Introduction
When NVIDIA were initially bringing their NV4x architecture to the market their CEO, Jen-Hsun Huang, stated that they would see a full line of top-to-bottom processors based on the NV4x architecture by the end of the year. Indeed, the mid-range 6600 boards, based on NV43, followed on from the initial high end 6800 boards, based on NV40, NV41, a dedicated 12 pipe 6800 part, came later and finally, towards the end of the year NVIDIA, completing the initial top-to-bottom line are the 6200 boards.
Due to the fact that graphics chips are getting increasingly complex, even at the low end, the old one-to-one mapping of chip to board level is getting increasingly blurred, and 6200 is no exception. So far NVIDIA have released two version of 6200 – the initial one as a standard discrete graphics board, and a later “TurboCache” versions that are able to utilise system memory in order to cut board costs by having less local memory. These two different 6200 solutions are based on different chips – in this article we’ll be taking a look at the first to be introduced, the GeForce 6200.
The Graphics Chip
We we mentioned, the 6200 series of boards are based on two chips - the more recent "TurboCache" versions use the later NV44 chip, but the standard 6200 actually uses a version of GeForce 6600's NV43 chip, dubbed NV43-V (Value):
Although the NV43 chip has 8 internal fragment (Pixel Shader) pipelines, the GeForce 6200 is designed to be an entry level part and as such as a reduced number of pipelines, down to 4 - we'll take a closer look at this later in the article. It would seem apparent that the 6200 configuration is primarily based on fall-out parts that don't operate in NV43's full 6600 configuration.
The NV43 chip is fully derived from the NV4x architecture and as such it inherits the same features, albeit at an expected lower performance, as the high end GeForce 6800 Series. Listed below are a few of the key 3D features that are present in the NV4x series:
- CineFX 3.0 Architecture
- Full DirectX9 Support
- DirectX9 Shader Model 3.0 Support
- Vertex Shader 3.0
- Pixel Shader 3.0
- Internal 128-bit Floating Point (FP32) Precisions
- Unlimited Shader Lengths
- Up to 16 textures per pass
- Support for FP16 Texture Formats with Filtering, FP32 without
- Intellisample 3.0
- Up to 4X Native Multi-sampling FSAA with rotated grid sampling
- Lossless color, texture, z-data compression
- Fast Z Clear
- Up to 16x Anisotropic Filtering
- UltraShadow Technology
- Advanced Display options
- Dedicated on-chip video processor
- nView Multi Display technology
- Digital Vibrance Control 3.0
Although NV43 has the all the 3D processing capabilities NV40 does, to match with the lower end NV44 (6200 TurboCache) the NV43 based 6200 has the 16-bit floating point blending and filtering capabilities and any Z compression bandwidth saving schemes disabled. For a more detailed look at the feature-set and architecture of the NV4x architecture take a look at our NV40 - GeForce 6800 Ultra preview.
