Reports suggest that the Taiwan Semiconductor Manufacturing Company (TSMC) could receive manufacturing contracts by AMD, and long-time customer NVIDIA, for graphics processors based on the newest 40nm silicon fabrication process. The two giants in the visual computing industry, are expected to skip the 45nm process altogether. This is seen as a move to cut down manufacturing costs, and use the most feasible manufacturing technologies. Miniaturizing circuitry at that scale would allow them to build even more complex silicon machinery, with much higher transistor counts, while maintaining size constraints.

AMD on its part would have to use the services of TSMC, since the foundry company's fab in Dresden is only capable of 45nm SOI fabrication process, and that it would take as long as 2010, by the time the expected 32nm process is ready at the New York fab. The next star-entry for AMD would be the ATI RV870, and as for NVIDIA, it would continue development of monolithic high-end GPUs with GT216, a successor to the G200 GPU. Meanwhile, the companies could continue aggressive competition, with projections of up to 20% price-cuts by Christmas 2008 shopping season. Their 40nm GPU designs could be ready by the end of this year, and 40nm GPUs could be supplied by June, 2009.

The United States Department of Justice had issued notices to both NVIDIA corporation and AMD for their alleged involvement in malpractices with fixing prices in the graphics processor market, and inducing staged competition. Both companies were to be penalized if the antitrust investigations found the charges to be true. In its most recent press release, AMD states that the Dept. of Justice closed its investigation into these charges, and that no action would be taken against the company, in this regard. Press release follows:

On October 10, 2008, the U.S. Department of Justice informed AMD [NYSE: AMD] and ATI Technologies that it has closed its investigation into ATI Technologies' pricing and marketing practices in the sale of graphics processing units ("GPUs"). On December 1, 2006, AMD announced that ATI had received a subpoena from the U.S. Department of Justice.

AMD spokesman reassured today that the deal between AMD and Advanced Technology Investment Company won't violate any of the patent agreements AMD made with its main competitor Intel. Let me remind you that earlier that day Intel said it would review its processor licensing agreement with AMD, especially the agreement which allows AMD to use Intel's x86 chip instruction set, to ensure the ATIC deal does not violate any patent agreements. Phil Hughes, an AMD spokesman, wrote in an e-mail today:
"We are completely confident the structure of this transaction takes into account our cross-license agreements,"
"Rest assured - we plan to continue respecting Intel's intellectual property rights, just as we expect them to respect ours."
Financial analyst Hans Mosesmann even believes that the dispute between the two giants may become a weapon for Intel to solve some of its problems in court. If Intel manages to find something is wrong with this deal it can use it against Advanced Micro Deviced and make AMD drop the long-standing anti-trust suits against the company. Continue reading the full story here.

Here's an interesting story I found today. It's short, but I'm sure it won't end just like that and there'll be a lot more to discuss later.

Intel Corp. said on Tuesday it has "serious questions" about a deal announced by Advanced Micro Devices to spin off its manufacturing business and focus on chip design, and is ready to defend its patents. Intel spokesman Chuck Mulloy said that AMD and Intel have a patent cross-licensing agreement under which AMD pays royalties to Intel. "Intel has serious questions about this transaction as it relates to the license and will vigorously protect Intel's intellectual property rights," Mulloy said of AMD's announcement. Mulloy said Intel has asked AMD to make the agreement public but he said it would not, so he was not at liberty to discuss the matter in detail.

AMD and the Advanced Technology Investment Company (ATIC) of Abu Dhabi today announced the creation of a U.S.-headquartered, leading-edge semiconductor manufacturing company to address growing demand for independent, leading-edge foundry production capabilities. The new global company, to be temporarily called "The Foundry Company", will serve this need by combining advanced process technology, industry-leading manufacturing facilities and aggressive plans to expand its global capacity footprint. At the same time, the Mubadala Development Company will increase its current investment in AMD to 19.3 percent on a fully diluted basis. AMD will contribute to The Foundry Company its manufacturing facilities, including two fabrication facilities in Dresden, Germany, as well as related assets and intellectual property rights.

AMD conducted a presentation at CEATEC Japan, where the company took a sneak-peak at how the role of GPUs would become critical to the PC of tomorrow. This of course revolved around the company's newly adopted "The Future is Fusion" slogan, integrating all of AMD's technological expertise into object and function oriented solutions for the PC industry.

Among the numerous slides that formed part of the presentation, one such slide, shows some very interesting points on what the year 2009 looks like, from AMD's perspective. It shows a lot of things slated for much later to make it to the industry. To begin with, the DirectX 11 API and Windows 7 (Vienna) operating system could make it to the industry in 2009. However, there's no mention of them being "released" as such, or if they could just be working prototypes, such as alpha releases for use by select parts of the industry for mutual technology development.

To combat Intel's Atom-embedded ITX motherboard linueup, and the platform itself, there has been an influx of mini-ITX motherboards based on the AMD 780G chipset lately. Perhaps the thermal design of the 780G allows implementation in form-factors this small. To combat the dual-core Atom, AMD had built up plans to release low-power processors based on the AM2 socket. Notably, the Athlon X2 22W and Athlon single core 15W chips. Perhaps it is these processors that are inviting companies to plan Mini-ITX motherboards based on the AM2+ socket with the 780G chipset.

Albatron is about to release the KI780G motherboard, and they seem to have packed quite a lot to it. To begin with, the board features an AM2+ socket (that supports processors only upto 65W TDP), two DDR2 SO-DIMM slots for using the more compact memory sticks notebooks use, four SATA II ports, an IDE connector, Gigabit LAN, 7.1 channel audio with SPDIF optical and co-axial pass, display connectors that include D-Sub, DVI and HDMI (HDCP capable), and finally a PCI-Express 2.0 x16 slot. The motherboard goes head-on against the feature-set that the MINIX comes with. This is meant to be a full-featured HTPC board. Its pricing is not known at the moment.

According to sources at server builders, AMD has notified that it would be releasing nine models of quad-core AMD Opteron 45nm processors this month. The processors would be based on the Shanghai core and mark the company's entry into the 45nm silicon fabrication for CPUs.

Out of the lineup of 20 processors that would come out, the ones slated for this month include CPUs of clock-speeds between 2.3 and 2.7 GHz, of which five of them are 2000 series (2-way) models and the rest, 8000 series (8-way) models. The processors are based on Socket F (s1207) design, and sport 6 MB level 3 caches. The sources also mentioned that in February 2009, AMD will launch five 55W energy-efficient enterprise CPUs, three 2-way and two 8-way; and two 105W high-performance models a 2-way and an 8-way each.

AMD Shanghai is the codename for a new enterprise processor the company is working on. It is a monolithic quad-core x86 processor that retains the design methodology of the K10h generation of AMD processors. It carries forward the Opteron brand. This time around, the company promises to have made numerous changes, big and small, learning from the shortfalls of the Barcelona chip. Shanghai will be the first 45nm AMD processor to make it to the market. Originally slated for Q1, 2009 its launch was pushed to Q4, 2008.

In a press event held yesterday, the company claimed that the company is all set to roll out the new chip. Initial volumes of the processor would run off existing core logic platforms by NVIDIA and Broadcom. It will retain the Socket F (aka s1207) package and support DDR2 memory. The lineup shows that most of these chips would retain the 2000 MT/s system bus, and all of them come with 6 MB of level 3 cache. A new company slide sourced by C|Net shows that the Opteron processor could embrace DDR3 memory only by 2010, with the introduction of the Magny-Corus 12-core processor architecture. Until then, the Istanbul 6-core processor the is slated for H2, 2009, could be the company's flagship product.

Performance projections for the upcoming Radeon HD 4830, made news yesterday, where it is projected to outperform the GeForce 9800 GT. Two new slides published by VR-Zone shed light on how the RV770LE graphics processor manages to do that, and there's a surprise: It was earlier known that the RV770LE comes equipped with 480 stream processors, and one of the earliest reports even suggested that it could have a narrower 192-bit memory bus. The slides published, show that the new GPU could carry as many as 640 stream processors, double that of the RV670. It comes with a proper 256-bit wide GDDR3 memory bus. Also included, is hardware support for 24x custom filter anti-aliasing (CFAA).

It is also known that the new card is gearing up for an October 21 launch and would be priced below US $150. AIBs would be able to sell custom designed cards right upon launch. While the new GPU does look "tougher" and more equipped than expected, here's something to ponder: with twice the amount of stream processors as the Radeon HD 3870, as projections suggest, is it going to be only 10% faster than the 9800 GT?

As reported earlier, AMD is preparing a scaled-down version of the RV770 graphics processor, called RV770LE. The new graphics core would go into making the Radeon HD 4830 SKU. The new graphics core was created to compete with the GeForce 8800/9800 GT, which seem to be in a comfortable market position owing to their price and price/performance. They are comfortably placed between the GeForce 9600 seties and the 8800 GTS 512M/9800 GTX/GTX+.

AMD has already started projecting the HD 4830 to outperform the the G92GT, in essence, the GeForce 9800 GT. In a slide titled "Best Graphics Card for <$150", the company puts up figures where the new card is projected 10% faster in most games tested. The only glitch, is that most of AMD's own increment projections are based around that 10% mark, which means it could be a very tight contest between the HD 4830 and 9800 GT. A prelude could be that for the 9800 GT, its CUDA and PhysX processing capabilities could bail it out, if it does fall into a tight spot, as in, if the two end up in a neck and neck contest. The new Radeon accelerator comes out sometime next month.

AMD would be releasing 45nm silicon fabricated series of the quad-core Opteron processor based on the newer Shanghai core. Like its desktop PC counterpart, the Deneb core, it features 6 MB of L3 cache along with some enterprise-specific features. They will be designed for high-uptime and energy efficiency. AMD banks heavily on the performance efficiency and virtualization capabilities of these processors. There are primarily two types of these processors based on the multi-processor capabilities: the 2000 series 2-way processors and the 8000 series 8-way processors. The former supports two-socket systems while the latter is designed for dense 2+ socket environments, where up to 8 Opteron processors could work in tandem. The processors continue to have Socket-F (s1207) interfaces, hence the entire lineup retains its DDR2 memory standard.

Interestingly, AMD did not rework the system interface for all its newer chips. Across the vast lineup of 20 models, only four support the newer HyperTransport 3.0 interface with 4400 MT/s interconnects. The rest continue to use the 2000 MT/s interconnects. It could be mid-2009 by the time AMD launches its own enterprise-segment chipsets, which could also be the time when the current leading player, NVIDIA launches its newer nForce Professional lineup for the Opteron. Since current chipsets from both NVIDIA and Broadcom feature only the older HyperTransport 2000 MT/s interfaces, it could explain why most of these chips feature it.

AMD would be releasing its first desktop processors based on the 45nm silicon fabrication process, based on the newer Deneb core. The company is said to have made several tweaks to the original K10 design and equipped the core with 300% the amount of L3 cache as its 65nm Agena parts. The only thing got us wondering was what would be its nomenclature like? Well, be surprised to know that after Phenom X4 9000 series, the company plans to continue the numbering with a 5-digit model number scheme with x1000 unit deviations between models. A rather confusing naming scheme, as suggested by the chart provided, seems to have been adopted.

It is now clear, that there will be two distinct kinds of Phenom X4 45nm chips: those which continue support for DDR2 memory on the existing AM2/AM2+ sockets, and those which are exclusive to the AM3 socket and feature support for DDR3 memory, DDR3 1333MHz at that. The processors would feature dual 64-bit memory controllers, which could be ganged for a single 128-bit wide memory interface, or un-ganged to step up multi-tasking efficiency.

IN the past few months, an interesting set of Mini-ITX motherboards made it to the market. There are two primary classes of ITX boards: embedded CPU and those with CPU sockets. VIA and Intel have boards based on their small-footprint embedded processors, while others made innovative solutions with CPU sockets. The Minix-ITX, an AM2+ socket based ITX board, took it to the next level with a motherboard that fulfills most requirements of HTPC users.

Magic Pro, a Chinese company that concentrates on Asian markets has now come up with another solution based on the AMD 780G chipset. The board, MP-NC81 comes with 128 MB of Sideport memory to assist the integrated graphics processor (IGP) that's housed in the 780G chipset. The IGP supports UVD, it accelerates H.264, VC1 and MPEG2 1080p playback. It comes with SB600 southbrige which routes to four SATA II ports with RAID support. The board features a single SO-DIMM slot to hold notebook DDR2 memory. It supports AM2+ processors with TDP of up to 95W. A striking feature however, is that the board optionally ships with a wireless ethernet riser-card, much in the same way as some high-end motherboards from ASUS do. The riser card is said to support 802.11n standard.

The AM3 socket finally gives AMD Processors access to DDR3 memory. The only complication is that processors must feature DDR3 integrated memory controllers (IMC). Earlier roadmap slides had shown that the RD890 and RS880 along with SB800 would be the first platform with AM3 compliance. However, sources tell Expreview that AMD would not be ready with the chipsets by the time AM3 processors come out. Hence, AMD would be using the existing high-end duo of 790 FX and SB750 which would be re-launched in H1 2009 that support AM3 sockets. It is known that 790 FX supports HyperTransport 3.1 links that would offer processors system interconnect bandwidths of up to 6400 MT/s.

Another chipset that has been pushed to H1 2009 is the AMD 760G, which is a mainstream chipset with integrated graphics. It was expected that this would come out sometime in Q4 2008.

AMD today launched a new corporate brand campaign under a new tagline, "The Future is Fusion." The campaign is intended to focus on how the unique AMD combination of technologies, coupled with close relationships with computer manufacturers and a deep understanding of customer needs, results in exciting next-generation capabilities and experiences at work, at home, and at play.

Every second counts when it comes to transitions in games! MSI GX711 provides with the enormous 17-inch LCD monitor with powerful Ati HD3650 Graphic with 512 MB VRAM, GX711 can achieve the ultimate realism experience whether you are playing games or watching a DVD and create the most astonishingly realistic audio and visual effects. Moreover, in order to keep the edge in the fast paced world of intense gaming competition, GX711 especially adds the colorful W, A, S, D buttons on the keyboard, so the player can find the direction controls in record time. To deliver real-life listening enjoyment, GX711 is recognized and approved by the highest standards of Dolby, theater-class 4 high quality speakers and built-in subwoofer.

GX711 features AMD Turion X2 Ultra dual-Core / Sempron Mobile Processor Technology which uses Ati RX690+SB600 chipset to deliver truly outstanding dual-core performance computing power, great battery life, and expanded wireless connectivity. Additionally GX711 is equipped with a 3 well distributed USB 2.0 ports, IEEE1394 connecting interface and a Muti-DVD burner for you to easy editing and saving high-quality photos/videos.With blistering speeds and true-to-life video and audio performance on-the-go,GX711 is gamers' first choice for gaming notebook. The newest Ati HD3650 Graphic with 512 MB VRAM delivers remarkable performance and an unsurpassed feature set for accelerating spectacular gaming. For a more fluid, more detailed, more realistic 3D high-definition while playing the game. Never a moment of delay to occur, feel the excitement.

The battle for supremacy in the current generation of consumer graphics hasn't calmed down as yet. Just as reports came in of NVIDIA giving the GeForce GTX 260 a boost with its shader compute power by enabling 24 shader units, taking on the Radeon HD 4870, GeForce GTX 280 seems to be enjoying its $420~$440 price-tag. For sure it isn't the fastest graphics accelerator but its compelling price tag is what is drawing buyers away from the HD 4870 X2.

AMD had already announced in its press release for the Radeon HD 4870 X2 launch that it would release a HD 4850 X2 at a roughly $400 price range. That moment seems to be coming closer, when ATI has a graphics card for $399, that outperforms the GeForce GTX 280. Three slides from the company have surfaced. The first one lists out its vital specifications, including its maximum power draw, rated at 230W. The memory bandwidth of this card is 128 GB/s, and it ends up with 2 GB of GDDR3 memory.

AMD/ATI's graphics partner Sapphire has released today Radeon HD 4870 TOXIC, an overclocked Radeon HD 4870 card with a custom cooling on it. What makes this card so special is the patented Vapor-X cooling system and its clock speeds higher than the reference boards. The cooling itself takes up two slots, makes use of a Vapour Chamber technology and features a thermally-controlled fan supported by three copper heatpipes. The new TOXIC model is also factory-overclocked to 780MHz GPU from 750MHz stock speed and 4000MHz memory from 3600MHz stock respectively. The card is equipped with 512MB GDDR5 memory. It offers 800 Stream Processors, has DirectX 10.1 and CrossFireX support, a PCI-Express 2.0 interface and two DVI connectors. Except that and maybe the fancy retail box it ships with, there's nothing else about this card that could interest you more. Pricing and availability are still unknown.

As much as releasing hardware that brings in competition and products at great prices is essential for AMD, backing it up with software is equally important. Company slides that point to tentative time-scales that pertain to AMD's software releases have surfaced. Some of the important software products AMD releases are Catalyst driver suite for ATI products and OverDrive, a tool that provides features to tweak AMD processors and graphics cards, as well as several motherboard parameters for motherboards equipped with AMD chipsets.

The slides also provide a sneak-peak into what could be in store with those releases. VR-Zone has published slides, from which the first one points to release dates for the Catalyst driver suite with respect to release candidates and public releases. The second slide points to time-scales at which software with vital changes are released. We are already past August and AMD has rolled out the feature pertaining to that release, allowing Hybrid Crossfire of AMD chipsets with integrated graphics working in tandem with Radeon HD 3400 series graphics accelerators.

Later this month, AMD is expected to launch the ATI Radeon HD 4550 graphics processor (GPU). Its board SKU details have surfaced. The RV710 graphics processor will make it to the grass-root of entry-level and high-definition media grpahics boards. It is based on the 55nm silicon fabrication process. The GPU has a transistor count of 242 million, it has 80 stream processors. It is engineered in a way that makes it draw less than 25W. In the form of reference designs, AMD will release two cards: a full height, passively cooled 512 MB DDR3 version, and a low-profile, 256MB fan-heatsink cooled card. Both models have core clock speeds of 600 MHz with the memory clocked at 800 MHz on a 64-bit wide memory bus. The GPU schematic shows it to have 8 texture address and 8 texture filter units. There's only one render back-end. The 512MB variant is expected to be priced at US $49~59, while the 256MB variant in the $39~49 range. These cards are expected to compete with the GeForce 9400 GT GPU.

Computer Aided Design (CAD) and Digital Content Creation (DCC) professionals who require maximum performance from their workstations can turn to AMD and the powerful new ATI FirePro V8700 graphics accelerator, which provides a 40 percent performance gain1 for memory intensive applications. Additionally, the new ATI FirePro V3750 delivers superior performance in an entry class professional graphics accelerator. Delivering on the AMD commitment to the professional graphics market, these products join the ATI FirePro family of professional graphics accelerators introduced last month at Siggraph 2008 in Los Angeles: ATI FirePro V5700 and ATI FirePro V3700.

AMD plans to offer two processors for the ULPC. Cost-effectiveness being the mantra, AMD plans to offer processors that provide users of ULPC with a level of performance that takes it closer to that of regular PCs and eradicate the netPC/netbook moniker as the processors offer more performance at a low cost than what the likes of Atom or Nano could. Of course this comes at the expense of much higher TDP and slightly higher price. Perhaps the performance to price ratio is what AMD is banking on.

There are two parts under the Athlon X2 banner. The 22W, 1.50 GHz dual-core 3250e processor aims to compete with the desktop variant of Atom dual core where the Athlon's architectural superiority aims to give it a leading edge. The issue of energy savings dampens when it comes to desktop. Another processor, the single-core Athlon 2650e that is available now is rated at 15W with a 1.60 GHz clock speed. It comes with 512 KB L2 cache.

The chipset division of AMD really is reminiscent of former ATI Technologies that jumped into the business of making chipsets looking at rival NVIDIA's success with nForce series. After being acquired by AMD, the chipset division started making chipsets under the AMD banner. AMD has had a good amount of success with its recent 7-series chipsets for the desktop platform. It has managed to snatch a 20% market share away from NVIDIA which even today leads in market share for platform core-logic technologies for AMD. Not only have they had a desktop line of nForce products but also an equally popular enterprise chipset for the Opteron platform in the form of nForce Professional series server and workstation core-logic.

Phenom X2 is intended to be the latest dual-core processor from AMD. It is based on the newer K10 architecture. The 65nm Kuma core is what sits inside the first to release models of the Phenom X2. Although based on K10 architecture and Kuma core, the initial batch being 65nm, AMD for some reason chooses to call it Athlon X2 saving the Phenom brand name perhaps for the 45nm batches? At least the sample Expreview got had the Athlon X2 etching on it. The chip carried the "AD6500" label and came with a 2.30 GHz clock speed.

Here's something to ponder: 6500 isn't a performance rating, it is just a model number. The Athlon 64 X2 6400+ had a 3.20 GHz clock speed. As performance evaluations will soon show, the Phenom/Athlon 6500 isn't anywhere close to the performance of the X2 6400+. However, architectural improvements meant that Kuma outperforms Brisbane (K8, 65nm) on a clock to clock basis. To ascertain this, Expreview used a Athlon 64 X2 5000+ Black Edition (Brisbane) that was underclocked to 2.30 GHz, the clock speed which AD6500 comes with. Super Pi 1M benchmark was run. While the Brisbane chip crunched it in 39.374 s, Kuma did it in 33.43 s indicating a performance increment of roughly 15%. The test-bed consisted of a NVIDIA nForce 750a SLI motherboard and GEIL 1GB DDR2-800 memory. Kuma AD6500 comes with a shared L3 cache of 2 MB apart from dedicated L2 caches of 512 KB per core. It uses a broader HyperTransport 3.0 system interface at 3600 MT/s. It supports DDR2-1066.