ATI X1600XT Video Card Review :: X1600XT Features

• 16 Pixel Shader Processors
• 4 Pixel Pipelines
• 5 Vertex Shader Pipelines
• 590MHz Core Clock
• 1.38GHz memory clock
• 90 nanometer process
• 157 Million Transistors
• 256MB memory
• Vertex Shader 3.0
• Pixel Shader 3.0
• HDR
• 3Dc+
• Up to 6 sample Multisampling full scene Anti-Aliasing
• Adaptive Anti-Aliasing
• Temporal Anti-Aliasing
• Lossless Color Compression
• Up to 16x Anisotropic Filtering
• 4096x4096 textures
• Ring Bus Architecture
• AVIVO
• Dual 10-bit color per component Display Pipelines
• Dual Integrated Dual Link TMDS Transmitters
• 2x 400MHz RAMDAC

Wow where to begin. The X1600XT is derived from the X1K series of video cards. Unlike the RADEON X1800 family, the X1600XT is most similar in architecture to the X1900 family introduced recently. The RV530 chip uses the 90 nanometer low-k process and has over 157 million transistors. It's pretty amazing that the mid-range chip of the line has nearly the same transistor count as the last generation x800 series which was the high end of early last year.

The RV530 has an interesting pipeline arrangement. Most video cards today have the same number of texturing units as pixel Shader processors. The X1600XT has 12 pixel Shader processors, 4 texture units, 4 ROPs (Rendered Pixels), 8 Z-Units and 5 Vertex Shader processors. ATI believes that games will use increasingly more complicated pixel Shader programs and increasing the number of pixel Shader processors is more important than the number of pixels outputted. In any event, the X1600XT has a fillrate of 2690 megapixels per second.

Pixel Shader 3.0 is actually the 4th iteration of Microsoft's Pixel Shader versions. Pixel Shader 3.0 has the following improvements over 2.0: nearly unlimited Shader program lengths, fast dynamic branching, and 128-bit floating point precision (FP32) processing, Multiple Render Target support, a facing register for two-sided lighting, and shadow volume rendering acceleration. ATI feels that dynamic branching is a very important feature and have spent a lot of resources on the entire R5xx family to implement efficient dynamic branching.

Vertex Shader 3.0 differs from 2.0 in the support for 1024 static Shader instructions per pass versus 256, dynamic branching via looping branching and subroutines, Single Cycle Trigonometric Operations like SIN and COS, and vertex texturing. ATI supports Render to Vertex Buffer, but not the full Vertex Texture capabilities of Vertex Shader 3.0. The RV530 has 5 Vertex Shader processors, less than the 8 found on the X1800 or x1900 families, but more than the 4 found on the previous x700 series.

High Dynamic Range Rendering is made possible by floating point hardware such as the X1600XT from ATI. The human eye can distinguish between thousands of contrasts. The modern video card without FP hardware is limited to a 255:1 contrast ratio. With HDR and FP16 hardware, you can do a variety of effects. For example, in HL2 Lost Coast, HDR is used to render the effect of sunlight on the water, on the sky and even the rocks.

NVIDIA first introduced Transparency Anti-aliasing with the GeForce 7800GTX card. When ATI introduced the X1K series, they introduced their equivalent feature, Adaptive Anti-aliasing. Transparent textures aren't antialiased with normal Multisampling Anti-Aliasing. With AAA, Super-sampling is applied to transparent textures such as the interior of a chain link fence or dense vegetation.

Anisotropic Filtering quality is a big concern for me as a gamer. High quality AF has been missing from video cards since the GeForce Ti 4600 with their angle independent AF. The default setting for NVIDIA video cards is Quality mode. In this mode some games like Battlefield 2 exhibit "shimmering". ATI decided to up the bar with the X1K series, offering high quality AF as an option.

Discuss This Article