DX8 and DX9 performance
The following is a brief description of the DX9 performance characteristics.
Vertex Shader 2.0:
In DX9, the vertex shading engine adds limited flow control (loops, judgments, and subroutines) so that game designers can use more complex sequences of instructions, or more streamlined instructions to achieve effects that used to require very complex instructions to achieve. In addition to this, Vertex Shader 2.0 adds new registers, methods for setting constants, new macros, and more, and the maximum instruction length per shader has been increased to 256.
Pixel Shader 2.0:
The biggest improvement in Pixel Shader 2.0 over older versions is the 64-bit and 128-bit floating-point color precision support, which provides a dynamic range of 2 to the 128th power of 2 for each color channel. Pixel Shader 2.0 also increases the maximum instruction length to 96 (64 for algorithms and 32 for materials), as well as new registers and floating point data types. And programmers can now mix material addresses and algorithmic instructions when designing their own Pixel Shader.
Displacement Mapping
Displacement mapping is a Matrox technology that Microsoft has integrated into DirectX 9 as an industry standard. Hardware Displacement Mapping is a powerful and simple new way to compute 3D objects in real-time, to interpret and draw complex 3D geometries. Deep adaptation of vertex tessellation and vertex texturing techniques, along with reorganization of the underlying mesh and displacement mapping, creates a very realistic 3D graphic. The use of displacement mapping produces high-resolution scenes that take up a small amount of storage space. This has a very positive effect on reducing graphics program complexity, memory requirements, and data transfer credit bandwidth. Therefore, the hardware implementation of displacement mapping has been called a breakthrough technology.
Scissor Planes:
Before rendering an image, eliminating the parts that are unlikely to be seen by the user effectively reduces the time and bandwidth consumed in rendering. The designer doesn't have to deal with clipping on their own.
Line Atialiasing:
Prior to DX9, curves were often drawn as a number of straight lines, with jagged edges that were clearly visible at high resolutions, and the jagged edges were often made to look less jarring by fogging the edges (FSAA), but then the image would become blurry. In DX9, game designers can just draw smooth lines to get the best results without the FSAA blurring.
I believe we can see the performance advantages of DX9 from the introduction above - it can bring us a more realistic picture, which is one of the reasons why the two major display chips are fully supported. But to experience the high-quality effect of DX9 must be thought with a huge amount of data computing. And in the 500 yuan below the graphics card almost do not have the ability to support such a large amount of data computing, which is the reason why DX9 in the low-end is not favored by ATI. But if this kind of massive data computing can be partially undertaken by high-performance CPUs, DX9 is still very advantageous in the low-end market. So if you're looking at DX9 and you're ready to buy a lower-end card, you might want to consider adding a bit more budget to your CPU.
Considering that there are not many games that support DX9, and that graphics cards that do not support DX9 cannot run games that support DX9, it is possible to choose a low-end graphics card that does not support DX9 if you don't have high graphics quality requirements for the game.