by Anton Shilov
03/03/2004 | 05:34 AM
Meanwhile the world is waiting for CeBIT 2004 to see the new visual processing units showing up, leading developers of graphics chips do not really want to demonstrate the new technology during the show for a number of reasons.
Sources close to ATI Technologies and NVIDIA Corporation indicate inability of both companies to ship new graphics chips in volumes to partners right after the show in
Code-named NV40 and R420/R423 chips from NVIDIA and ATI are projected to deliver much higher performance as well as feature-set improvements compared to the RADEON 9800 XT and the GeForce FX 5950 Ultra that are selling now at $499 price-points. Given that more and more graphics cards based on the NV38 and R360 technology hit the market, it seems that the companies are generally satisfied with their sales
ATI’s new graphics technology may apparently see the light of the day in the very late-March to the very late-April timeframe. NVIDIA is a bit behind with its novelty and is likely to roll it out in mid-April – late-May timeframe, unofficial sources indicated.
In fact, ATI Technologies already has the new VPU functional, as last month the R420 was displayed at Intel Developer Forum.
While graphics companies investigate on the matter of PCI Express x16 ramp in order to be in a position to offer graphics cards for new slot, it does not seem that NVIDIA or ATI align schedules for the new generation architectures with chipset makers, such as VIA or Intel. NVIDIA officially said that it would ship its GeForce PCX processors for PEG x16 in the second half of the year; ATI remained tight-lipped on timeframes for its PCI Express VPUs, but is likely to align the availability of its PCI Express x16 lineup with availability of i915 (Grantsdale) mainboards.
Microsoft Corporation has not yet set the release date for the new version of DirectX 9 that should support additional functionality of the upcoming graphics processors, such as Pixel Shaders 3.0 and Vertex Shaders 3.0. But this is not something crucial, as historically, graphics companies launched their chips even if not all their capabilities could be utilized by available APIs.