally proceeding apace with the introduction of two new 2.4GHz and two new 2.5GHz parts. The 2.4GHz quad-core Opteron 2358 SE is priced at $873 on AMD's Website, while the new top-of-the-line 2-socket released Monday by AMD, the 2.5GHz 2360 SE, has a price tag of $1,165.
Adding clock speed to its 8-socket quad-core portfolio are the new 2.4GHz 8358 SE, priced at $1,865, and the 2.5GHz 8360 SE, which runs a cool $2,149.
"In the 2-socket space, we're talking about the HPC folks, people doing genome-mapping for instance, where they need every last bit of clock speed and performance from each core," said AMD server chip manager Steve Demski last week.
"In the 8-socket space, the demographic is customers doing the really intense number-crunching against those big enterprise databases."
Demski said AMD expects Sun and Hewlett-Packard, who already offer 8-way servers built on previously released quad-core Opterons, to lead the way with new configurations featuring the 8358 SE and 8360 SE devices.
The new devices are benchmarking particularly well in terms of floating point performance, a traditional Opteron strength, Demski said, but the two new 8300 parts have also nailed the highest SAP-SD two-tier score in eight-socket servers and AMD is touting the 8358 SE and 8360 SE for data center scalability.
AMD also released three new Barcelona processors, codenamed Budapest, for single-socket servers earlier in June, bringing the chip maker's range of quad-core Opteron products a lot closer to the comprehensiveness anticipated last year, before the TLB errata was discovered in the original silicon.
The three single-socket chips that make up AMD's new Opteron 1300 series are all quad-cores, comprising the 2.1GHz Opteron 1352 ($209), the 2.2GHz 1354 ($255) and the 2.3GHz 1356 ($377).
Following a highly publicized setback late last year in its initial ramp of quad-core processors caused by a silicon glitch, Sunnyvale, Calif.-based AMD and OEM partners finally hit the market with B3-revision chips in March.
After splashy launch parties around the globe way back in September of last year, the ramp of the product codenamed Barcelona is finally proceeding apace with the introduction of two new 2.4GHz and two new 2.5GHz parts. The 2.4GHz quad-core Opteron 2358 SE is priced at $873 on AMD's Website, while the new top-of-the-line 2-socket released Monday by AMD, the 2.5GHz 2360 SE, has a price tag of $1,165.
Adding clock speed to its 8-socket quad-core portfolio are the new 2.4GHz 8358 SE, priced at $1,865, and the 2.5GHz 8360 SE, which runs a cool $2,149.
"In the 2-socket space, we're talking about the HPC folks, people doing genome-mapping for instance, where they need every last bit of clock speed and performance from each core," said AMD server chip manager Steve Demski last week.
"In the 8-socket space, the demographic is customers doing the really intense number-crunching against those big enterprise databases."
Demski said AMD expects Sun and Hewlett-Packard, who already offer 8-way servers built on previously released quad-core Opterons, to lead the way with new configurations featuring the 8358 SE and 8360 SE devices.
The new devices are benchmarking particularly well in terms of floating point performance, a traditional Opteron strength, Demski said, but the two new 8300 parts have also nailed the highest SAP-SD two-tier score in eight-socket servers and AMD is touting the 8358 SE and 8360 SE for data center scalability.
AMD also released three new Barcelona processors, codenamed Budapest, for single-socket servers earlier in June, bringing the chip maker's range of quad-core Opteron products a lot closer to the comprehensiveness anticipated last year, before the TLB errata was discovered in the original silicon.
The three single-socket chips that make up AMD's new Opteron 1300 series are all quad-cores, comprising the 2.1GHz Opteron 1352 ($209), the 2.2GHz 1354 ($255) and the 2.3GHz 1356 ($377).
Intel waiting for a hit product to get MID market rolling
Intel is counting on the emergence of a hit product to kick-start demand for mobile Internet devices (MID) and spur sales of its Centrino Atom chip platform.
"By the end of this year, you will have seen a whole bunch of new MIDs coming out and we'll see which ones are hits," said Sean Maloney, executive vice president and general manager of Intel's Sales and Marketing Group. "You only need one hit."
Centrino Atom is based on the Silverthorne version of the Atom processor and includes Intel's System Controller Hub, a single-chip chipset that functions as the device's central nervous system. The chips were designed to be used in handheld computers that can be used to access the Internet and play multimedia files.
So far, very few devices based on Centrino Atom have been announced and given a release date; the few devices that have been announced are priced far above the $500 target price that Intel set for these devices. The rarity of these devices means Centrino Atom shipments are also low.
The sluggish pace of the MID market stands in sharp contrast to low-cost laptops and desktops based on the Diamondville version of Atom, which uses a traditional two-chip chipset. At the Computex exhibition in Taipei last week, hardware makers were falling over themselves to showcase their latest laptops and desktops based on Atom -- and complaining they can't get enough of the chips from Intel.
There wasn't the same level of activity in the MID segment, although several prototype devices were on display. Despite that slower pace, Maloney said the MID category will get traction in the market eventually.
"As yet, we don't have any hit devices.The product's only just come out and there are a bunch of people announcing products, but the smaller a device gets the more it becomes a fashion item," Maloney said, alluding to the challenge hardware makers face when trying to gauge consumer tastes.
"It's very difficult for a CPU company to pick winners there. It's not really so much in our gene pool," he said.
Laptops and desktops based on Atom -- which Intel calls netbooks and nettops, respectively -- are a different matter. Hardware makers and Intel can draw on years of experience with these products, including the first version of Asustek Computers' Eee PC, to create devices that appeal to users.
"The netbook is more predictable, it's really like a notebook. With a MID, there's much more experimental design in it," Maloney said.
"By the end of this year, you will have seen a whole bunch of new MIDs coming out and we'll see which ones are hits," said Sean Maloney, executive vice president and general manager of Intel's Sales and Marketing Group. "You only need one hit."
Centrino Atom is based on the Silverthorne version of the Atom processor and includes Intel's System Controller Hub, a single-chip chipset that functions as the device's central nervous system. The chips were designed to be used in handheld computers that can be used to access the Internet and play multimedia files.
So far, very few devices based on Centrino Atom have been announced and given a release date; the few devices that have been announced are priced far above the $500 target price that Intel set for these devices. The rarity of these devices means Centrino Atom shipments are also low.
The sluggish pace of the MID market stands in sharp contrast to low-cost laptops and desktops based on the Diamondville version of Atom, which uses a traditional two-chip chipset. At the Computex exhibition in Taipei last week, hardware makers were falling over themselves to showcase their latest laptops and desktops based on Atom -- and complaining they can't get enough of the chips from Intel.
There wasn't the same level of activity in the MID segment, although several prototype devices were on display. Despite that slower pace, Maloney said the MID category will get traction in the market eventually.
"As yet, we don't have any hit devices.The product's only just come out and there are a bunch of people announcing products, but the smaller a device gets the more it becomes a fashion item," Maloney said, alluding to the challenge hardware makers face when trying to gauge consumer tastes.
"It's very difficult for a CPU company to pick winners there. It's not really so much in our gene pool," he said.
Laptops and desktops based on Atom -- which Intel calls netbooks and nettops, respectively -- are a different matter. Hardware makers and Intel can draw on years of experience with these products, including the first version of Asustek Computers' Eee PC, to create devices that appeal to users.
"The netbook is more predictable, it's really like a notebook. With a MID, there's much more experimental design in it," Maloney said.
AMD Wins Another Round Against Intel
The FTC has subpoenaed the chipmakers to seek information about their customers, including Dell, HP, and Apple. At issue: Intel's pricingScore another victory for Advanced Micro Devices (AMD) in its crusade to sick regulators on bigger rival Intel (INTC). The U.S. Federal Trade Commission has opened an investigation into chipmaker Intel and how its conduct affects AMD.
Intel and AMD said on June 6 that they had received subpoenas that give the FTC the leeway to seek information from the chipmakers' customers, including computer makers Dell (DELL), Hewlett-Packard (HPQ), Toshiba, and Apple (AAPL).
The investigation is focused on Intel's dominance of the market for microprocessors, which make up the heart of a PC. In particular, the FTC wants more information on Intel's practice of offering favorable pricing on chips to certain customers.
The move follows a 2005 lawsuit filed by AMD (BusinessWeek.com, 6/28/05) that accuses Intel of engaging in an aggressive and global campaign to shut AMD out of the marketplace. AMD says Intel's main weapon of choice is the regular use of volume discounts that it claims Intel has sometimes rescinded when PC makers begin adding AMD chips to their product lineup.
That case is expected to come to trial in a Delaware court in early 2010.
Intel Cites Fierce Competition
Intel was already the subject of an informal FTC inquiry in September, 2007, and said it would cooperate with the commission's more formal investigation. In a statement, Intel described its business practices as "well within U.S. law," and said the microprocessor business is so "fiercely competitive" that prices on chips have fallen by more than 42% from 2000 to 2007.
Intel says AMD's problems in the marketplace have more to do with its ability to ship price-competitive products. "There are only two ways to look at this," says Intel Chief Counsel Bruce Sewell. "There is nothing in the record to suggest that Intel is pricing its products below cost. So if Intel is pricing above cost, then either AMD can meet those prices because it is equally efficient or it can't because it is less efficient."
That argument drew a sharp rebuke from Tom McCoy, AMD's executive vice-president for legal affairs. "This case isn't about discounts or even about pricing," McCoy says. "It's about market foreclosure. If it were only about pricing and discounts then Intel would be winning with the regulators around the world."
Korea Imposes $25 Million Fine on Intel
Indeed, some regulators haven't been too sympathetic to Intel. The Korea Fair Trade Commission ruled on June 5 that Intel had violated that country's antitrust laws by offering rebates to Korean PC makers Samsung and Trigem between 2002 and 2005 in exchange for their agreement not to use AMD chips. The commission imposed a $25 million fine and ordered Intel to stop offering the rebates. Intel has said it will likely appeal the decision. Antitrust regulators in Japan and the European Union also have come out against Intel, saying the company violated laws by paying PC makers not to use AMD chips.
Intel's defense may not hold much water with the FTC either, says David Balto, a former policy director for the FTC who's now an attorney in Washington. "These discounts are handcuffs dressed up as benefits, and they have prevented retailers and others from making choices they would have otherwise made in a free market," Balto says.
During 2005 and 2006, AMD won a good portion of business away from Intel in the server chip market (BusinessWeek.com, 3/6/06). It also landed its chips in computers from both Dell and Toshiba, both of which for many years had used only Intel chips (BusinessWeek.com, 5/30/07).
PC Vendors Not So Fearful of Intel
McCoy says at least some of those successes came in the wake of regulatory action. PC vendors are proving more willing to use AMD chips because they're less afraid of retaliation from Intel, he says. "There is an emerging consensus around the world that Intel is coercing customers to avoid doing business with us and that this is hurting consumer choice," McCoy says. More recently, AMD has floundered (BusinessWeek.com, 4/8/08).
This is not the first time Intel has tangled with the FTC. In the late 1990s, the FTC accused Intel of withholding key technical information about its chips from Digital Equipment Corp. and Compaq, (both now units of Hewlett-Packard) as well as Intergraph. The three all had key patents related to microprocessors for which Intel had sought licenses. When the companies tried to enforce their patents, Intel held back information the companies would need to build new computers as a means of coercing them to license the patents in question.
Intel settled in 1999 with only a day before trial was scheduled to begin. Settlement talks began after a tennis game between Intel attorney Michael Sohn and then FTC antitrust chief William Baer.
Since then, especially in the wake of the Justice Dept.'s case against Microsoft (MSFT), antitrust case law has been altered in such a way that the FTC is more likely to play hardball with Intel: "This time I think they're going to take a cold hard look at Intel's conduct," Balto says. "The first time antitrust law didn't recognize the potential anticompetitive harm from this kind of conduct and how it might be illegal."
Intel and AMD said on June 6 that they had received subpoenas that give the FTC the leeway to seek information from the chipmakers' customers, including computer makers Dell (DELL), Hewlett-Packard (HPQ), Toshiba, and Apple (AAPL).
The investigation is focused on Intel's dominance of the market for microprocessors, which make up the heart of a PC. In particular, the FTC wants more information on Intel's practice of offering favorable pricing on chips to certain customers.
The move follows a 2005 lawsuit filed by AMD (BusinessWeek.com, 6/28/05) that accuses Intel of engaging in an aggressive and global campaign to shut AMD out of the marketplace. AMD says Intel's main weapon of choice is the regular use of volume discounts that it claims Intel has sometimes rescinded when PC makers begin adding AMD chips to their product lineup.
That case is expected to come to trial in a Delaware court in early 2010.
Intel Cites Fierce Competition
Intel was already the subject of an informal FTC inquiry in September, 2007, and said it would cooperate with the commission's more formal investigation. In a statement, Intel described its business practices as "well within U.S. law," and said the microprocessor business is so "fiercely competitive" that prices on chips have fallen by more than 42% from 2000 to 2007.
Intel says AMD's problems in the marketplace have more to do with its ability to ship price-competitive products. "There are only two ways to look at this," says Intel Chief Counsel Bruce Sewell. "There is nothing in the record to suggest that Intel is pricing its products below cost. So if Intel is pricing above cost, then either AMD can meet those prices because it is equally efficient or it can't because it is less efficient."
That argument drew a sharp rebuke from Tom McCoy, AMD's executive vice-president for legal affairs. "This case isn't about discounts or even about pricing," McCoy says. "It's about market foreclosure. If it were only about pricing and discounts then Intel would be winning with the regulators around the world."
Korea Imposes $25 Million Fine on Intel
Indeed, some regulators haven't been too sympathetic to Intel. The Korea Fair Trade Commission ruled on June 5 that Intel had violated that country's antitrust laws by offering rebates to Korean PC makers Samsung and Trigem between 2002 and 2005 in exchange for their agreement not to use AMD chips. The commission imposed a $25 million fine and ordered Intel to stop offering the rebates. Intel has said it will likely appeal the decision. Antitrust regulators in Japan and the European Union also have come out against Intel, saying the company violated laws by paying PC makers not to use AMD chips.
Intel's defense may not hold much water with the FTC either, says David Balto, a former policy director for the FTC who's now an attorney in Washington. "These discounts are handcuffs dressed up as benefits, and they have prevented retailers and others from making choices they would have otherwise made in a free market," Balto says.
During 2005 and 2006, AMD won a good portion of business away from Intel in the server chip market (BusinessWeek.com, 3/6/06). It also landed its chips in computers from both Dell and Toshiba, both of which for many years had used only Intel chips (BusinessWeek.com, 5/30/07).
PC Vendors Not So Fearful of Intel
McCoy says at least some of those successes came in the wake of regulatory action. PC vendors are proving more willing to use AMD chips because they're less afraid of retaliation from Intel, he says. "There is an emerging consensus around the world that Intel is coercing customers to avoid doing business with us and that this is hurting consumer choice," McCoy says. More recently, AMD has floundered (BusinessWeek.com, 4/8/08).
This is not the first time Intel has tangled with the FTC. In the late 1990s, the FTC accused Intel of withholding key technical information about its chips from Digital Equipment Corp. and Compaq, (both now units of Hewlett-Packard) as well as Intergraph. The three all had key patents related to microprocessors for which Intel had sought licenses. When the companies tried to enforce their patents, Intel held back information the companies would need to build new computers as a means of coercing them to license the patents in question.
Intel settled in 1999 with only a day before trial was scheduled to begin. Settlement talks began after a tennis game between Intel attorney Michael Sohn and then FTC antitrust chief William Baer.
Since then, especially in the wake of the Justice Dept.'s case against Microsoft (MSFT), antitrust case law has been altered in such a way that the FTC is more likely to play hardball with Intel: "This time I think they're going to take a cold hard look at Intel's conduct," Balto says. "The first time antitrust law didn't recognize the potential anticompetitive harm from this kind of conduct and how it might be illegal."
British Agency Reverses Stand on Restricting Lucentis for AMD
LONDON, January 2008 — A British health advisory agency has reversed an earlier cost-saving proposal that allows government reimbursement for Lucentis treatments only for the second eye of individuals diagnosed with advanced or "wet" age-related macular degeneration (AMD), which can cause blindness.
The National Institute for Health and Clinical Excellence (NICE) has issued a revised proposal, with "draft" status ending Jan. 14, which would allow Lucentis to be injected in any affected eye of British citizens when treatments are capped at 14 per eye.
If additional treatments are needed, NICE officials say the Lucentis manufacturer (Genentech, San Francisco) has agreed to provide additional doses at no cost.
NICE officials say Lucentis is an effective though expensive drug capable of stabilizing vision loss in most people with advanced AMD. About 25 percent of affected eyes achieve vision improvement after regular Lucentis treatments, which cost more than $2,000 per dose.
In another recent development, Genentech announced that a much cheaper version of Lucentis, known as Avastin, will be sold directly to any physicians who choose to use the drug off-label for treating AMD. Avastin is not FDA-approved as an AMD treatment.
Late last year, Genentech made the controversial decision to cease selling Avastin directly to compounding pharmacies that had been repackaging the cancer drug for use as an AMD treatment.
In its recent announcement, Genentech said physicians choosing to use Avastin as an AMD treatment can purchase the drug directly and designate where they want it shipped — including to compounding pharmacies.
The National Institute for Health and Clinical Excellence (NICE) has issued a revised proposal, with "draft" status ending Jan. 14, which would allow Lucentis to be injected in any affected eye of British citizens when treatments are capped at 14 per eye.
If additional treatments are needed, NICE officials say the Lucentis manufacturer (Genentech, San Francisco) has agreed to provide additional doses at no cost.
NICE officials say Lucentis is an effective though expensive drug capable of stabilizing vision loss in most people with advanced AMD. About 25 percent of affected eyes achieve vision improvement after regular Lucentis treatments, which cost more than $2,000 per dose.
In another recent development, Genentech announced that a much cheaper version of Lucentis, known as Avastin, will be sold directly to any physicians who choose to use the drug off-label for treating AMD. Avastin is not FDA-approved as an AMD treatment.
Late last year, Genentech made the controversial decision to cease selling Avastin directly to compounding pharmacies that had been repackaging the cancer drug for use as an AMD treatment.
In its recent announcement, Genentech said physicians choosing to use Avastin as an AMD treatment can purchase the drug directly and designate where they want it shipped — including to compounding pharmacies.
Smoking Increases Macular Degeneration Risk by Nearly 50 Percent
BEAVER DAM, Wis., January 2008 — Smoking boosts risks of developing age-related macular degeneration (AMD) by nearly 50 percent, reveals a recent analysis of about 3,500 Beaver Dam, Wis. residents undergoing eye-related studies since 1987. The finding was reported in the January issue of Archives of Ophthalmology.
Study results also show that, of people who have macular degeneration, smokers on average were more likely to develop the eye disease at a younger age:
• Current smokers (69.2 years)
• Former smokers (72.3 years)
• Non-smokers who have never smoked (74.4 years)
Researchers noted that it's difficult in modern times to assess eye damage caused by secondary exposure to smoke, partly because of new laws that now restrict smoking in public areas.
Study results also show that, of people who have macular degeneration, smokers on average were more likely to develop the eye disease at a younger age:
• Current smokers (69.2 years)
• Former smokers (72.3 years)
• Non-smokers who have never smoked (74.4 years)
Researchers noted that it's difficult in modern times to assess eye damage caused by secondary exposure to smoke, partly because of new laws that now restrict smoking in public areas.
African Americans May Have Less Risk for
BALTIMORE, February 2008 — African Americans may have less risk of developing advanced macular degeneration (AMD) than Caucasians, according to recent findings by Johns Hopkins School of Medicine scientists.
Early signs of AMD include formation of yellowish spots (drusen) in the inner back of the eye (retina), where vision processing occurs. In a study involving 2,520 participants of average ages of 73.5 years, researchers found that larger drusen and those closer to the eye's center were more likely to occur in Caucasians than in African Americans.
Study authors concluded that more Caucasians than African Americans were likely to develop advanced and potentially blinding forms of macular degeneration. Researchers said more research is needed to determine why African Americans are less likely to have significant degenerative changes associated with AMD, such as those that occur in central portions of the retina.
Early signs of AMD include formation of yellowish spots (drusen) in the inner back of the eye (retina), where vision processing occurs. In a study involving 2,520 participants of average ages of 73.5 years, researchers found that larger drusen and those closer to the eye's center were more likely to occur in Caucasians than in African Americans.
Study authors concluded that more Caucasians than African Americans were likely to develop advanced and potentially blinding forms of macular degeneration. Researchers said more research is needed to determine why African Americans are less likely to have significant degenerative changes associated with AMD, such as those that occur in central portions of the retina.
Discovery May Lead to Therapies for "Dry" AMD
MONTREAL, February 2008 — Researchers in Montreal have identified an underlying cause of the "dry" form of age-related macular degeneration (AMD), a discovery that could lead to development of new therapies.
Sainte-Justine Hospital and Universite de Montreal scientists said they now have a better understanding of processes that cause accumulations of lipids (fats) in the inner back of the eye (retina), which can lead to eye damage associated with macular degeneration.
"This discovery brings us one step closer to treating dry AMD, which could significantly improve the quality of life of seniors who are most affected by this eye disease," said Universite de Montreal professor and study co-author Huy Ong, PhD.
Study results were published in the February issue of the medical journal PLoS Medicine.
Sainte-Justine Hospital and Universite de Montreal scientists said they now have a better understanding of processes that cause accumulations of lipids (fats) in the inner back of the eye (retina), which can lead to eye damage associated with macular degeneration.
"This discovery brings us one step closer to treating dry AMD, which could significantly improve the quality of life of seniors who are most affected by this eye disease," said Universite de Montreal professor and study co-author Huy Ong, PhD.
Study results were published in the February issue of the medical journal PLoS Medicine.
T2K Open Supercomputer Systems with Quad-Core AMD Opteron(TM) Processors Now in Operation
AMD Japan (NYSE: AMD) today announced the official start of operation of the T2K supercomputer systems featuring Quad-Core AMD Opteron™ processors and with specifications developed jointly by the University of Tsukuba, the University of Tokyo, and Kyoto University. These T2K systems were delivered to the University of Tsukuba by Cray Japan Inc. and Sumisho Computer Systems, to the University of Tokyo by Hitachi Ltd., and to Kyoto University by Fujitsu Ltd. "In July 2006, these three universities began jointly developing common specifications for each university’s next-generation supercomputer with an eye toward using their supercomputers collaboratively,” said Shunsuke Yoshizawa, director, Regional Marketing, AMD Japan. "The universities aimed to adopt the most advanced and highest performance technology as early as possible, and have engaged in three pillars of openness that make up this new shared specifications experiment: the use of open source hardware architecture, open source system software, and their desire to make the supercomputers widely available amongst the three institutions. The performance, scalability and advance floating point processing capability of the Quad-Core AMD Opteron processor easily matched their requirements.” The University of Tokyo supercomputer system is comprised of 952 nodes of the Hitachi HA8000-tc/RS425 technical server with four Quad-Core AMD Opteron processors per node. The system as designed achieves a theoretical peak performance of approximately 140 teraflops. A teraflop or "tflop” is the computing power required to process one trillion floating point operations per second. This theoretical peak performance is the fastest in Japan at the time of launch. The University of Tsukuba’s system was built by Cray Japan Inc. and Sumisho Computer Systems and is based on Appro International’s Xtreme-X Supercomputer, while the Kyoto University’s system was built around the Fujitsu HX600 HPC server. Both supercomputers are equipped with four Quad-Core AMD Opteron processors per node. The University of Tsukuba’s system can achieve a theoretical peak performance of approximately 95 tflops, while the Kyoto University system tops out at a theoretical peak performance of approximately 61 tflops. The T2K supercomputers will provide these universities and their researchers with the outstanding system performance and advanced floating point processing enabled by Quad-core AMD Opteron processors. The selection of AMD technology for these highly advanced systems can also provide the universities with much-needed energy efficiency in the form of performance-per-watt and innovative processor-level power management features such as AMD CoolCore™ Technology and Independent Dynamic Core Technology. The synergy of advanced research expertise in computer science developed over many years at these three universities and the computing power of the new AMD processor-based T2K supercomputers is expected to help address increasingly diverse needs in industry and academia including the large-scale scientific calculations for researching subatomic particles and nuclear energy, astronomy, climate modeling and weather forecasting, and genetics and biomedical advancements, among many others. Additional information about T2K can be found at http://www.open-supercomputer.org. Visit AMD on the Web For more information on the Quad-Core AMD Opteron processor family, including FAQs, development tools, and other general information, please visit http://multicore.amd.com. Information on Quad-Core AMD Opteron processor pricing can be found at http://www.amd.com/pricing. An online press kit is available at http://www.amd.com/quadcore/presskit. About AMD Advanced Micro Devices (NYSE: AMD) is a leading global provider of innovative processing solutions in the computing, graphics and consumer electronics markets. AMD is dedicated to driving open innovation, choice and industry growth by delivering superior customer-centric solutions that empower consumers and businesses worldwide. For more information, visit http://www.amd.com. AMD, the AMD Arrow logo, AMD Opteron and combinations thereof, are trademarks of Advanced Micro Devices, Inc. Other names are for informational purposes only and may be trademarks of their respective owners.
NComputing, AMD and BRAC Join Forces to Expand Computing in Bangladesh
NComputing, AMD (NYSE: AMD) and BRAC announced today the completion of a collaborative project to deploy innovative technology solutions in 10 new school-based 50x15 Initiative Learning Labs throughout Bangladesh. BRAC, one of the world's largest non-governmental organizations, deployed NComputing virtual desktops and AMD Athlon(TM) 64 X2 Dual-Core processor-based desktop computers in 10 schools around the country as part of the 50x15 Initiative, a global initiative founded by AMD with the goal of providing computing capabilities and Internet connectivity to 50 percent of the world's population by 2015.
NComputing, a leading provider of desktop virtualization software and hardware, and AMD, a leading global provider of innovative processing solutions in computing, graphics and consumer electronics, are collaborating to empower people in emerging countries with tools and skills to help them succeed in today's modern world. Similar to previous 50x15 deployments, BRAC followed AMD's proven learning lab methodology, which includes providing tools, techniques and training to maximize educational impact. BRAC, NComputing, and AMD plan to apply the successes and lessons learned in these labs to future deployments.
The AMD/NComputing solution makes sense for budget-constrained areas. AMD is known for its work in energy efficiency throughout the world, and NComputing virtual desktop software and hardware allow the power -- and cost -- to be shared by multiple users.
BRAC organized the deployment of the Learning Labs, which now provide computing access to thousands of children in 10 cities. BRAC's Internet Service Provider arm, BRACNet, provided connectivity solutions to provide the locations with wireless Internet service.
"The deployment of 50x15 Learning Labs throughout Bangladesh will help this country to become a bigger part of the information society," said Abdul-Muyeed Chowdhury, chairman of BRAC BDMail Network Ltd (BBN). "Our goal is to provide PCs to 1,000 libraries and 15,000 non-formal primary schools all over the country. With NComputing's virtual desktop technology and the 50x15 Learning Lab model, we will be one step closer to bridging the digital divide that exists between developed countries and emerging nations, such as Bangladesh."
With the NComputing X300, seven users can simultaneously share a single computer, while the company's L-series supports up to 10 users on a basic computer, 30 on a mid-range system, and hundreds on enterprise-class servers. A breakthrough in green computing, NComputing access devices consume as little as 1.16 watts per user at idle, versus as much as 115 watts for some PCs. Support costs can be expected to plummet because fewer PCs are used (compared to one PC per user) and the solid-state NComputing access devices require little or no maintenance.
BRAC selected the AMD/NComputing solution because of the low costs of deployment, maintenance, support and upgrade along with the energy efficiency of the solution.
"NComputing believes that providing PC access to the next billion users is the single biggest challenge facing our industry today," said Stephen Dukker, chairman and CEO of NComputing. "Emerging countries make up the majority of this under-served mass market and we are gratified that NComputing's technology can be an important part of the solution in Bangladesh and around the world."
NComputing solutions have already been deployed by 15,000 organizations in over 70 countries, helping to slash their computing and electric consumption costs. Calculators for budget, total costs of ownership, and environmental impact are available on NComputing's website at www.ncomputing.com. One scenario is as follows: based on the one-PC-per-user model, seven computers that cost $700 each would cost a total of $4,900 for seven users. But a single $700 computer can support 7 users with 2 NComputing X300 kits. At a price of about $210 per X300 kit, that means a total cost for the hardware and NComputing virtualization software of $1,120 for 7 users ($700 for the computer plus $210 each for the X300 kits, for a total of $1,120). This calculation does not include keyboards, mice, monitors and software licenses, but it shows the potential for enormous savings.
"Learning Labs are deliberate, measured, strategic deployments of technology solutions used to gain insight and knowledge on how best to foster digital inclusion worldwide," said Dan Shine, vice president, 50x15 Initiative AMD and president, 50x15 Foundation. "This expansive project to create 10 new learning labs truly captures the spirit and vision of the 50x15 Initiative by bringing together global collaborators and developing an innovative solution based on leading-edge technology to provide Internet connectivity and computing capability to a new generation."
About NComputing
Winner of The Wall Street Journal's Technology Innovation Award, NComputing, Inc. was founded with the goal of making desktop computing affordable for everyone. Headquartered in Redwood City, CA, NComputing is a privately held virtualization software and hardware company. The company's patented technology can help lower desktop computing costs, improves manageability, and reduces both energy consumption and e-waste.
About BRAC
Established in 1972 as a small-scale relief and rehabilitation organization, BRAC today is one of the largest NGOs in the world, pioneering a multi-faceted approach to development with programs in health, education and social development along with its economic development initiatives, which include microfinance. BRAC reaches an estimated 110 million people in all 64 districts of Bangladesh and in recent years has stepped beyond the borders to countries in distress like Afghanistan, Sri Lanka, Uganda and Tanzania.
NComputing, a leading provider of desktop virtualization software and hardware, and AMD, a leading global provider of innovative processing solutions in computing, graphics and consumer electronics, are collaborating to empower people in emerging countries with tools and skills to help them succeed in today's modern world. Similar to previous 50x15 deployments, BRAC followed AMD's proven learning lab methodology, which includes providing tools, techniques and training to maximize educational impact. BRAC, NComputing, and AMD plan to apply the successes and lessons learned in these labs to future deployments.
The AMD/NComputing solution makes sense for budget-constrained areas. AMD is known for its work in energy efficiency throughout the world, and NComputing virtual desktop software and hardware allow the power -- and cost -- to be shared by multiple users.
BRAC organized the deployment of the Learning Labs, which now provide computing access to thousands of children in 10 cities. BRAC's Internet Service Provider arm, BRACNet, provided connectivity solutions to provide the locations with wireless Internet service.
"The deployment of 50x15 Learning Labs throughout Bangladesh will help this country to become a bigger part of the information society," said Abdul-Muyeed Chowdhury, chairman of BRAC BDMail Network Ltd (BBN). "Our goal is to provide PCs to 1,000 libraries and 15,000 non-formal primary schools all over the country. With NComputing's virtual desktop technology and the 50x15 Learning Lab model, we will be one step closer to bridging the digital divide that exists between developed countries and emerging nations, such as Bangladesh."
With the NComputing X300, seven users can simultaneously share a single computer, while the company's L-series supports up to 10 users on a basic computer, 30 on a mid-range system, and hundreds on enterprise-class servers. A breakthrough in green computing, NComputing access devices consume as little as 1.16 watts per user at idle, versus as much as 115 watts for some PCs. Support costs can be expected to plummet because fewer PCs are used (compared to one PC per user) and the solid-state NComputing access devices require little or no maintenance.
BRAC selected the AMD/NComputing solution because of the low costs of deployment, maintenance, support and upgrade along with the energy efficiency of the solution.
"NComputing believes that providing PC access to the next billion users is the single biggest challenge facing our industry today," said Stephen Dukker, chairman and CEO of NComputing. "Emerging countries make up the majority of this under-served mass market and we are gratified that NComputing's technology can be an important part of the solution in Bangladesh and around the world."
NComputing solutions have already been deployed by 15,000 organizations in over 70 countries, helping to slash their computing and electric consumption costs. Calculators for budget, total costs of ownership, and environmental impact are available on NComputing's website at www.ncomputing.com. One scenario is as follows: based on the one-PC-per-user model, seven computers that cost $700 each would cost a total of $4,900 for seven users. But a single $700 computer can support 7 users with 2 NComputing X300 kits. At a price of about $210 per X300 kit, that means a total cost for the hardware and NComputing virtualization software of $1,120 for 7 users ($700 for the computer plus $210 each for the X300 kits, for a total of $1,120). This calculation does not include keyboards, mice, monitors and software licenses, but it shows the potential for enormous savings.
"Learning Labs are deliberate, measured, strategic deployments of technology solutions used to gain insight and knowledge on how best to foster digital inclusion worldwide," said Dan Shine, vice president, 50x15 Initiative AMD and president, 50x15 Foundation. "This expansive project to create 10 new learning labs truly captures the spirit and vision of the 50x15 Initiative by bringing together global collaborators and developing an innovative solution based on leading-edge technology to provide Internet connectivity and computing capability to a new generation."
About NComputing
Winner of The Wall Street Journal's Technology Innovation Award, NComputing, Inc. was founded with the goal of making desktop computing affordable for everyone. Headquartered in Redwood City, CA, NComputing is a privately held virtualization software and hardware company. The company's patented technology can help lower desktop computing costs, improves manageability, and reduces both energy consumption and e-waste.
About BRAC
Established in 1972 as a small-scale relief and rehabilitation organization, BRAC today is one of the largest NGOs in the world, pioneering a multi-faceted approach to development with programs in health, education and social development along with its economic development initiatives, which include microfinance. BRAC reaches an estimated 110 million people in all 64 districts of Bangladesh and in recent years has stepped beyond the borders to countries in distress like Afghanistan, Sri Lanka, Uganda and Tanzania.
AMD Foundation Launches ''AMD Changing the Game''
NEW YORK--(BUSINESS WIRE)--AMD (NYSE: AMD) today announced the launch of AMD Changing the Game, the first initiative of the newly formed AMD Foundation. The AMD Changing the Game program is intended to improve critical technical and life skills by teaching kids to develop digital games with social content. The program is rooted in AMD’s commitment and experience in supporting education with the company’s passion and expertise in the gaming industry. The program’s launch accompanies AMD’s sponsorship and participation at the Fifth Annual Games for Change Festival to be held June 3 - 4 at Parsons The New School for Design in New York City.
AMD Changing the Game is a natural fit for AMD, which features products powering the visual experience of the two most popular gaming consoles in the world today1 and which recently launched AMD GAME!, a program designed to help consumers select perfectly suited PCs for high-definition gaming.
“We have a tremendous opportunity to harness the passion that kids have for gaming while teaching the skills they need to be successful in our 21st Century digital economy,” said Dirk Meyer, AMD president and chief operating officer.
In addition to technical skills such as science, technology, engineering and math, digital games can be used to help teach youth how to be more engaged citizens, to see conflict from another’s viewpoint and find positive ways to respond to challenging social issues such as poverty, hunger, disease, energy conservation, water use and global warming.
“Today’s youth are highly concerned about social issues and the current generation of youth gamers is among the most socially conscious in history,” said Suzanne Seggerman, co-founder and president of Games for Change. “The movement toward educating and engaging youth through digital games for change not only raises awareness of the importance of social issues, but gives youth an opportunity to make a difference. As this movement continues to build momentum, we believe that partnerships and grant support from leading technology companies like AMD will be critical to its success.”
Through AMD Changing the Game, AMD Foundation grants will go to nonprofit organizations that inspire young people learn while creating games with social content. Employees will also support the initiative through volunteer opportunities. In its pilot year, the following organizations will be funded:
Girlstart, is an Austin, TX-based nonprofit organization created to empower middle and high school girls to excel in math, science, and technology. In the summer of 2008, the AMD Foundation’s grant will enable 60 Girlstart participants to attend a program focused on games with social content. As a capstone project, girls will be creating a social awareness event in Teen Second Life, a virtual gathering place for teens 13-17 all over the world to make friends, play, learn and create. The Girlstart team will identify a social issue of importance to them and create an event in Teen Second Life that will help raise awareness and inspire action around the issue.
Global Kids, is a Brooklyn, NY-based nonprofit organization that seeks to transform urban youth into successful students and global and community leaders. Through its grant to Global Kids’ Playing for Keeps program, AMD has joined The Microsoft Corporation in enabling 20 young people from underserved communities to work with game developers to develop, create and distribute a game about the heroic role of residents following Hurricane Katrina. Last year, young people worked with developers in the Playing for Keeps program to create the game Ayiti: The Cost of Life which allows players to assume the role of impoverished people living in rural Haiti with the goal of meeting some key health, education and quality of life challenges.
Institute for Urban Game Design, is a Washington, DC-based nonprofit organization teaching science, technology, engineering and math (STEM) skills through the hands-on creation of digital games. Beginning in the summer of 2008, the AMD Foundation’s grant will enable IUGD participants to apply their learning in 3-D modeling, animation and computer programming to the development of a game focused on the issue of energy usage. Students will learn about and explore the social issues associated with different types of energy.
Science Buddies is a national, non-profit organization based in California's Silicon Valley offering a variety of web-based tools that help K-12 students explore science through research-based projects often done at Science Fairs and other school and community events. AMD Foundation’s grant will enable Science Buddies to launch a Video and Computer Games Interest Area on its site aimed at helping students understand and practice what is required to design digital games. AMD volunteers will work with Science Buddies staff scientists to develop project ideas to spark student interest in exploring topics such as human behavior in games, ergonomics, game design and programming and the incorporation of social or educational content in games.
As part of launching the Games for Change Festival, the AMD Foundation, in partnership with the John G. and Catherine T. MacArthur Foundation, is sponsoring “Let the Games Begin,” a day-long workshop for nonprofit organizations focusing on how to create social issue games. The workshop will feature interactive lectures by some of the nation’s leading authorities on social issue game development and cover fundamentals such as game design, fundraising, evaluation, youth participation, distribution and press strategies.
Following the Festival, the AMD Foundation and Games for Change plan to co-produce a how-to digital toolkit for nonprofits that includes examples of games with social content, interviews with key experts and additional guidance for nonprofits creating social issue games for the first time.
AMD is also working with PETLab, a joint project of Games for Change and Parsons The News School, to create a social issue game development curriculum for youth. The curriculum is expected to be piloted in the fall of 2008.
AMD Changing the Game is a natural fit for AMD, which features products powering the visual experience of the two most popular gaming consoles in the world today1 and which recently launched AMD GAME!, a program designed to help consumers select perfectly suited PCs for high-definition gaming.
“We have a tremendous opportunity to harness the passion that kids have for gaming while teaching the skills they need to be successful in our 21st Century digital economy,” said Dirk Meyer, AMD president and chief operating officer.
In addition to technical skills such as science, technology, engineering and math, digital games can be used to help teach youth how to be more engaged citizens, to see conflict from another’s viewpoint and find positive ways to respond to challenging social issues such as poverty, hunger, disease, energy conservation, water use and global warming.
“Today’s youth are highly concerned about social issues and the current generation of youth gamers is among the most socially conscious in history,” said Suzanne Seggerman, co-founder and president of Games for Change. “The movement toward educating and engaging youth through digital games for change not only raises awareness of the importance of social issues, but gives youth an opportunity to make a difference. As this movement continues to build momentum, we believe that partnerships and grant support from leading technology companies like AMD will be critical to its success.”
Through AMD Changing the Game, AMD Foundation grants will go to nonprofit organizations that inspire young people learn while creating games with social content. Employees will also support the initiative through volunteer opportunities. In its pilot year, the following organizations will be funded:
Girlstart, is an Austin, TX-based nonprofit organization created to empower middle and high school girls to excel in math, science, and technology. In the summer of 2008, the AMD Foundation’s grant will enable 60 Girlstart participants to attend a program focused on games with social content. As a capstone project, girls will be creating a social awareness event in Teen Second Life, a virtual gathering place for teens 13-17 all over the world to make friends, play, learn and create. The Girlstart team will identify a social issue of importance to them and create an event in Teen Second Life that will help raise awareness and inspire action around the issue.
Global Kids, is a Brooklyn, NY-based nonprofit organization that seeks to transform urban youth into successful students and global and community leaders. Through its grant to Global Kids’ Playing for Keeps program, AMD has joined The Microsoft Corporation in enabling 20 young people from underserved communities to work with game developers to develop, create and distribute a game about the heroic role of residents following Hurricane Katrina. Last year, young people worked with developers in the Playing for Keeps program to create the game Ayiti: The Cost of Life which allows players to assume the role of impoverished people living in rural Haiti with the goal of meeting some key health, education and quality of life challenges.
Institute for Urban Game Design, is a Washington, DC-based nonprofit organization teaching science, technology, engineering and math (STEM) skills through the hands-on creation of digital games. Beginning in the summer of 2008, the AMD Foundation’s grant will enable IUGD participants to apply their learning in 3-D modeling, animation and computer programming to the development of a game focused on the issue of energy usage. Students will learn about and explore the social issues associated with different types of energy.
Science Buddies is a national, non-profit organization based in California's Silicon Valley offering a variety of web-based tools that help K-12 students explore science through research-based projects often done at Science Fairs and other school and community events. AMD Foundation’s grant will enable Science Buddies to launch a Video and Computer Games Interest Area on its site aimed at helping students understand and practice what is required to design digital games. AMD volunteers will work with Science Buddies staff scientists to develop project ideas to spark student interest in exploring topics such as human behavior in games, ergonomics, game design and programming and the incorporation of social or educational content in games.
As part of launching the Games for Change Festival, the AMD Foundation, in partnership with the John G. and Catherine T. MacArthur Foundation, is sponsoring “Let the Games Begin,” a day-long workshop for nonprofit organizations focusing on how to create social issue games. The workshop will feature interactive lectures by some of the nation’s leading authorities on social issue game development and cover fundamentals such as game design, fundraising, evaluation, youth participation, distribution and press strategies.
Following the Festival, the AMD Foundation and Games for Change plan to co-produce a how-to digital toolkit for nonprofits that includes examples of games with social content, interviews with key experts and additional guidance for nonprofits creating social issue games for the first time.
AMD is also working with PETLab, a joint project of Games for Change and Parsons The News School, to create a social issue game development curriculum for youth. The curriculum is expected to be piloted in the fall of 2008.
Intel and Nvidia headed for licensing stand-off
This started out as a rumor today, but Intel has since more-or-less confirmed that licensing discussions between Intel and Nvidia for Intel's next-generation processors are not going well and the resulting conflict could have implications for high-end gaming PCs.
The story is that Intel and Nvidia are currently negotiating technology licenses for Nehalem, Intel's next-generation desktop CPU due out at the end of this year. As we were told from multiple desktop vendors who wish to remain nameless, Intel wants to license SLI from Nvidia for its Nehalem chipsets, and if Nvidia won't, Intel will withhold the license that would enable Nvidia to support Nehalem's memory controller, and thus Nehalem, on its own chipsets.
We have no official confirmation from Nvidia on this, and Intel's statement from PR manager Dan Snyder is vague, but it lends credence to the story:
"There is a disagreement between Intel and Nvidia as to the scope of Nvidia's license from Intel to make chipsets compatible with Intel microprocessors. Intel is trying to resolve the disagreement privately with Nvidia and therefore we will not provide additional details. It is our hope that this dispute will be resolved amicably and that it will not impact other areas of our companies' working relationship."
Intel has been after SLI support for its chipsets for years, but has thus far only been able to build it into its ultra high-end Skulltrail motherboards, seemingly a one-off. With SLI available across all of its chipset lines, Intel would be able to sell motherboards that support both AMD's and Nvidia's multigraphics card technologies. Right now Intel boards (with the exception of Skulltrail) only support AMD's CrossFire.
Nvidia, on the other hand, has kept SLI close, often citing compatibility and certification concerns as the reason why no other chipset vendor has been able to offer SLI-capability. But if Nvidia loses out on Nehalem for its next-generation chipsets, the high-end desktop market will become more fragmented than it's been in years. Nvidia has been able to offer SLI-supporting chipsets for both AMD and Intel processors, but if this split happens, on one side we'll have Nehalem and CrossFire-based systems, the other will offer SLI (and possibly CrossFire, if hacks used in the past continue to work) and AMD CPUs.
In light of this rumor, Nvidia's recent marketing push encouraging upgraders to pick a graphics card before a quad-core CPU takes on new significance. If Nvidia knows high-end PC gamers will have to make a choice later this year, better to plant the seeds in its favor early. Intel probably has less to worry about, because gamers who demand SLI with a fast Intel processor can still use Nvidia's NForce 790i chipset, which supports the current generation of Intel Core 2 Extreme chips.
The story is that Intel and Nvidia are currently negotiating technology licenses for Nehalem, Intel's next-generation desktop CPU due out at the end of this year. As we were told from multiple desktop vendors who wish to remain nameless, Intel wants to license SLI from Nvidia for its Nehalem chipsets, and if Nvidia won't, Intel will withhold the license that would enable Nvidia to support Nehalem's memory controller, and thus Nehalem, on its own chipsets.
We have no official confirmation from Nvidia on this, and Intel's statement from PR manager Dan Snyder is vague, but it lends credence to the story:
"There is a disagreement between Intel and Nvidia as to the scope of Nvidia's license from Intel to make chipsets compatible with Intel microprocessors. Intel is trying to resolve the disagreement privately with Nvidia and therefore we will not provide additional details. It is our hope that this dispute will be resolved amicably and that it will not impact other areas of our companies' working relationship."
Intel has been after SLI support for its chipsets for years, but has thus far only been able to build it into its ultra high-end Skulltrail motherboards, seemingly a one-off. With SLI available across all of its chipset lines, Intel would be able to sell motherboards that support both AMD's and Nvidia's multigraphics card technologies. Right now Intel boards (with the exception of Skulltrail) only support AMD's CrossFire.
Nvidia, on the other hand, has kept SLI close, often citing compatibility and certification concerns as the reason why no other chipset vendor has been able to offer SLI-capability. But if Nvidia loses out on Nehalem for its next-generation chipsets, the high-end desktop market will become more fragmented than it's been in years. Nvidia has been able to offer SLI-supporting chipsets for both AMD and Intel processors, but if this split happens, on one side we'll have Nehalem and CrossFire-based systems, the other will offer SLI (and possibly CrossFire, if hacks used in the past continue to work) and AMD CPUs.
In light of this rumor, Nvidia's recent marketing push encouraging upgraders to pick a graphics card before a quad-core CPU takes on new significance. If Nvidia knows high-end PC gamers will have to make a choice later this year, better to plant the seeds in its favor early. Intel probably has less to worry about, because gamers who demand SLI with a fast Intel processor can still use Nvidia's NForce 790i chipset, which supports the current generation of Intel Core 2 Extreme chips.
AMD Stream SDK now supports GPU programming for Linux
The new beta version 1.1 of the AMD Stream development environment incorporates support for Linux. The software development kit (SDK) allows current AMD graphic chips to be used for carrying out processor-intensive tasks. The new Stream version is already on AMD's FTP servers, but there is no link to it yet on the corresponding web site. Besides the conventional Windows version, AMD is also offering version 1.1 beta in 32- and 64-bit versions for the Red Hat Enterprise 5.1 and Suse Linux Enterprise Server 10 SP1 Linux distributions.
AMD has added support not just for Linux, but also for the FireStream 9170 stream-processing graphics card, which is tailored for dealing with complex calculations. Microsoft's Visual Studio 2008 is also now supported, at least provisionally. The Brook+ compiler now handles integer mathematical operations and enables line numbering to simplify BR data debugging. According to AMD, several errors (not described in detail) have been eliminated, and speed has been optimized. AMD Stream contains all the tools required for programming Stream processors.
Nvidia is offering a similar suite with its CUDA (Compute Unified Device Architecture), that supports programming current Nvidia GPUs. According to Nvidia's CUDA Zone developer portal, most of the applications so far published fall within the areas of astrophysics and molecular dynamics. (jk/c't)
AMD has added support not just for Linux, but also for the FireStream 9170 stream-processing graphics card, which is tailored for dealing with complex calculations. Microsoft's Visual Studio 2008 is also now supported, at least provisionally. The Brook+ compiler now handles integer mathematical operations and enables line numbering to simplify BR data debugging. According to AMD, several errors (not described in detail) have been eliminated, and speed has been optimized. AMD Stream contains all the tools required for programming Stream processors.
Nvidia is offering a similar suite with its CUDA (Compute Unified Device Architecture), that supports programming current Nvidia GPUs. According to Nvidia's CUDA Zone developer portal, most of the applications so far published fall within the areas of astrophysics and molecular dynamics. (jk/c't)
EC denies 'provisional decision' on Intel anti-trust probe
The anti-trust arm of the European Commission today denied that it had reached any provisional decision in its ongoing probe of Intel’s marketing and sales practices.
EC competition spokesman Jonathan Todd said: "There is no provisional or internal decision on this case. The investigation is very much active and ongoing," reports Reuters.
He slammed The Financial Times Deutschland newspaper, which cited Brussels sources, for providing misleading information and described the report as "irresponsible journalism".
The newspaper had said that the European Union anti-trust authority planned to take action against Intel’s marketing and sales practices by this summer. It reckoned that decision had already been made by EU commissioner Neelie Kroes because the EC had gathered enough evidence against the world’s biggest chip maker following a lawsuit by rival AMD.
The paper also claimed that the EU would prevent Intel from marketing its microprocessors to PC vendors at a marked down price, and that the authority would forbid the company from subsidising retailer's advertising costs.
The EU has been sniffing around Intel's business practices in Europe for six years, but the investigation got a second wind in 2005 when AMD formally accused its arch-rival of using illegal tactics to dominate the PC and server markets.
In July last year the EU filed formal anti-trust charges against Intel.
Intel spokesman Chuck Mulloy told Reuters yesterday: "We continue to work with the Commission to attempt to convince them that our business practices are well within the law." ®
EC competition spokesman Jonathan Todd said: "There is no provisional or internal decision on this case. The investigation is very much active and ongoing," reports Reuters.
He slammed The Financial Times Deutschland newspaper, which cited Brussels sources, for providing misleading information and described the report as "irresponsible journalism".
The newspaper had said that the European Union anti-trust authority planned to take action against Intel’s marketing and sales practices by this summer. It reckoned that decision had already been made by EU commissioner Neelie Kroes because the EC had gathered enough evidence against the world’s biggest chip maker following a lawsuit by rival AMD.
The paper also claimed that the EU would prevent Intel from marketing its microprocessors to PC vendors at a marked down price, and that the authority would forbid the company from subsidising retailer's advertising costs.
The EU has been sniffing around Intel's business practices in Europe for six years, but the investigation got a second wind in 2005 when AMD formally accused its arch-rival of using illegal tactics to dominate the PC and server markets.
In July last year the EU filed formal anti-trust charges against Intel.
Intel spokesman Chuck Mulloy told Reuters yesterday: "We continue to work with the Commission to attempt to convince them that our business practices are well within the law." ®
IBM Unveils New x86 Servers With Quad-Core AMD Opteron(TM) Processors
IBM (NYSE: IBM) today announced three IBM System x(TM) servers featuring the Quad-Core AMD Opteron(TM) processor.
All of the new servers provide the IT efficiency needed in a new enterprise data center, with the System x3755
delivering 15 percent better performance than its predecessor.
IBM is delivering new Quad-Core AMD Opteron processor-based systems that offer exceptional power-efficiency,
performance and virtualization capabilities for today's most demanding datacenters. These include:
-- The System x3455 offers leading memory performance with 48 GB of
memory, making it ideal for technical and financial applications. With
today's announcement of the addition of Quad-Core processors, the x3455 is
now more competitive for high performance computing.
-- The System x3655 is now more effective for business performance
computing and memory-intensive applications such as database,
virtualization and dynamic web serving.
-- The new System x3755 leverages a unique modular design to scale to
four sockets more affordably. Its flexible design allows customers to
increase the size of their system as their databases, virtualization
environments and collaboration applications grow. The x3755 provides a
three socket configuration that outperforms many competitive four socket
configurations.
The new servers are an important element in helping clients develop a new enterprise data center, which offers
dramatic improvements in IT efficiency and provides for rapid deployment of new IT services to support future
business growth. IBM is helping clients move to new enterprise data centers by focusing on best practices around
virtualization, green IT, service management and cloud computing.
"IBM continues to deliver innovation and choice in the x86 market with today's introduction of System x servers
based on AMD's new Quad-Core processors," said James Northington, vice president, IBM System x. "The new System
x3755 allows clients to grow the system along with their business, affordably scaling from the standard two socket
system to three and four socket configurations while delivering industry leading price and performance."
"The addition of Quad-Core AMD Opteron processor-based servers will help continue IBM's tradition of enhancing its
customers' competitiveness through innovative, high-performing computing solutions," said Randy Allen, Senior Vice
President of CSG, AMD. "Datacenter managers are increasingly seeking a balance of performance, energy-efficiency,
and advanced virtualization functionality in order to optimize server resources amidst skyrocketing power, cooling
and space costs. The Quad-Core AMD Opteron processor is at the forefront of addressing this new paradigm of
datacenter performance."
Pricing and Availability
The IBM System x3455, x3655 and x3755 will be available with Quad-Core AMD Opteron(TM) processors in June, starting
at $1,809, $2,499 and $3,407 respectively.
AMD, the AMD Arrow logo, AMD Opteron and combinations thereof, are trademarks of Advanced Micro Devices, Inc. Other
names are for informational purposes only and may be trademarks of their respective owners.
All of the new servers provide the IT efficiency needed in a new enterprise data center, with the System x3755
delivering 15 percent better performance than its predecessor.
IBM is delivering new Quad-Core AMD Opteron processor-based systems that offer exceptional power-efficiency,
performance and virtualization capabilities for today's most demanding datacenters. These include:
-- The System x3455 offers leading memory performance with 48 GB of
memory, making it ideal for technical and financial applications. With
today's announcement of the addition of Quad-Core processors, the x3455 is
now more competitive for high performance computing.
-- The System x3655 is now more effective for business performance
computing and memory-intensive applications such as database,
virtualization and dynamic web serving.
-- The new System x3755 leverages a unique modular design to scale to
four sockets more affordably. Its flexible design allows customers to
increase the size of their system as their databases, virtualization
environments and collaboration applications grow. The x3755 provides a
three socket configuration that outperforms many competitive four socket
configurations.
The new servers are an important element in helping clients develop a new enterprise data center, which offers
dramatic improvements in IT efficiency and provides for rapid deployment of new IT services to support future
business growth. IBM is helping clients move to new enterprise data centers by focusing on best practices around
virtualization, green IT, service management and cloud computing.
"IBM continues to deliver innovation and choice in the x86 market with today's introduction of System x servers
based on AMD's new Quad-Core processors," said James Northington, vice president, IBM System x. "The new System
x3755 allows clients to grow the system along with their business, affordably scaling from the standard two socket
system to three and four socket configurations while delivering industry leading price and performance."
"The addition of Quad-Core AMD Opteron processor-based servers will help continue IBM's tradition of enhancing its
customers' competitiveness through innovative, high-performing computing solutions," said Randy Allen, Senior Vice
President of CSG, AMD. "Datacenter managers are increasingly seeking a balance of performance, energy-efficiency,
and advanced virtualization functionality in order to optimize server resources amidst skyrocketing power, cooling
and space costs. The Quad-Core AMD Opteron processor is at the forefront of addressing this new paradigm of
datacenter performance."
Pricing and Availability
The IBM System x3455, x3655 and x3755 will be available with Quad-Core AMD Opteron(TM) processors in June, starting
at $1,809, $2,499 and $3,407 respectively.
AMD, the AMD Arrow logo, AMD Opteron and combinations thereof, are trademarks of Advanced Micro Devices, Inc. Other
names are for informational purposes only and may be trademarks of their respective owners.
Intel Confirms Chip Component Delay
Intel(INTC - Cramer's Take - Stockpickr) confirmed that it is delaying the release of certain components at the heart of its forthcoming notebook PC package of chips.
The Santa Clara, Calif., chipmaker said the delay affects the integrated graphics processing and wireless chips in its so-called Montevina product, and will result in certain parts shipping a few weeks behind schedule.
The move doesn't affect Intel's financial forecast for the current quarter, and the company doesn't anticipate any change to its full-year forecast, said company spokesman Tom Beerman.
Shares of Intel were off 40 cents at $23.20 in recent trading Wednesday.
Earlier this week, analysts reported signs of trouble in Montevina, the code name of the new version of Centrino, the bundle of chips that Intel offers as a package of key electronic ingredients for notebook PCs.
Intel's Beerman confirmed that the company is retesting the graphics processing chips to address certain issues, while a "paperwork error" involving filings with the Federal Communications Commission means that the wireless component of Montevina will be also held up slightly.
Intel had initially promised to ship Montevina in the first half of 2008. The company now says it will ship the notebook microprocessor and other related Montevina chips on July 14. The rest of the components will ship a few weeks later, sometime in the first week of August, said Beerman.
He said that PC makers could fill out Intel's Montevina by incorporating graphics and WiFi components from other suppliers.
Whether or not other chipmakers have components that are compatible with Montevnia, which is a new product, isn't clear. If the chips are not compatible, PC makers may have to go with an entirely different bundle of chips until Intel can offer the complete package, says analyst Jon Peddie of Jon Peddie Research.
And the fact that Intel is missing the component for integrated graphics, which accounts for the bulk of the notebook PC market, could present rival Advanced Micro Devices(AMD - Cramer's Take - Stockpickr) with an opportunity to grab some market share.
"Every company in the industry has a hiccup from time to time. It's just the nature of the business" says Peddie. "The only thing that makes this one annoying and maybe a little scary is the fact that there is this timing for the back-to-school [sales] rollout."
The Santa Clara, Calif., chipmaker said the delay affects the integrated graphics processing and wireless chips in its so-called Montevina product, and will result in certain parts shipping a few weeks behind schedule.
The move doesn't affect Intel's financial forecast for the current quarter, and the company doesn't anticipate any change to its full-year forecast, said company spokesman Tom Beerman.
Shares of Intel were off 40 cents at $23.20 in recent trading Wednesday.
Earlier this week, analysts reported signs of trouble in Montevina, the code name of the new version of Centrino, the bundle of chips that Intel offers as a package of key electronic ingredients for notebook PCs.
Intel's Beerman confirmed that the company is retesting the graphics processing chips to address certain issues, while a "paperwork error" involving filings with the Federal Communications Commission means that the wireless component of Montevina will be also held up slightly.
Intel had initially promised to ship Montevina in the first half of 2008. The company now says it will ship the notebook microprocessor and other related Montevina chips on July 14. The rest of the components will ship a few weeks later, sometime in the first week of August, said Beerman.
He said that PC makers could fill out Intel's Montevina by incorporating graphics and WiFi components from other suppliers.
Whether or not other chipmakers have components that are compatible with Montevnia, which is a new product, isn't clear. If the chips are not compatible, PC makers may have to go with an entirely different bundle of chips until Intel can offer the complete package, says analyst Jon Peddie of Jon Peddie Research.
And the fact that Intel is missing the component for integrated graphics, which accounts for the bulk of the notebook PC market, could present rival Advanced Micro Devices(AMD - Cramer's Take - Stockpickr) with an opportunity to grab some market share.
"Every company in the industry has a hiccup from time to time. It's just the nature of the business" says Peddie. "The only thing that makes this one annoying and maybe a little scary is the fact that there is this timing for the back-to-school [sales] rollout."
Intel's Atom to be out in Q3
KOCHI: Intel’s new processor Atom would be launched in the third quarter (July-September) of 2008. This was announced at the Intel Channel Conference (ICC) held here on Tuesday. Atom is the name given to a new series of low-power processors being launched by Intel, which would find application in affordable computers designed specifically for internet-related use and mobile internet devices.
“Our studies showed that a good number of people used PCs for checking e-mails and other such internet-related use,” Ashok Nair, Intel’s channel business manager for south Asia, said. “For them, many advanced features of a PC are not necessary. At a time when companies are trying to reduce the cost of PCs, these studies prompted the company to work on processors that power PCs for internet use,” he said.
ICC is being held at 15 to 20 cities across the country. It is usually held two times a year. These conferences are followed up by mini-conferences in around 50 cities. The details about the upcoming products are shared with channel partners during the conferences. The company has 2000 channel partners in the country.
The new internet PCs based on the Atom processor could take the form of net-tops, netbooks or mobile internet devices. The new technology could lead to revolutionary designs for PCs based on low power use and higher efficiency. It could also translate into lower costs.
“Our studies showed that a good number of people used PCs for checking e-mails and other such internet-related use,” Ashok Nair, Intel’s channel business manager for south Asia, said. “For them, many advanced features of a PC are not necessary. At a time when companies are trying to reduce the cost of PCs, these studies prompted the company to work on processors that power PCs for internet use,” he said.
ICC is being held at 15 to 20 cities across the country. It is usually held two times a year. These conferences are followed up by mini-conferences in around 50 cities. The details about the upcoming products are shared with channel partners during the conferences. The company has 2000 channel partners in the country.
The new internet PCs based on the Atom processor could take the form of net-tops, netbooks or mobile internet devices. The new technology could lead to revolutionary designs for PCs based on low power use and higher efficiency. It could also translate into lower costs.
Intel launching six-core CPUs this year
While most of us are still stuck with dual-core microchips (if that), Intel is prepping to launch a six-core processor in the second half of 2008. While the company currently offers a few four-core, or quad-core, models of its chips, most currently-sold machines still come with the more affordable dual-core CPUs installed.
The six-core chip is code-named Dunnington and will be built using the same 45nm process that Intel is using for its just-released Penryn-class chips. Another key feature is the chip's large level of level 3 (or L3) cache, 16MB, which is one of the memory buffers between the processor and RAM that are used to speed the overall performance of the chip. Most current Intel chips don't use any L3 cache, instead using only L2 cache: L3 has historically been limited to server-class computer chips, though AMD's Phenom chip has 2MB of L3 onboard. (Bottom line for those who don't understand any of this stuff: This chip ought to be blazing fast, at least on paper.)
The bigger question now becomes what Joe Average might actually do with six cores of processing power available to him. Numerous performance benchmarks have shown little real-world advantage in moving from two cores to four, even in high-end applications. Users who spend most of their time browsing the web and replying to email will find even less of a performance boost. That said, other computer components tend to get faster and more capable alongside new CPUs, so a Dunnington-based computer ought to feel very fast at booting, loading applications, and running complex processes like spell-checking a large document. Running an antivirus scan in the background which bogs down most computers should be no problem on a six-core CPU, too.
Expect Dunnington to arrive first for desktop PCs later this year. Intel hasn't said whether laptops are part of the plan for the technology, but given quad-core's limited introduction on notebook computers, I'd expect this to remain hard to find in a portable format for some time. More details as they emerge.
The six-core chip is code-named Dunnington and will be built using the same 45nm process that Intel is using for its just-released Penryn-class chips. Another key feature is the chip's large level of level 3 (or L3) cache, 16MB, which is one of the memory buffers between the processor and RAM that are used to speed the overall performance of the chip. Most current Intel chips don't use any L3 cache, instead using only L2 cache: L3 has historically been limited to server-class computer chips, though AMD's Phenom chip has 2MB of L3 onboard. (Bottom line for those who don't understand any of this stuff: This chip ought to be blazing fast, at least on paper.)
The bigger question now becomes what Joe Average might actually do with six cores of processing power available to him. Numerous performance benchmarks have shown little real-world advantage in moving from two cores to four, even in high-end applications. Users who spend most of their time browsing the web and replying to email will find even less of a performance boost. That said, other computer components tend to get faster and more capable alongside new CPUs, so a Dunnington-based computer ought to feel very fast at booting, loading applications, and running complex processes like spell-checking a large document. Running an antivirus scan in the background which bogs down most computers should be no problem on a six-core CPU, too.
Expect Dunnington to arrive first for desktop PCs later this year. Intel hasn't said whether laptops are part of the plan for the technology, but given quad-core's limited introduction on notebook computers, I'd expect this to remain hard to find in a portable format for some time. More details as they emerge.
AMD accuses Intel in vendor payoff scheme
There's no love lost between rival chipmakers AMD and Intel, and this week's news isn't going to improve the relationship. According to legal filings and news reports, AMD alleges that Intel has paid major computer manufacturers not to use AMD products. The company filed a 108-page document with the court this week containing more details of the alleged scheme, but many of the points remain redacted. (According to Intel, AMD isn't alleging any new complaints in its latest filing, it's just making a few more of those complaints available to to the public.)
AMD and Intel have been fighting it out in court since 2005, when AMD first accused Intel of being a monopoly under anti-trust law. Intel has long been the market share leader in computer CPUs (actually, it's been the leader ever since PCs were invented), but AMD began to challenge that leadership in 2003-'04 as AMD innovated some then-unprecedented features into its chips, releasing the first 64-bit CPU and the first dual-core chips for consumer-level computers.
AMD's lawsuit alleges that rather than innovate a competitive response, Intel instead leveraged its vast power in the market to have AMD pushed to the sidelines by paying Dell, Gateway, Acer, HP, and others not to use AMD chips. (Of course, Intel did eventually mount a response: When the Core series of chips were released in 2006, Intel regained its place as the producer of the best chips on the market, a spot it's held ever since.)
Though I am not a lawyer, pay-for-placement schemes like this, if the allegations are true, are not immediately and universally illegal (after all, this is a big reason why Blu-ray ultimately beat HD DVD), but it will be up to the courts to decide whether legal lines were crossed. Intel has denied any wrongdoing, saying in its filing that any payments made simply amount to rebates and price cuts.
I'll also note that all of the companies mentioned, to my knowledge, have always offered products with AMD chips, even though they may not make up a large part of the product portfolio. That said, AMD accuses Intel of a whole range of wrongdoing, including predatory pricing and even tampering with benchmarks to make its products look better. Again, Intel denies all such claims in its response.
There clearly won't be a resolution to this issue until it is dealt with by the courts, and that seems some years off, unless a settlement is brokered. (The trial is already three years old: The current kerfuffle is over how many depositions each side can call for, so we're in for a long haul.)
AMD and Intel have been fighting it out in court since 2005, when AMD first accused Intel of being a monopoly under anti-trust law. Intel has long been the market share leader in computer CPUs (actually, it's been the leader ever since PCs were invented), but AMD began to challenge that leadership in 2003-'04 as AMD innovated some then-unprecedented features into its chips, releasing the first 64-bit CPU and the first dual-core chips for consumer-level computers.
AMD's lawsuit alleges that rather than innovate a competitive response, Intel instead leveraged its vast power in the market to have AMD pushed to the sidelines by paying Dell, Gateway, Acer, HP, and others not to use AMD chips. (Of course, Intel did eventually mount a response: When the Core series of chips were released in 2006, Intel regained its place as the producer of the best chips on the market, a spot it's held ever since.)
Though I am not a lawyer, pay-for-placement schemes like this, if the allegations are true, are not immediately and universally illegal (after all, this is a big reason why Blu-ray ultimately beat HD DVD), but it will be up to the courts to decide whether legal lines were crossed. Intel has denied any wrongdoing, saying in its filing that any payments made simply amount to rebates and price cuts.
I'll also note that all of the companies mentioned, to my knowledge, have always offered products with AMD chips, even though they may not make up a large part of the product portfolio. That said, AMD accuses Intel of a whole range of wrongdoing, including predatory pricing and even tampering with benchmarks to make its products look better. Again, Intel denies all such claims in its response.
There clearly won't be a resolution to this issue until it is dealt with by the courts, and that seems some years off, unless a settlement is brokered. (The trial is already three years old: The current kerfuffle is over how many depositions each side can call for, so we're in for a long haul.)
Intel Delays Centrino 2 Launch Over Chipset, Antenna Issues
Intel will delay the launch of its upcoming Centrino 2 platform for laptop computers to resolve issues related to the chipset and the antenna used with the wireless chipset.
Centrino 2, also called Montevina, is the next version of Intel's popular Centrino platform for laptops and was previously slated for a June launch. Computer makers are now expected to begin shipping Centrino 2 systems to retail shops in July, with production reaching high volumes during the first week of August, said Elvin Ong, an Intel spokesman in Singapore.
The Centrino 2 platform includes new Core 2 Duo and Core 2 Extreme Mobile processors, as well as an updated chipset with integrated graphics. On the wireless side, Centrino 2 will offer support for Wi-Fi, and WiMax is being offered as an option.
"We are taking the extra days to address two issues that require us to re-screen our chipsets with integrated graphics, and attend to some terms-and-conditions mistakes while filing and testing our wireless antennas," Ong said, adding that Intel expects Centrino 2 laptops to be available in time for the crucial back-to-school selling season.
Ong declined to detail the chipset issues that require Intel to re-screen these products, but said the move was necessary to insure the quality of the product. "We are not going to ship a sub-par product into the market," he said.
There is no problem with the processors used in Centrino 2, Ong said, calling these chips "healthy."
Centrino 2, also called Montevina, is the next version of Intel's popular Centrino platform for laptops and was previously slated for a June launch. Computer makers are now expected to begin shipping Centrino 2 systems to retail shops in July, with production reaching high volumes during the first week of August, said Elvin Ong, an Intel spokesman in Singapore.
The Centrino 2 platform includes new Core 2 Duo and Core 2 Extreme Mobile processors, as well as an updated chipset with integrated graphics. On the wireless side, Centrino 2 will offer support for Wi-Fi, and WiMax is being offered as an option.
"We are taking the extra days to address two issues that require us to re-screen our chipsets with integrated graphics, and attend to some terms-and-conditions mistakes while filing and testing our wireless antennas," Ong said, adding that Intel expects Centrino 2 laptops to be available in time for the crucial back-to-school selling season.
Ong declined to detail the chipset issues that require Intel to re-screen these products, but said the move was necessary to insure the quality of the product. "We are not going to ship a sub-par product into the market," he said.
There is no problem with the processors used in Centrino 2, Ong said, calling these chips "healthy."
Intel Dual-core Xeon Processors
The Dual-Core Intel Xeon processor 5000 series is a simple purchase decision for a business to make: Dual core performance at single core price. Starting today all dual processor servers should be multi-core, from high-performance to value. Adding to exciting new levels of processor performance come processor features: Intel Virtualization Technology, Intel Hyperthreading Technology, Intel Enhanced Speedstep Technology and Execute Disable Bit.The Dual-Core Intel Xeon processor 5000 series is compatible with Intel 5000P, 5000V and 5000X series of chipsets. Server boards based on these chipsets are optimized for dual core with advanced platform features of Dual Independent Bus and Fully Buffered DIMM memory technology offering up to 4X and 3X available bandwidth of current generation platforms. Improved I/O performance is possible with PCI Express and the option of Intel I/O Acceleration TechnologySpecifically designed for and validated with server boards based on the Intel 5000P, 5000V and 5000X chipsets Available in two thermal solution options.....
Intel® Atom™ Processor smallest chip
Intel’s smallest and lowest power processor², the Intel Atom processor will enable the industry to design new Mobile Internet Devices (MIDs) and affordable Internet-focused notebooks (netbooks), and desktops (nettops). This new processor also serves as the foundation for the all new Intel® Centrino® Atom™ processor technology, a collection of chips that enables amazing Internet experiences in pocketable devices.
Newly designed from the ground up, 45nm Intel Atom processors pack an astounding 47 million transistors on a single chip measuring less than 25mm², making them Intel’s smallest and lowest power processors.¹ All this, and while still delivering the power
* Get a new range of power-efficient devices with excellent performance enabled by all new hafnium-infused 45nm high-k silicon technology
* Increase energy efficiency in smaller more compact designs with a thermal design power specification ranging from subwatt to 2.5 watts for mobile devices
* Extend battery life in select devices with an incredibly low idle power as low as 30 mW allowing the device to stay powered on while also conserving energy
Powering the next generation of Internet-centric devices, Intel Atom processors will power the latest in pocketable, portable devices. Based on an entirely new microarchitecture, the Intel Atom processor was developed specifically for performance and low power while maintaining full Intel® Core™ microarchitecture instruction set compatibility. Some Intel Atom processors will also feature multiple threads for better performance and increased system responsiveness.
Devices powered by Intel Atom processors allow you to stay in touch on-the-go, connect to business and enjoy entertainment, remain connected affordably with a new series of netbooks and nettops, and so much more.
source:-intel.com
Newly designed from the ground up, 45nm Intel Atom processors pack an astounding 47 million transistors on a single chip measuring less than 25mm², making them Intel’s smallest and lowest power processors.¹ All this, and while still delivering the power
* Get a new range of power-efficient devices with excellent performance enabled by all new hafnium-infused 45nm high-k silicon technology
* Increase energy efficiency in smaller more compact designs with a thermal design power specification ranging from subwatt to 2.5 watts for mobile devices
* Extend battery life in select devices with an incredibly low idle power as low as 30 mW allowing the device to stay powered on while also conserving energy
Powering the next generation of Internet-centric devices, Intel Atom processors will power the latest in pocketable, portable devices. Based on an entirely new microarchitecture, the Intel Atom processor was developed specifically for performance and low power while maintaining full Intel® Core™ microarchitecture instruction set compatibility. Some Intel Atom processors will also feature multiple threads for better performance and increased system responsiveness.
Devices powered by Intel Atom processors allow you to stay in touch on-the-go, connect to business and enjoy entertainment, remain connected affordably with a new series of netbooks and nettops, and so much more.
source:-intel.com
Key Architectural Features of AMD Phenom™ X4 Quad-Core Processors
The industry's first true Quad core x86 processor
* True quad-core designed from the ground up for better communication between cores.
o BENEFIT : Cores can communicate on die rather than on package for better performance
AMD64 with Direct Connect Architecture
* Helps improve system performance and efficiency by directly connecting the memory controller and I/O to the CPU.
* Designed to enable simultaneous 32- and 64-bit computing
* Integrated DDR2 Memory Controller
* BENEFITS :
o Increases application performance by reducing memory latency
o Scales memory bandwidth and performance to match compute needs
o HyperTransport™ Technology provides up to 16.0GB/s peak bandwidth per processor—reducing I/O bottlenecks
o Up to 33.1GB/s total delivered processor-to-system bandwidth (HyperTransport bus + memory bus)
AMD Balanced Smart Cache
* Shared L3 cache
* In addition to the 512K L2 cache per core, up to 2MB of L3 cache shared by up to 4 cores.
o BENEFIT : Shortened access times to highly accessed data for better performance.
AMD Wide Floating Point Accelerator
* 128-bit floating point unit (FPU)
* High performance (128bit internal data path) floating point unit per core.
o BENEFIT : Larger data paths for quicker floating point calculations and better performance.
HyperTransport™ technology
* One 16-bit link up to 4000MT/s
* Up to 8.0 GB/s HyperTransport™ I/O bandwidth; Up to 16.0GB/s in HyperTransport Generation 3.0 mode
* Up to 33.1GB/s total delivered processor-to-system bandwidth (HyperTransport bus + memory bus)
o BENEFIT : Quick access times to system resources for better performance.
Integrated DDR2 DRAM Controller with AMD Memory Optimizer Technology
* A high-bandwidth, low-latency integrated DDR2 memory controller
* Supports PC2-8500 (DDR2-1066); PC2-6400 (DDR2-800), PC2-5300 (DDR2-667), PC2-4200 (DDR2-533) or PC2-3200 (DDR2-400) SDRAM unbuffered DIMMs
* Support for 64-bit DDR2 SDRAM memory
* Up to 17.1GB/s memory bandwidth
o BENEFIT : Quick access to system memory for better performance.
AMD Virtualization™ (AMD-V™) With Rapid Virtualization Indexing
* Silicon feature-set enhancements designed to improve the performance, reliability, and security of existing and future virtualization environments by allowing virtualized applications with direct and rapid access to their allocated memory.
o BENEFIT : Helps virtualization software to run more securely and efficiently enabling a better experience when dealing with virtual systems
AMD Cool'n'Quiet™ 2.0 technology
* Enhanced power management features which automatically and instantaneously adjusts performance states and features based on processor performance requirements
* For quieter operation and reduced power requirements
o BENEFIT : Enables platform designs providing less heat and noise efficient performance and energy usage.
AMD CoolCore™ Technology
* Reduces processor energy consumption by turning off unused parts of the processor. For example, the memory controller can turn off the write logic when reading from memory, helping reduce system power.
* Works automatically without the need for drivers or BIOS enablement.
* Power can be switched on or off within a single clock cycle, saving energy without comprimised performance.
o BENEFIT: Helps users get more efficient performance by dynamically activating or turning off parts of the processor.
Dual Dynamic Power Management™
* Enables more granular power management capabilities to reduce processor energy consumption.
* Separate power planes for cores and memory controller, for optimum power consumption and performance, creating more opportunities for power savings within the cores and memory controller.
o BENEFIT Helps improve platform efficiency by providing on demand memory performance while still allowing for decreased system power consumption
Key Architectural Features of AMD Phenom™ X3 Triple-Core Processors
45286-A_ Phenom triple core architectual diagram-1066 FINAL.jpg
The Industry's First True Triple-Core x86 Processor
* True triple-core designed from the ground up for better communication between cores.
o Benefit: Cores can communicate on die rather than on package for better performance.
AMD64 with Direct Connect Architecture
* Helps improve system performance and efficiency by directly connecting the memory controller and I/O to the CPU.
* Designed to enable simultaneous 32- and 64-bit computing
* Integrated DDR2 Memory Controller
* Benefits:
o Increases application performance by reducing memory latency
o Scales memory bandwidth and performance to match compute needs
o HyperTransport™ Technology provides up to 16.0GB/s peak bandwidth per processor — reducing I/O bottlenecks
o Up to 33.1GB/s total delivered processor-to-system bandwidth (HyperTransport bus + memory bus)
AMD Balanced Smart Cache
* Shared L3 cache
* In addition to the 512K L2 cache per core, up to 2MB of L3 cache shared by up to 3 cores.
o Benefit: Shortened access times to highly accessed data for better performance.
AMD Wide Floating Point Accelerator
* 128-bit floating point unit (FPU)
* High performance (128bit internal data path) floating point unit per core.
o Benefit: Larger data paths for quicker floating point calculations and better performance.
HyperTransport™ Technology
* One 16-bit link up to 3600MT/s
* Up to 8 .0 GB/s HyperTransport™ I/O bandwidth; Up to 16.0GB/s in HyperTransport Generation 3.0 mode
* Up to 33.1GB/s total delivered processor-to-system bandwidth (HyperTransport bus + memory bus)
o Benefit: Quick access times to system resources for better performance.
Integrated DDR2 DRAM Controller with AMD Memory Optimizer Technology
* A high-bandwidth, low-latency integrated DDR2 memory controller
* Supports PC2-8500 (DDR2-1066); PC2-6400 (DDR2-800), PC2-5300 (DDR2-667), PC2-4200 (DDR2-533) or PC2-3200 (DDR2-400) SDRAM unbuffered DIMMs
* Support for 64-bit DDR2 SDRAM memory
* Up to 17.1GB/s memory bandwidth
o Benefit: Quick access to system memory for better performance.
AMD Virtualization™ (AMD-V™) With Rapid Virtualization Indexing
* Silicon feature-set enhancements designed to improve the performance, reliability, and security of existing and future virtualization environments by allowing virtualized applications with direct and rapid access to their allocated memory.
o Benefit: Helps virtualization software to run more securely and efficiently enabling a better experience when dealing with virtual systems.
AMD Cool'n'Quiet™ 2.0 Technology
* Enhanced power management features which automatically and instantaneously adjusts performance states and features based on processor performance requirements
* For quieter operation and reduced power requirements
o Benefit: Enables platform designs providing less heat and noise efficient performance and energy usage.
AMD CoolCore™ Technology
* Reduces processor energy consumption by turning off unused parts of the processor. For example, the memory controller can turn off the write logic when reading from memory, helping reduce system power.
* Works automatically without the need for drivers or BIOS enablement.
* Power can be switched on or off within a single clock cycle, saving energy without compromised performance.
o Benefit: Helps users get more efficient performance by dynamically activating or turning off parts of the processor.
Dual Dynamic Power Management™
* Enables more granular power management capabilities to reduce processor energy consumption.
* Separate power planes for cores and memory controller, for optimum power consumption and performance, creating more opportunities for power savings within the cores and memory controller.
o Benefit: Helps improve platform efficiency by providing on demand memory performance while still allowing for decreased system power consumption.
* True quad-core designed from the ground up for better communication between cores.
o BENEFIT : Cores can communicate on die rather than on package for better performance
AMD64 with Direct Connect Architecture
* Helps improve system performance and efficiency by directly connecting the memory controller and I/O to the CPU.
* Designed to enable simultaneous 32- and 64-bit computing
* Integrated DDR2 Memory Controller
* BENEFITS :
o Increases application performance by reducing memory latency
o Scales memory bandwidth and performance to match compute needs
o HyperTransport™ Technology provides up to 16.0GB/s peak bandwidth per processor—reducing I/O bottlenecks
o Up to 33.1GB/s total delivered processor-to-system bandwidth (HyperTransport bus + memory bus)
AMD Balanced Smart Cache
* Shared L3 cache
* In addition to the 512K L2 cache per core, up to 2MB of L3 cache shared by up to 4 cores.
o BENEFIT : Shortened access times to highly accessed data for better performance.
AMD Wide Floating Point Accelerator
* 128-bit floating point unit (FPU)
* High performance (128bit internal data path) floating point unit per core.
o BENEFIT : Larger data paths for quicker floating point calculations and better performance.
HyperTransport™ technology
* One 16-bit link up to 4000MT/s
* Up to 8.0 GB/s HyperTransport™ I/O bandwidth; Up to 16.0GB/s in HyperTransport Generation 3.0 mode
* Up to 33.1GB/s total delivered processor-to-system bandwidth (HyperTransport bus + memory bus)
o BENEFIT : Quick access times to system resources for better performance.
Integrated DDR2 DRAM Controller with AMD Memory Optimizer Technology
* A high-bandwidth, low-latency integrated DDR2 memory controller
* Supports PC2-8500 (DDR2-1066); PC2-6400 (DDR2-800), PC2-5300 (DDR2-667), PC2-4200 (DDR2-533) or PC2-3200 (DDR2-400) SDRAM unbuffered DIMMs
* Support for 64-bit DDR2 SDRAM memory
* Up to 17.1GB/s memory bandwidth
o BENEFIT : Quick access to system memory for better performance.
AMD Virtualization™ (AMD-V™) With Rapid Virtualization Indexing
* Silicon feature-set enhancements designed to improve the performance, reliability, and security of existing and future virtualization environments by allowing virtualized applications with direct and rapid access to their allocated memory.
o BENEFIT : Helps virtualization software to run more securely and efficiently enabling a better experience when dealing with virtual systems
AMD Cool'n'Quiet™ 2.0 technology
* Enhanced power management features which automatically and instantaneously adjusts performance states and features based on processor performance requirements
* For quieter operation and reduced power requirements
o BENEFIT : Enables platform designs providing less heat and noise efficient performance and energy usage.
AMD CoolCore™ Technology
* Reduces processor energy consumption by turning off unused parts of the processor. For example, the memory controller can turn off the write logic when reading from memory, helping reduce system power.
* Works automatically without the need for drivers or BIOS enablement.
* Power can be switched on or off within a single clock cycle, saving energy without comprimised performance.
o BENEFIT: Helps users get more efficient performance by dynamically activating or turning off parts of the processor.
Dual Dynamic Power Management™
* Enables more granular power management capabilities to reduce processor energy consumption.
* Separate power planes for cores and memory controller, for optimum power consumption and performance, creating more opportunities for power savings within the cores and memory controller.
o BENEFIT Helps improve platform efficiency by providing on demand memory performance while still allowing for decreased system power consumption
Key Architectural Features of AMD Phenom™ X3 Triple-Core Processors
45286-A_ Phenom triple core architectual diagram-1066 FINAL.jpg
The Industry's First True Triple-Core x86 Processor
* True triple-core designed from the ground up for better communication between cores.
o Benefit: Cores can communicate on die rather than on package for better performance.
AMD64 with Direct Connect Architecture
* Helps improve system performance and efficiency by directly connecting the memory controller and I/O to the CPU.
* Designed to enable simultaneous 32- and 64-bit computing
* Integrated DDR2 Memory Controller
* Benefits:
o Increases application performance by reducing memory latency
o Scales memory bandwidth and performance to match compute needs
o HyperTransport™ Technology provides up to 16.0GB/s peak bandwidth per processor — reducing I/O bottlenecks
o Up to 33.1GB/s total delivered processor-to-system bandwidth (HyperTransport bus + memory bus)
AMD Balanced Smart Cache
* Shared L3 cache
* In addition to the 512K L2 cache per core, up to 2MB of L3 cache shared by up to 3 cores.
o Benefit: Shortened access times to highly accessed data for better performance.
AMD Wide Floating Point Accelerator
* 128-bit floating point unit (FPU)
* High performance (128bit internal data path) floating point unit per core.
o Benefit: Larger data paths for quicker floating point calculations and better performance.
HyperTransport™ Technology
* One 16-bit link up to 3600MT/s
* Up to 8 .0 GB/s HyperTransport™ I/O bandwidth; Up to 16.0GB/s in HyperTransport Generation 3.0 mode
* Up to 33.1GB/s total delivered processor-to-system bandwidth (HyperTransport bus + memory bus)
o Benefit: Quick access times to system resources for better performance.
Integrated DDR2 DRAM Controller with AMD Memory Optimizer Technology
* A high-bandwidth, low-latency integrated DDR2 memory controller
* Supports PC2-8500 (DDR2-1066); PC2-6400 (DDR2-800), PC2-5300 (DDR2-667), PC2-4200 (DDR2-533) or PC2-3200 (DDR2-400) SDRAM unbuffered DIMMs
* Support for 64-bit DDR2 SDRAM memory
* Up to 17.1GB/s memory bandwidth
o Benefit: Quick access to system memory for better performance.
AMD Virtualization™ (AMD-V™) With Rapid Virtualization Indexing
* Silicon feature-set enhancements designed to improve the performance, reliability, and security of existing and future virtualization environments by allowing virtualized applications with direct and rapid access to their allocated memory.
o Benefit: Helps virtualization software to run more securely and efficiently enabling a better experience when dealing with virtual systems.
AMD Cool'n'Quiet™ 2.0 Technology
* Enhanced power management features which automatically and instantaneously adjusts performance states and features based on processor performance requirements
* For quieter operation and reduced power requirements
o Benefit: Enables platform designs providing less heat and noise efficient performance and energy usage.
AMD CoolCore™ Technology
* Reduces processor energy consumption by turning off unused parts of the processor. For example, the memory controller can turn off the write logic when reading from memory, helping reduce system power.
* Works automatically without the need for drivers or BIOS enablement.
* Power can be switched on or off within a single clock cycle, saving energy without compromised performance.
o Benefit: Helps users get more efficient performance by dynamically activating or turning off parts of the processor.
Dual Dynamic Power Management™
* Enables more granular power management capabilities to reduce processor energy consumption.
* Separate power planes for cores and memory controller, for optimum power consumption and performance, creating more opportunities for power savings within the cores and memory controller.
o Benefit: Helps improve platform efficiency by providing on demand memory performance while still allowing for decreased system power consumption.
AMD Geode GX S Best processors
The AMD Geode™ GX Single Board Computer (SBC) Reference Design Kit (RDK) is a compact, low-power, high-performance tool developed to enable the next generation of single-board computing and network appliances. Virtually manufacture-ready, the RDK is designed to deliver low total cost of ownership and provide a complete system design package that gives designers flexibility, versatility, and enhanced capabilities in a small-sized package (approximately 5.5" x 5"). The RDK is driven by the AMD Geode GX 466@0.9W processor and complemented by the CS5536 companion chip which provides general I/O support, including USB 2.0, audio, and ATA 100. The RDK is supported by General Software's industry leading Embedded BIOS® with StrongFrame™ Technology.Embedded BIOS® with StrongFrame™ Technology, provides maximum configurability with high-level firmware tools and full source code for the core BIOS. With modular architecture, custom adaptations of the BIOS can be made with ease, speed, and power.Embedded BIOS® with StrongFrame™ Technology FEATURES:• Over 850 configuration options• Advanced firmware applications, available as options, provide boot security, high availability and platform updates• USB: USB 2.0, Legacy keyboard and mouse, USB boot (optional
AMD Duron 650 MHz
AMD presents Duron processor, once again setting the standard for business and home users in the value PC space. Optimized for the value-conscious user, the AMD Duron processor reinforces AMD's leadership position in delivering better computing solutions to the value PC space.??The AMD Duron processor offers value conscious buyers access to technology and performance that stands out among other processors in its class. Employing an innovative design, the AMD Duron processor features sophisticated cache architecture, a high-speed front-side bus, and a super-scalar floating-point unit with enhanced 3DNow! technology. Designed to help prolong the life of a buyer's investment, the AMD Duron processor provides the capability and flexibility to meet the computing needs of value conscious users today and tomorrow. AMD Duron processor-based systems are ideal for applications typically employed by such users, including surfing the Internet, business and personal productivity suites.
IBM Processor Amd Opteron Model
This AMD Opteron processor option provides a highly scalable architecture that delivers next-generation performance as well as a flexible upgrade path from 32- to 64-bit computing. With a single architecture designed to meet current and future business needs, the AMD Opteron processor can help to minimize the integration complexities presented by current business environments and answer future business growth needs. This evolutionary processor provides a dramatic leap forward in compatibility, performance and investment protection and can help significantly reduce the total cost of ownership (TCO) variable. The AMD Opteron processor is offered in three models, the 244, 246 and the 248, ranging in processor speeds and is supported by 1MB of L2 cache on all modelsSupports large memory addressability for data-intensive applications, allowing for optimal performance and improved productivity in your work environments. Flexibility Minimizes integration complexities by providing simultaneous high performance 32- and 64-bit computing Provides the flexibility and scalability for demanding enterprise-class applications. Provides flexibility with a scalable design, delivering the headroom needed for future applications. Investment Protection Provides lower cost of ownership, helping with long-term IT investment protection Provides an investment with a long-span potential thanks to the performance upgrade capabilities inherent in the AMD Opteron processor Allows you to better manage the transition cost and timing of your application suite's 64-bit migration .
AMD talks Shanghai performance, features, roadmap to 2010
Randy Allen, corporate VP and general manager for AMD's Server/Workstation division presented AMD's server roadmap for the rest of 2008 today, hosted a conference call today on what the company is planning for 2009 and 2010. Barcelon's nasty stumble in the fall of 2007 raised questions about AMD's ability to execute effectively in the server space, so the company spent the first part of today's presentation talking up recent Barcelona design wins and reassuring journalists that the processor is shipping in volume now that the TLB erratum has been put to rest.
Unsurprisingly, AMD's upcoming 45nm Shanghai processor was the unofficial star of the day. Barcelona, despite AMD's best efforts, simply isn't capable of scaling to the frequencies required to aggressively compete against Yorkfield-based Xeon processors, to say nothing of Intel's upcoming Nehalem architecture. To date, AMD's guidance on Shanghai's ramp has been strong; the company reiterated its plan to have its 45nm process technology in volume production by the end of 2008, with Shanghai processors presumably shipping very late in '08 or early in 2009.
According to AMD, Shanghai will consume approximately 20 percent less power at idle than Barcelona, and will offer a full 6MB of L3 cache, as opposed to Barcelona's 2MB of L3. AMD claims that Shanghai has approximately twice as much cache as Barcelona overall, which implies that the new processor retains the same 128K L1/512K L2 that Barcelona currently features. Other features include AMD-V (meant to improve virtualization performance) and support for DDR2-800 memory. Allen characterized Shanghai as a "tweak" of Barcelona's architecture, but predicts that we should see overall performance increase by 20 percent when comparing the two product families. He did not, however, break down how much of this performance might be delivered by higher processor efficiency as opposed to clockrate boosts.
If all goes according to plan, Shanghai will be superceded by Istanbul in the second half of 2009. Istanbul is a six-core processor built on 45nm technology whose feature set and cache loadout should otherwise mirror Shanghai's. Like Shanghai, Istanbul will also be a monolithic core design, though it may be the company's last monolithic architecture in the top end of the server space. In 2010, AMD will introduce Magny-Cours, a twelve-core design (duodecim-core if you prefer the Latin) with support for DDR3 and what AMD is calling a "probe filter," which will reduce the amount of cross-processor cache snooping necessary to maintain cache coherency. 2010 will also see the introduction of Sao Paulo, a new six-core design with the same features as Magny-Cours, and the launch of the new "Maranello" platform, aka Socket G34.
When it comes to performance in the here-and-now, AMD isn't saying much. While the company did feature some Barcelona vs. Xeon comparisons, most of the information was presented as a performance-per-watt comparison rather than in terms of absolute (or even percentile) performance. There's nothing wrong with performance-per-watt—it is, in fact, increasingly important—but it's also something of a refuge for companies and products that don't compare well "by the numbers." It doesn't help that AMD claims to have tested against Intel's fastest 80W part, but included only a Xeon 5440 (2.83GHz, 1333MHz FSB, 12MB L3) instead of a Xeon 5450 (3GHz, 1333MHz FSB, 12MB L3).
AMD's new information on Shanghai leaves me feeling cautiously optimistic, but there are still a lot of empty spots in the company's roadmap.. Sunnyvale refused to address any questions related to its "asset smart" strategy, beyond saying that asset smart, whatever it means, will have no impact on the roadmap as published today. Allen also dodged questions on when AMD will begin its conversion to 32nm process technology; the roadmap we saw today does not mention 32nm, despite stretching through 2010.
The pressure on Shanghai at this point is immense; all of the CPUs on AMD's server/workstation roadmap for the next 2.5 years are Shanghai derivatives. AMD's current financial situation and market share have left the company with no room for error; its next-generation CPU will need to shine from the get-go.
Unsurprisingly, AMD's upcoming 45nm Shanghai processor was the unofficial star of the day. Barcelona, despite AMD's best efforts, simply isn't capable of scaling to the frequencies required to aggressively compete against Yorkfield-based Xeon processors, to say nothing of Intel's upcoming Nehalem architecture. To date, AMD's guidance on Shanghai's ramp has been strong; the company reiterated its plan to have its 45nm process technology in volume production by the end of 2008, with Shanghai processors presumably shipping very late in '08 or early in 2009.
According to AMD, Shanghai will consume approximately 20 percent less power at idle than Barcelona, and will offer a full 6MB of L3 cache, as opposed to Barcelona's 2MB of L3. AMD claims that Shanghai has approximately twice as much cache as Barcelona overall, which implies that the new processor retains the same 128K L1/512K L2 that Barcelona currently features. Other features include AMD-V (meant to improve virtualization performance) and support for DDR2-800 memory. Allen characterized Shanghai as a "tweak" of Barcelona's architecture, but predicts that we should see overall performance increase by 20 percent when comparing the two product families. He did not, however, break down how much of this performance might be delivered by higher processor efficiency as opposed to clockrate boosts.
If all goes according to plan, Shanghai will be superceded by Istanbul in the second half of 2009. Istanbul is a six-core processor built on 45nm technology whose feature set and cache loadout should otherwise mirror Shanghai's. Like Shanghai, Istanbul will also be a monolithic core design, though it may be the company's last monolithic architecture in the top end of the server space. In 2010, AMD will introduce Magny-Cours, a twelve-core design (duodecim-core if you prefer the Latin) with support for DDR3 and what AMD is calling a "probe filter," which will reduce the amount of cross-processor cache snooping necessary to maintain cache coherency. 2010 will also see the introduction of Sao Paulo, a new six-core design with the same features as Magny-Cours, and the launch of the new "Maranello" platform, aka Socket G34.
When it comes to performance in the here-and-now, AMD isn't saying much. While the company did feature some Barcelona vs. Xeon comparisons, most of the information was presented as a performance-per-watt comparison rather than in terms of absolute (or even percentile) performance. There's nothing wrong with performance-per-watt—it is, in fact, increasingly important—but it's also something of a refuge for companies and products that don't compare well "by the numbers." It doesn't help that AMD claims to have tested against Intel's fastest 80W part, but included only a Xeon 5440 (2.83GHz, 1333MHz FSB, 12MB L3) instead of a Xeon 5450 (3GHz, 1333MHz FSB, 12MB L3).
AMD's new information on Shanghai leaves me feeling cautiously optimistic, but there are still a lot of empty spots in the company's roadmap.. Sunnyvale refused to address any questions related to its "asset smart" strategy, beyond saying that asset smart, whatever it means, will have no impact on the roadmap as published today. Allen also dodged questions on when AMD will begin its conversion to 32nm process technology; the roadmap we saw today does not mention 32nm, despite stretching through 2010.
The pressure on Shanghai at this point is immense; all of the CPUs on AMD's server/workstation roadmap for the next 2.5 years are Shanghai derivatives. AMD's current financial situation and market share have left the company with no room for error; its next-generation CPU will need to shine from the get-go.
AMD to spin-off fab operations?
SAN JOSE, Calif. -- Shares in Advanced Micro Devices Inc. (AMD) have climbed amid reports that the company will split itself into two parts, including the spin-off of its fab operations.
AMD (Sunnyvale, Calif.) also filed a new antitrust suit against rival Intel Corp., according to reports. The filing is part of the company's 2005 antitrust suit against Intel, the reports said.
Meanwhile, for some time, AMD has talked about a new fab-lite strategy, but it has failed to provide any details. But according to reports from Reuters, AMD could spin-off its manufacturing arm. AMD, in turn, would keep its design group.
"What's been bandied about is the asset-light part which could be the partial separation of AMD's design and manufacturing arms," said CRT Capital Group analyst Ashok Kumar, told Reuters. "On the manufacturing side, they would likely partner with Chartered, but maintain a 51 percent ownership of any such venture."
AMD (Sunnyvale, Calif.) also filed a new antitrust suit against rival Intel Corp., according to reports. The filing is part of the company's 2005 antitrust suit against Intel, the reports said.
Meanwhile, for some time, AMD has talked about a new fab-lite strategy, but it has failed to provide any details. But according to reports from Reuters, AMD could spin-off its manufacturing arm. AMD, in turn, would keep its design group.
"What's been bandied about is the asset-light part which could be the partial separation of AMD's design and manufacturing arms," said CRT Capital Group analyst Ashok Kumar, told Reuters. "On the manufacturing side, they would likely partner with Chartered, but maintain a 51 percent ownership of any such venture."
AMD says Intel paid PC makers to boycott rival
Advanced Micro Devices has expanded the allegations in its antitrust suit against Intel, saying its larger rival paid computer makers not to do business with AMD.
In court papers filed May 1 in U.S. District Court in Delaware, AMD said Intel paid off Acer, Dell, Gateway, major Japanese manufacturers, and system builders and distributors "to close their doors to AMD." Where it couldn't buy exclusivity from computer makers, Intel focused its payments on keeping computers based on AMD microprocessors away from large business customers, AMD said.
Most of the details in the 108-page document, unsealed Monday, have been blacked out, due to a protective order restricting the disclosure of trade secrets and other confidential dealings involving the parties in the suit. AMD sued Intel in June 2005, accusing the chipmaker of illegally using its monopoly to shut AMD out of the market.
The latest filing said Intel paid customers to abandon development of computer models that were essential to the success of a new line of AMD processors. The tactic also was used "to nip in the bud AMD inroads into sectors Intel viewed as critical."
In addition, Intel used discount schemes that made it uneconomical for AMD to compete for a customer's business, and deployed various strong-arm tactics against partners that refused to go along. Those retaliatory tactics included withdrawing marketing funds, withholding critical technical and road map information, allocating scarce products away from disloyal companies, and generally scaling back the level of customer support, the filing said.
In its own brief, Intel denied any anticompetitive practices. The company said that AMD was seeking court protection from discounting and other legal practices within a competitive environment.
"AMD's complaint about Intel's discounting boils down to a complaint that Intel is a more efficient competitor," Intel said.
For a few years before 2006, AMD managed to take market share from Intel, particularly in the market for processors powering server computers. About two years ago, Intel embarked on a reorganization that saw the chipmaker get several important products out before its smaller rival, including a quad-core processor and next-generation 45-nanometer chips that deliver higher power-to-performance ratios.
In court papers filed May 1 in U.S. District Court in Delaware, AMD said Intel paid off Acer, Dell, Gateway, major Japanese manufacturers, and system builders and distributors "to close their doors to AMD." Where it couldn't buy exclusivity from computer makers, Intel focused its payments on keeping computers based on AMD microprocessors away from large business customers, AMD said.
Most of the details in the 108-page document, unsealed Monday, have been blacked out, due to a protective order restricting the disclosure of trade secrets and other confidential dealings involving the parties in the suit. AMD sued Intel in June 2005, accusing the chipmaker of illegally using its monopoly to shut AMD out of the market.
The latest filing said Intel paid customers to abandon development of computer models that were essential to the success of a new line of AMD processors. The tactic also was used "to nip in the bud AMD inroads into sectors Intel viewed as critical."
In addition, Intel used discount schemes that made it uneconomical for AMD to compete for a customer's business, and deployed various strong-arm tactics against partners that refused to go along. Those retaliatory tactics included withdrawing marketing funds, withholding critical technical and road map information, allocating scarce products away from disloyal companies, and generally scaling back the level of customer support, the filing said.
In its own brief, Intel denied any anticompetitive practices. The company said that AMD was seeking court protection from discounting and other legal practices within a competitive environment.
"AMD's complaint about Intel's discounting boils down to a complaint that Intel is a more efficient competitor," Intel said.
For a few years before 2006, AMD managed to take market share from Intel, particularly in the market for processors powering server computers. About two years ago, Intel embarked on a reorganization that saw the chipmaker get several important products out before its smaller rival, including a quad-core processor and next-generation 45-nanometer chips that deliver higher power-to-performance ratios.
Analysis: Is asset-lite a smart move for AMD?--2
R&D expenses
It's not clear though that AMD's research and development expenses would be impacted by the move towards an asset-light manufacturing strategy. In its bid to catch up with Intel, AMD jacked up its R&D budget in 2007 to $1.9 billion from $1.2 billion in 2006.
While this amount is dwarfed by Intel's $5.8 billion R&D budget for 2007, last year's figure represented an outsized 31 percent of AMD's annual revenue compared with 15 percent for Intel.
While outsourced manufacturing might be a viable option for AMD, it's not that clear the company would greatly improve its competitiveness by outsourcing R&D, a key driver of differentiation and market share gain in the semiconductor market.
AMD is locked into a microprocessor manufacturing process development agreement with IBM Corp. focused on the development of leading-edge 45-nanometer, 32-nanometer and 22-nanometer technologies. The agreement is expected to continue for several additional years and AMD isn't likely to scuttle the engagement with IBM even if it spins off its manufacturing operation.
Likely partners
AMD already has outsourced manufacturing agreements with several wafer foundries including Chartered Semiconductor Manufacturing Co. Ltd. for microprocessors and on the graphics IC and chipset products side with Taiwan Semiconductor Manufacturing Co. Ltd., United Microelectronics Corp. and Chartered.
Chartered is the most 8likely candidate for a manufacturing joint venture company if AMD were to move in that direction, according to Ashok Kumar, an analyst with CRT Capital Group speaking with Thomson Reuters News. Kumar believes AMD would still maintain a controlling interest in the company.
AMD currently operates five manufacturing facilities, two of them microprocessor wafer fabrication plants located in Dresden, Germany. The company also runs three microprocessor assembly and test facilities in Malaysia, Singapore and China that it could pool into a new company with a joint venture partner.
Other production agreements and operations the company might also have to review as it prepares for the asset-smart strategy are test, assembly and packaging as well as board-level services it currently outsources to contract manufacturing partners, including Amkor, Advanced Semiconductor Engineering Group, King Yuan Electronics, Silicon Precision Industries, Stats-Chippac,Celestica and Foxconn
It's not clear though that AMD's research and development expenses would be impacted by the move towards an asset-light manufacturing strategy. In its bid to catch up with Intel, AMD jacked up its R&D budget in 2007 to $1.9 billion from $1.2 billion in 2006.
While this amount is dwarfed by Intel's $5.8 billion R&D budget for 2007, last year's figure represented an outsized 31 percent of AMD's annual revenue compared with 15 percent for Intel.
While outsourced manufacturing might be a viable option for AMD, it's not that clear the company would greatly improve its competitiveness by outsourcing R&D, a key driver of differentiation and market share gain in the semiconductor market.
AMD is locked into a microprocessor manufacturing process development agreement with IBM Corp. focused on the development of leading-edge 45-nanometer, 32-nanometer and 22-nanometer technologies. The agreement is expected to continue for several additional years and AMD isn't likely to scuttle the engagement with IBM even if it spins off its manufacturing operation.
Likely partners
AMD already has outsourced manufacturing agreements with several wafer foundries including Chartered Semiconductor Manufacturing Co. Ltd. for microprocessors and on the graphics IC and chipset products side with Taiwan Semiconductor Manufacturing Co. Ltd., United Microelectronics Corp. and Chartered.
Chartered is the most 8likely candidate for a manufacturing joint venture company if AMD were to move in that direction, according to Ashok Kumar, an analyst with CRT Capital Group speaking with Thomson Reuters News. Kumar believes AMD would still maintain a controlling interest in the company.
AMD currently operates five manufacturing facilities, two of them microprocessor wafer fabrication plants located in Dresden, Germany. The company also runs three microprocessor assembly and test facilities in Malaysia, Singapore and China that it could pool into a new company with a joint venture partner.
Other production agreements and operations the company might also have to review as it prepares for the asset-smart strategy are test, assembly and packaging as well as board-level services it currently outsources to contract manufacturing partners, including Amkor, Advanced Semiconductor Engineering Group, King Yuan Electronics, Silicon Precision Industries, Stats-Chippac,Celestica and Foxconn
Analysis: Is asset-lite a smart move for AMD?
If Advanced Micro Devices Inc. were to spin off or combine its manufacturing operations into a standalone company with any of its current production partners as is being speculated (see: AMD to spin-off fab operations?), the microprocessor supplier could dramatically alter its cost structure and reduce this enough to improve its competitive stance against rival Intel Corp.
While industry observers believe the asset-smart strategy could overall be positive for AMD, it could also distract management attention as it executes what would amount to a delicate transition at a time the company is struggling to vend off intense competition from Intel in the microprocessor area and Nvidia Corp. in the graphics IC sector.
The Sunnyvale, Calif., company isn't disclosing details of the asset-smart manufacturing strategy it first announced more than one year ago but industry sources expect an announcement soon from AMD, possibly ahead of or during its annual shareholders' meeting on Thursday (May 8.)
AMD's depressed shares initially surged in after-hour trading on Tuesday and rose more than one percent in early trading Wednesday to $7.20 from $7.12 before retreating again following unconfirmed reports the company plans to merge its manufacturing operations with a foundry partner and spin off the business as a separate company.
The remaining business would continue to design, market and sell the company's microprocessor and graphics IC products, according to analysts. Such a move would help to sharply reduce the company's operating costs, giving it the much needed financial relief in its bruising fight against Intel although AMD would still face tremendous obstacles keeping up with its archrival in rolling out advanced technology products.
One area where AMD would see immediate cost improvements if it pursues a spinoff of its manufacturing operation is in its selling, general and 8administrative costs, which in 2007 and in the first quarter represented approximately 23 percent of total revenue, compared with 14 percent for Intel.
The added savings could be poured into stabilizing the company's shaky finances by paying down long-term debts of approximately $5 billion.
While industry observers believe the asset-smart strategy could overall be positive for AMD, it could also distract management attention as it executes what would amount to a delicate transition at a time the company is struggling to vend off intense competition from Intel in the microprocessor area and Nvidia Corp. in the graphics IC sector.
The Sunnyvale, Calif., company isn't disclosing details of the asset-smart manufacturing strategy it first announced more than one year ago but industry sources expect an announcement soon from AMD, possibly ahead of or during its annual shareholders' meeting on Thursday (May 8.)
AMD's depressed shares initially surged in after-hour trading on Tuesday and rose more than one percent in early trading Wednesday to $7.20 from $7.12 before retreating again following unconfirmed reports the company plans to merge its manufacturing operations with a foundry partner and spin off the business as a separate company.
The remaining business would continue to design, market and sell the company's microprocessor and graphics IC products, according to analysts. Such a move would help to sharply reduce the company's operating costs, giving it the much needed financial relief in its bruising fight against Intel although AMD would still face tremendous obstacles keeping up with its archrival in rolling out advanced technology products.
One area where AMD would see immediate cost improvements if it pursues a spinoff of its manufacturing operation is in its selling, general and 8administrative costs, which in 2007 and in the first quarter represented approximately 23 percent of total revenue, compared with 14 percent for Intel.
The added savings could be poured into stabilizing the company's shaky finances by paying down long-term debts of approximately $5 billion.
AMD discloses 12-core server chip
SAN JOSE, Calif. — Advanced Micro Devices disclosed a conservative road map for 45nm server processors using six and twelve cores through 2010. Lifting a page from the playbook of archrival Intel, the 12-core CPU will use two six-core die in a multichip package.
Analysts said AMD's updated road map should help it remain competitive with Intel while remaining appropriately conservative in the wake of AMD's heavy financial losses and delays for its four-core Barcelona server CPU.
In a briefing today (May 7), AMD said it will ship in early 2010 its Magny-Cours processor. It will use two six-core die of a 2010-class design called Sao Paolo liked via a novel interconnect the company would not discuss.
Randy Allen, general manager of AMD's server and workstation division, said the company rejected the approach Intel took to linking two die on a chip. The method it came up with achieves "nearly monolithic levels of performance," he said.
AMD claims its Barcelona server CPU has an average 13 percent performance advantage over Intel's chips, in part because Barcelona packs all four cores on one die while Intel's parts link two die in a package.
The current road map showed no new core designs. AMD has discussed a new high performance core called Bulldozer as well as a low power core, but neither appeared on the server road map through 2010.
The next step for AMD is its move from 65nm to 45nm process technology. A four-core server CPU dubbed Shanghai now sampling and set to ship before the end of the year will be among the first parts made in the 45nm process.
Shanghai will deliver a performance boost of about 20 percent over Barcelona thanks to the new process as well as a doubling of overall cache size and a move to 800 MHz DDR-2 memory.
In late 2009, AMD plans to ship Istanbul, a newly disclosed six-core processor, also made in 45nm, and likely to deliver a 20 percent performance boost over the Shanghai design. Analysts expect the 45nm technology could help AMD reduce silicon costs and get back to profitability after several loss-making quarters.
Allen would not comment on reports AMD might sell its fabs or form a joint venture around them. He did say the current road map takes into account any business changes with the fabs.
The 2010-class Sao Paolo will add to the six-core Istanbul design support for DDR-3 memory and support for four links of AMD's HyperTransport version 3.0 interconnect. Sao Paolo will also include a new probe filter to increase performance by reducing coherency traffic on the device. Alan said Sao Paolo will likely have a smaller than 20 percent performance increase over the 2009 Istanbul chip.
In 2010, AMD will introduce its own core logic chip sets for the server processors. They will support the 5 GHz version of PCI Express as well as a new spec for I/O virtualization over Express. Currently AMD servers use chip sets from Broadcom or NVidia.
For its part, Intel is expected to ship this fall its first server processors that put four and six cores on a single die, all using 45nm technology. The six-core Intel Dunnington design is expected to use more cache memory than the six-core AMD server chips.
"Overall I think you will see AMD remain competitive with Intel," said Nathan Brookwood, principal of market watcher Insight64 (Saratoga, Calif.) "They are not knocking the ball out of the park as they did a few years ago with the original Opteron design, but they are staying in the game, and that's the best they can do right now," he said.
Dean McCarron, principal of Mercury Research (Cave Creek, Ariz.), agreed.
With its move to a multichip 12-core CPU "AMD is making the same business decision as Intel for getting to the most performance and revenue per silicon area," said McCarron. "In light of AMD's financial position that makes a lot of sense. They are being judicious," he added.
The current Barcelona chip, originally scheduled for release late last year, has been shipping about six weeks, Allen said. Dell, Hewlett-Packard, IBM and Sun are expected to roll out systems using it before June.
"We are optimistic we have resolved all the execution issues" with Barcelona that could impact future products, he said.
Allen said the Shanghai design was created by teams in several locations with significant blocks designed in Austin and India. "This was the first product where our India design center took substantial leadership in chip integration and shows this center is becoming a big part of our strategy," he said.
Analysts said AMD's updated road map should help it remain competitive with Intel while remaining appropriately conservative in the wake of AMD's heavy financial losses and delays for its four-core Barcelona server CPU.
In a briefing today (May 7), AMD said it will ship in early 2010 its Magny-Cours processor. It will use two six-core die of a 2010-class design called Sao Paolo liked via a novel interconnect the company would not discuss.
Randy Allen, general manager of AMD's server and workstation division, said the company rejected the approach Intel took to linking two die on a chip. The method it came up with achieves "nearly monolithic levels of performance," he said.
AMD claims its Barcelona server CPU has an average 13 percent performance advantage over Intel's chips, in part because Barcelona packs all four cores on one die while Intel's parts link two die in a package.
The current road map showed no new core designs. AMD has discussed a new high performance core called Bulldozer as well as a low power core, but neither appeared on the server road map through 2010.
The next step for AMD is its move from 65nm to 45nm process technology. A four-core server CPU dubbed Shanghai now sampling and set to ship before the end of the year will be among the first parts made in the 45nm process.
Shanghai will deliver a performance boost of about 20 percent over Barcelona thanks to the new process as well as a doubling of overall cache size and a move to 800 MHz DDR-2 memory.
In late 2009, AMD plans to ship Istanbul, a newly disclosed six-core processor, also made in 45nm, and likely to deliver a 20 percent performance boost over the Shanghai design. Analysts expect the 45nm technology could help AMD reduce silicon costs and get back to profitability after several loss-making quarters.
Allen would not comment on reports AMD might sell its fabs or form a joint venture around them. He did say the current road map takes into account any business changes with the fabs.
The 2010-class Sao Paolo will add to the six-core Istanbul design support for DDR-3 memory and support for four links of AMD's HyperTransport version 3.0 interconnect. Sao Paolo will also include a new probe filter to increase performance by reducing coherency traffic on the device. Alan said Sao Paolo will likely have a smaller than 20 percent performance increase over the 2009 Istanbul chip.
In 2010, AMD will introduce its own core logic chip sets for the server processors. They will support the 5 GHz version of PCI Express as well as a new spec for I/O virtualization over Express. Currently AMD servers use chip sets from Broadcom or NVidia.
For its part, Intel is expected to ship this fall its first server processors that put four and six cores on a single die, all using 45nm technology. The six-core Intel Dunnington design is expected to use more cache memory than the six-core AMD server chips.
"Overall I think you will see AMD remain competitive with Intel," said Nathan Brookwood, principal of market watcher Insight64 (Saratoga, Calif.) "They are not knocking the ball out of the park as they did a few years ago with the original Opteron design, but they are staying in the game, and that's the best they can do right now," he said.
Dean McCarron, principal of Mercury Research (Cave Creek, Ariz.), agreed.
With its move to a multichip 12-core CPU "AMD is making the same business decision as Intel for getting to the most performance and revenue per silicon area," said McCarron. "In light of AMD's financial position that makes a lot of sense. They are being judicious," he added.
The current Barcelona chip, originally scheduled for release late last year, has been shipping about six weeks, Allen said. Dell, Hewlett-Packard, IBM and Sun are expected to roll out systems using it before June.
"We are optimistic we have resolved all the execution issues" with Barcelona that could impact future products, he said.
Allen said the Shanghai design was created by teams in several locations with significant blocks designed in Austin and India. "This was the first product where our India design center took substantial leadership in chip integration and shows this center is becoming a big part of our strategy," he said.
AMD says Intel paid PC makers to boycott rival
Advanced Micro Devices has expanded the allegations in its antitrust suit against Intel, saying its larger rival paid computer makers not to do business with AMD.
In court papers filed May 1 in U.S. District Court in Delaware, AMD said Intel paid off Acer, Dell, Gateway, major Japanese manufacturers, and system builders and distributors "to close their doors to AMD." Where it couldn't buy exclusivity from computer makers, Intel focused its payments on keeping computers based on AMD microprocessors away from large business customers, AMD said.
Most of the details in the 108-page document, unsealed Monday, have been blacked out, due to a protective order restricting the disclosure of trade secrets and other confidential dealings involving the parties in the suit. AMD sued Intel in June 2005, accusing the chipmaker of illegally using its monopoly to shut AMD out of the market.
The latest filing said Intel paid customers to abandon development of computer models that were essential to the success of a new line of AMD processors. The tactic also was used "to nip in the bud AMD inroads into sectors Intel viewed as critical."
In addition, Intel used discount schemes that made it uneconomical for AMD to compete for a customer's business, and deployed various strong-arm tactics against partners that refused to go along. Those retaliatory tactics included withdrawing marketing funds, withholding critical technical and road map information, allocating scarce products away from disloyal companies, and generally scaling back the level of customer support, the filing said.
In its own brief, Intel denied any anticompetitive practices. The company said that AMD was seeking court protection from discounting and other legal practices within a competitive environment.
"AMD's complaint about Intel's discounting boils down to a complaint that Intel is a more efficient competitor," Intel said.
For a few years before 2006, AMD managed to take market share from Intel, particularly in the market for processors powering server computers. About two years ago, Intel embarked on a reorganization that saw the chipmaker get several important products out before its smaller rival, including a quad-core processor and next-generation 45-nanometer chips that deliver higher power-to-performance ratios.
In court papers filed May 1 in U.S. District Court in Delaware, AMD said Intel paid off Acer, Dell, Gateway, major Japanese manufacturers, and system builders and distributors "to close their doors to AMD." Where it couldn't buy exclusivity from computer makers, Intel focused its payments on keeping computers based on AMD microprocessors away from large business customers, AMD said.
Most of the details in the 108-page document, unsealed Monday, have been blacked out, due to a protective order restricting the disclosure of trade secrets and other confidential dealings involving the parties in the suit. AMD sued Intel in June 2005, accusing the chipmaker of illegally using its monopoly to shut AMD out of the market.
The latest filing said Intel paid customers to abandon development of computer models that were essential to the success of a new line of AMD processors. The tactic also was used "to nip in the bud AMD inroads into sectors Intel viewed as critical."
In addition, Intel used discount schemes that made it uneconomical for AMD to compete for a customer's business, and deployed various strong-arm tactics against partners that refused to go along. Those retaliatory tactics included withdrawing marketing funds, withholding critical technical and road map information, allocating scarce products away from disloyal companies, and generally scaling back the level of customer support, the filing said.
In its own brief, Intel denied any anticompetitive practices. The company said that AMD was seeking court protection from discounting and other legal practices within a competitive environment.
"AMD's complaint about Intel's discounting boils down to a complaint that Intel is a more efficient competitor," Intel said.
For a few years before 2006, AMD managed to take market share from Intel, particularly in the market for processors powering server computers. About two years ago, Intel embarked on a reorganization that saw the chipmaker get several important products out before its smaller rival, including a quad-core processor and next-generation 45-nanometer chips that deliver higher power-to-performance ratios.
AMD: Next-Generation Microprocessors’ Will Solve Problems that Cannot Be Addressed with Hardware. :::
Advanced Micro Devices has been struggling to deliver higher-performance quad-core central processing units for months to address the market of enthusiasts, unfortunately, without much luck. But things may be more complicated than one may imagine and the chipmaker will continue to have hard times with higher-performance chips until the new code-named Bulldozer micro-architecture arrives.
AMD’s Athlon 64 and Opteron processors substantially reshaped the markets of desktop, server and workstation chips thanks to serious performance advantage over competing solutions from Intel, however, the new-generation code-named K10 central processing units (CPUs) can hardly leave Intel’s chips behind in terms of performance, but consume considerably more power. Apparently, AMD has problems with the design of its quad-core AMD Phenom and AMD Opteron processors that cannot be cured easily enough by tweaking the current hardware based on the current micro-architecture.
“If I look at the next-generation architecture of our CPU, then it will definitely not be, how can I say, comparable with the Phenom. It will look completely different. [It will] solve problems that today we think can never be addressed by hardware,” said Giuseppe Amato, technical director of sales and marketing for EMEA region at AMD in an interview with custom PC web-site .
AMD admitted in the past that it had issues with “wedding” the chip design with process technology, but denied the fact that it had issues with the micro-architecture or its current implementation when it comes to volume production. However, judging by the fact that the company still cannot deliver quad-core microprocessors clocked at 2.60GHz or higher, it has issues with volume manufacturing of higher-end K10 generation chips. If Mr. Amato’s comments touch upon the problems of higher-end chips’ volume ramp, then it will be pretty hard for AMD to compete against Intel for higher-end CPU market.
Bulldozer is the next-generation micro-architecture and processor design developed from the ground up by AMD. It is expected that the next-generation micro-processors will offer considerably higher performance than current-generation chips. AMD Bulldozer CPUs will feature
SSE5 instruction set. The first Bulldozer processors are projected to emerge on the market in very late 2009 at the earliest or early 2011 at the latest.
AMD’s Athlon 64 and Opteron processors substantially reshaped the markets of desktop, server and workstation chips thanks to serious performance advantage over competing solutions from Intel, however, the new-generation code-named K10 central processing units (CPUs) can hardly leave Intel’s chips behind in terms of performance, but consume considerably more power. Apparently, AMD has problems with the design of its quad-core AMD Phenom and AMD Opteron processors that cannot be cured easily enough by tweaking the current hardware based on the current micro-architecture.
“If I look at the next-generation architecture of our CPU, then it will definitely not be, how can I say, comparable with the Phenom. It will look completely different. [It will] solve problems that today we think can never be addressed by hardware,” said Giuseppe Amato, technical director of sales and marketing for EMEA region at AMD in an interview with custom PC web-site .
AMD admitted in the past that it had issues with “wedding” the chip design with process technology, but denied the fact that it had issues with the micro-architecture or its current implementation when it comes to volume production. However, judging by the fact that the company still cannot deliver quad-core microprocessors clocked at 2.60GHz or higher, it has issues with volume manufacturing of higher-end K10 generation chips. If Mr. Amato’s comments touch upon the problems of higher-end chips’ volume ramp, then it will be pretty hard for AMD to compete against Intel for higher-end CPU market.
Bulldozer is the next-generation micro-architecture and processor design developed from the ground up by AMD. It is expected that the next-generation micro-processors will offer considerably higher performance than current-generation chips. AMD Bulldozer CPUs will feature
SSE5 instruction set. The first Bulldozer processors are projected to emerge on the market in very late 2009 at the earliest or early 2011 at the latest.
Subscribe to:
Comments (Atom)
