Sunnyvale (CA) – AMD today announced its RS690 chipset which is not only AMD’s first-ever integrated graphics chipset solution, but also the first product that takes advantage of technologies from both AMD and ATI. The 690 will attack Intel’s dominance in the high-volume mainstream market of desktop PCs and position the chipset as a product that is superior to Nvidia chipsets.A standalone graphics card may be more exciting than an integrated graphics chipset (IGCs) in a PC, but there is no denying that many PC buyers believe that integrated is good enough for their needs. Analysts estimate that roughly two thirds of all PCs sold include fairly cheap integrated graphics that will do just about anything that users throw at them: Some chipsets already have begun supporting Windows Vista’s demanding AeroGlass interface and these low-end graphic devices will only capitulate if you load the latest 3D games or run HD video.
Integrated graphics is a key business for several chip manufacturers, foremost ATI, Intel and Nvidia. The segment is strong enough that Intel’s dominance in selling IGCs has allowed the company to become the unchallenged overall graphics chip supplier. In the fourth quarter of last year, Intel held a commanding 37.4% share in the graphics market, according to Jon Peddie Research. Nvidia, despite its 50%+ market share in discrete desktop and mobile graphics, landed at only at 28.5% and ATI, now part of AMD, came in at 23.0%.
The $5.4 billion acquisition of ATI allows AMD to take aim at the chipset market, for the first time in the company’s history. First out the gate are two variants of the RS690 chipset that will try to conquer ground in the area of mainstream entertainment PCs as well as the very-low end of the market - sub-$500 PCs. Both versions - the higher-end 690G as well as the low-end 690V - are built in a 80 nm process by TSMC and incorporate the graphics engine of ATI’s Radeon X700 products, which are now called Radeon “X1200” series.
The chipsets will share core features – the capability to run Windows Vista Premium and all its eye candy as well as support for the Hypertransport bus and the Avivo engine, which, according to AMD, will bring increased image quality in video playback: Instead of displaying 16.7 million colors in standard processing, Avivo can pump out 1.07 billion (10-bit) colors, which results in smoother color transitions within an image. Both chipsets also share support for DirectX 9, Shader Model 2.0, acceleration for H.264 and VC-1 HD video, as well as the connected SB600 Southbridge with 10 USB and four SATA ports as well as support for PATA-based hard drives.
The 690V and its X1200 graphics chip does not offer much more beyond that, but AMD promises long-term support for the platform which should make the chipset an attractive alternative for business PCs, especially with AMD’s claim that 690-based motherboard/processor combinations will be about 10% cheaper than a comparable Intel platform consisting of a Core 2 Duo processor and a G965-based motherboard. We expect the 690V to be paired mainly with single-core Athlon 64 and Sempron processors.
For some extra cash, the 690G will offer enough horsepower for entry-level and mainstream entertainment center PCs. 690G boards will be based on slightly upgraded X1250 graphics and will be offering separate DVI/HDMI outputs as their main differentiator. AMD promises that 690G-based motherboards will provide more performance at lower cost when compared to Intel 965-based solutions.
What is interesting about the RS690 chipset is its positioning against Nvidia. Up until now, Nvidia had the most convincing chipset solution for AMD-based systems and AMD wasn’t shy to promote the benefits of Nforce chipsets for its systems. Apparently, this scenario is about to change. The product launch presentation puts the RS690 right against Nvidia’s MCP61 and AMD claims that its ATI-based device will outrun not only Intel, but also Nvidia in video quality, 3DMark05, Aquamark, Half-Life 2 and FarCry benchmarks.
AMD says it will keep its platforms open and the company still considers Nvidia a partner, but we do feel that this launch message may not go down too well with Nvidia. In fact, it would not be too surprising, if we saw Nvidia searching for a closer relationship with Intel very soon. It is the rather strange portion of this announcement, given the fact that AMD’s primary target is Intel at this time and Nvidia could be an ally who would be able to make AMD platforms even more compelling.
To be fair, let’s also talk about Intel’s near-future chipset plans. In the second quarter, the company will be rolling out its Crestline chipset for mobile computers and by that time, AMD will have a 690 part for notebooks as well. Our sources are telling us that Intel is heavily testing its “Bearlake” chipset for desktop computers, which will be named Q33 and Q35 when the platform debuts in Q3 of this year. The chipset will support DDR2-800 memory and FSB1333 and will launch with a few new processors supporting the chipset, which will also include the new Southbridge ICH9.
We have no word on performance yet, but we do know that Crestline and Bearlake will be supporting DirectX 10 and integrate very capable graphics engines. If AMD’s current numbers are close to reality, then it looks like the company will be able to claim performance and value leadership for one or two quarters, which may be enough to capture critical market shares. However, at least on the performance side, Intel appears to be well on track to reclaim that crown in Q2 and Q3 of this year.