HyperZ
HyperZ is the name of a collection of various techniques for improving the efficiency of graphics chips that ATI Technologies has developed and is used in the Radeon graphics chips.
HyperZ was introduced with the first Radeon, and ATI Technologies claimed that HyperZ allowed a 20% increase in rendering efficiency. The Radeon can deliver 1.5 GigaTexel / s fill rate, although theoretically it can only achieve 1.2 GigaTexel / s. Practical tests actually showed an increase in speed, so that the actually weaker Radeon could catch up with the GeForce 2 GTS .
With each new graphics core, ATI revised and improved these techniques. HyperZ II was introduced with the Radeon 8500 , HyperZ III + was used from the Radeon 9700 and the latest development stage, HyperZ HD , came onto the market with the Radeon X800.
With the GeForce 3, Nvidia introduced a similar technology called Lightspeed Memory Architecture (LMA).
functionality
HyperZ consists of three mechanisms:
Z Compression : The Z-buffer is saved with a lossless compression process in order to minimize the necessary data rate for reading or writing the Z-buffer.
Fast Z Clear : Instead of completely filling the Z-buffer with zeros when deleting and using the data rate for writing, an entire block of the Z-buffer is simply marked for deletion.
Hierarchical Z-Buffer : This technique allows the pixels to be rendered to be checked to see whether they are actually visible in the finished scene. These unnecessary pixels are then discarded before they burden the render pipeline.
Web links
- ATI HyperZ Glossary
- Beyond3D's explanation of HyperZ from Radeon 8500
- Tom's Hardware Guide pages matching "HyperZ"
swell
- ↑ Anandtech: ATI Radeon 64MB DDR , July 17, 2000