Intel D945GCLF (with integrated Atom 1.6Ghz CPU) BOX
- Intel Atom Silverthorne !
- 1.6Ghz !
- Intel Rapid BIOS Boot
- 4x Rear USB 2.0
Intel D945GCLF (with integrated Atom 1.6Ghz CPU) BOX
Intel's 'Tolapai': for workstations, not tellies
The QuickAssist Accelerator ties the SoC into Intel's QuickAssist Architecture, its framework for application-specific co-processors that connect to the host computer's CPU via the standard frontside bus. This is not so very different from what AMD is proposing with its Torrenza programme, which, in part, is about establishing a framework for application accelerators that connect to the AMD CPU via the HyperTransport bus.
Intel's own take on HyperTransport, QuickPath Interconnect, is due to debut late next year as a key element of its 45nm 'Nehalem' processor architecture, and the use of the 'Quick' prefix - QuickPath, QuickAssist - may not be entirely coincidental.
Incidentally, Intel and IBM are working to deliver the same kind of approach, but using PCI Express add-in cards. That project, codename fans, is called 'Geneseo'.
Canmore, of course, doesn't need any of this, so is likely to come to market much sooner than Tolapai, especially if Intel's keeping the enterprise SoC waiting in the wings until Nehalem-based machines debut at the end of 2008.
Intel News Release
Researchers from Intel Corporation and the University of California, Santa Barbara (UCSB) have built the world’s first electrically powered Hybrid Silicon Laser using standard silicon manufacturing processes. This breakthrough addresses one of the last major barriers to producing low-cost, high-bandwidth silicon photonics devices for use inside and around future computers and data centers.
The researchers were able to combine the light-emitting properties of Indium Phosphide with the light-routing capabilities of silicon into a single hybrid chip. When voltage is applied, light generated in the Indium Phosphide enters the silicon waveguide to create a continuous laser beam that can be used to drive other silicon photonic devices. A laser based on silicon could drive wider use of photonics in computers because the cost can be greatly reduced by using high-volume silicon manufacturing techniques.
“This could bring low-cost, terabit-level optical ‘data pipes’ inside future computers and help make possible a new era of high-performance computing applications," said Mario Paniccia, director of Intel’s Photonics Technology Lab. "While still far from becoming a commercial product, we believe dozens, maybe even hundreds of hybrid silicon lasers could be integrated with other silicon photonic components onto a single silicon chip.”
"Our research program with Intel highlights how industry and academia can work together to advance the state of science and technology," said John Bowers, a professor of electrical and computer engineering at UC Santa Barbara. “By combining UCSB’s expertise with Indium Phosphide and Intel’s silicon photonics expertise, we have demonstrated a novel laser structure based on a bonding method that can be used at the wafer-, partial-wafer or die-level, and could be a solution for large-scale optical integration onto a silicon platform. This marks the beginning of highly integrated silicon photonic chips that can be mass produced at low cost.”
Turbo Memory
Turbo Memory, formerly code-named Robson, is a NAND flash hard disk cache that's designed to reduce boot times, boost performance and conserve battery life.
More performance, and lower power consumption, comes from a flash-based hard disk cache called Turbo Memory. Previously codenamed Robson, this devotes 512MB or 1GB of NAND flash memory to holding system files and other data that the operating system requires at start-up, restoring from sleep or during application switches. Intel claims that with a suitable operating system -- and Windows Vista is designed for this -- Turbo Memory can more than halve application load and wake-from-sleep times, while adding up to twenty minutes to battery life. Like any cache, its performance can vary according to how it's used, the mixture of reads and writes and the nature of the data sets it handles. However, with luck and application, the advantages will increase as programmers learn how to write optimal code and as flash memory gets bigger and better.
Inside Intel's Santa Rosa platform
Processor
The processor has changed the least. It's a 65nm Merom-based Core 2 Duo with a few tweaks, and will be the last of these CPUs before the arrival of 45nm Penryn processors in the next update. The major change between Santa Rosa's processor and its predecessors is the acceleration of the frontside bus (FSB) from 667MHz to 800MHz , together with the ability to slow it down during light loads. This technology, called Super Low Frequency Mode, knocks the bus speed down to 400MHz, the CPU's clock speed to 600MHz and reduces the core voltage. Another tweak, called Enhanced Deeper Sleep, means that the memory controller in the chipset won't wake up the processor to check on cache status if the system knows the cache is empty and the CPU is sufficiently comatose not to be able to change things.
A rather unusual performance boost comes in the form of Enhanced Dynamic Acceleration Technology. The Santa Rosa processor can run one of its two cores at full tilt while the other is in one of a number of sleep states, at which time the total thermal energy is somewhat less than the maximum allowed when both cores are running. Under conditions where only one core is running, Intel can effectively overclock it, taking advantage of the increased thermal headroom left by the other core ticking over. This will give single-threaded tasks a boost, says Intel, without risking overheating.
Centrino Pro is the business version of the Santa Rosa platform
Improvements to the processor, chipset and wireless components of Intel's latest mobile platform should result in a new generation of faster notebooks with longer battery life. Business systems will also get Intel's Active Management Technology (AMT) for the first time, while Turbo Memory should reduce the frequency of hard disk accesses, saving power and boosting performance
Intel's AMD Antitrust Rebuttal: We're Not Bad, You're Just Stupid
Advanced Micro Devices Inc.'s failure to compete effectively with Intel in the microprocessor market is a "direct result of AMD's own actions or inaction," and weren't caused by any illegal actions by Intel, the microprocessor market leader said in court documents filed Thursday.
The filing in U.S. District Court was Intel's first formal response to AMD's lawsuit filed in June that charged Intel with using bribery and coercion of computer makers and retailers to limit the use of AMD processors.
Specifically, Intel points to AMD's "playing it safe ... with anemic investment in manufacturing capacity, leaving Intel to shoulder the burden of investment to enhance the usefulness of computers and enhance the market." In addition, Intel claims AMD has been "dogged" by a reputation of being unreliable as a supplier, has traditionally lagged in innovation, and has seen products delayed well beyond original launch dates.
When AMD has been able to deliver competitive products, such as with its Opteron server processors, it has seen share gains, Intel said.
AMD's allegations center primarily on the PC market, where chairman and CEO Hector Ruiz says the company has been unable to increase market share despite having highly competitive products in the past year. According to Mercury Research, AMD has seen its share of all x86 processors shipped, including those used in PCs and servers, rise from 15.1% in the second quarter of 2004 to 16.2% in the second quarter of 2005. When counting only x86 processors used in server applications, AMD has fared better, with its market share increasing 51% in the second quarter, from 7.4% to 11.2%
Intel argued that it hasn't violated any law or committed any wrongdoing. Intel said the AMD lawsuit is full of contradictions, including the claim that Intel's competitive actions could render it "nonviable." Ruiz has stated that the company "is in the strongest position we've ever been in."
Intel said it has demonstrated over a 30-year period that it can deliver faster, better, and cheaper products to consumers, in part because of its continued multibillion-dollar investment in new capacity and research and development even during downturns in the semiconductor market.
"In short," Intel wrote in its response, "AMD's colorful language and fanciful claims cannot obscure its goal of shielding the company from price competition."
Tom McCoy, executive of legal affairs and chief administrative officer for AMD, says Intel's response is "an attempt to divert attention from what this case is really about. Crossing swords of rhetoric is fun, but it's time to put the truth on the table in the courtroom. The industry problem is that [computer manufacturers and retailers] are under threats from retaliatory and potentially lethal price increases by Intel if they give too much business to AMD."
Evidence of Intel's illegal practices has already been substantiated by a prior ruling by the Fair Trade Commission of Japan that found Intel guilty of antitrust violations, and by ongoing investigations in Europe, McCoy says.
"There is an global regulatory focus on how Intel controls the IT industry, and it is plainly a serious industry problem that requires the light of truth," he says.
McCoy dismissed Intel charges that AMD's difficulty in competing was of its own making, pointing to a 25-year history of AMD consistently being a low-price leader in the x86 processor market, and the company's plans to open a 300-mm wafer fab in Dresden, Germany in October.
AMD was quick to respond to the charges. "It is true that with our early K6 processor, we had difficulty ramping one of our steppers. That hurt us for a couple of quarters," Thomas McCoy, AMD's executive vice president of legal affairs told internetnews.com. "But it's overwhelmingly true that we've been a worthy competitor for over twenty years, and now we have the technology lead. The problem is that Intel is using its monopoly power to force the world to accept Intel monopoly prices and is illegally insulated from having to compete on price and performance. We know that Intel is not the price leader, AMD is. Intel maintains a very constant monopoly margin it's proud of."
Intel did furnish a backhanded compliment, however, giving AMD credit for its success with the 64-bit Opteron processor for the server market, where it brought out a dual-core offering ahead of Intel. "When AMD is able to combine competitive products with reliable supply, the market responds," said Intel.
"Let's not debate it anymore," said AMD's McCoy. "We brought this case to put light on the truth. We'll put it in a courtroom and the world will see for itself."
GigaByte GA-EP45-UD3P Motherboard
(Intel Socket T LGA775, P45 Express, ATX, 16GB DDR2, 1600MHz Bus - MPN: GAEP45UD3P)
Price Range: $134.99 - $151.07 from 6 Sellers
Description: The GA-EP45-UD3P features a 2 oz copper PCB to help reduce system temperature and dramatically enhance performance. Based on the Intel P45 chipset, the GA-EP45-UD3P delivers a 1600MHz FSB for Intel Core 2 multi-core processors, and supports dual channel DDR2 memory at speeds of up to 1366MHz and even higher. The GA-EP45-UD3P has dual PCI Express 2.0 VGA card slots for ATI CrossFireX support with smooth 3D rendering and fast frame rates. Additionally, the board incorporates GIGABYTE DES Advanced technology for industry leading energy saving and enhanced system efficiency.
Intel's Core 2 Duo processors
Intel's Core 2 Duo processors have been available for about six months now, and the response has been very favorable for Intel. In fact, many people send us emails asking our views on the best platform for Core 2 Duo systems. Until now the answer has been an Intel chipset, but competition has emerged: Nvidia's nForce 680i SLI core logic wants to be the undisputed champion for Core 2 Duo processors. Of course, we wanted to know whether it really is or not...
First of all, we have to make very clear that although today's chipsets support varying features and target different user segments, you won't notice any difference in performance between a motherboard using a P965, 975X or nForce 680i SLI chipset when you start Windows to attend to your daily business. Only if you belong to the feature-aware, performance-hungry or overclocking-savvy crowds will you appreciate the huge overclocking margins and plethora of tweaking options.
Intel's P965 and 975X chipsets represent the establishment, both because most Intel-based systems are powered by Intel's very own chipsets, and because Intel processors have traditionally run best with an Intel chipset. Although the 975X has been available for over a year, it is still the high-end product - it is paired with the ICH7 southbridge, and thus offers a nice range of interface options. However, the P965 is the latest core logic product, utilizing the ICH8 southbridge family with even more goodies. There is a large variety of 965 and 975 motherboards available today.
Nvidia has been in the chipset arena since 2001. Although its first nForce chipset failed, the nForce2 and following generations were increasingly successful. Today, the nForce 4 and 5 series are considered among the best choices for Athlon 64 solutions, and Nvidia is hoping that the nForce 6 will give it a big share of the Intel market. It is the larger business space, but also more difficult to get into: Nvidia already tried its luck with the nForce 4 Intel Edition, but even though the product was great, it couldn't take significant market share from Intel.
The nForce 6 chipset is supposed to be better than the Intel chipsets in every way: better overclocking, better storage subsystem, better dual x16 PCI Express SLI graphics as opposed to Intel's dual x8 PCI Express Crossfire support, better networking support and better memory auto-configuration. Let's dig into this new chipset and find out if it delivers.
abit launched ther latest motherboard, the I-G31!
abit launches its latest "Multimedia Platform" motherboard, the I-G31. With the Intel G31 chipset, the abit I-G31 supports Intel Core 2 Extreme/Quad/Duo and Pentium Dual-Core processors. The Intel Graphics Media Accelerator 3100 doesn't Do DirectX 10, but it does offer support of Microsoft DirectX 9.0c, Shader Model 2.0 and OpenGL 1.4 technology.
Subscribe to:
Posts (Atom)
