Due to the emergence of the Ageia physicx PPU Physical accelerator card, the NVIDIA-initiated GPU concept is broken, but due to insufficient support from vendors, it is not considered the future mainstream, with ATI RD600 and NVIDIA nforce 6801 chipset release (built in three pci-e slots), we were surprised to find that the two big display giants have been the future of dual-card interconnection technology into three cards joint work, one of the graphics card is mainly responsible for physical acceleration, below let us look at the 2007 Physical Acceleration graphics technology.
The birth of the physical accelerator card
The initial understanding of physical computing is the GDC2005 of the Ageia company, the introduction of the first multithreaded API physical processor PhysX renowned, and then due to the related game lag, and no more actual action. In 2006, with the support of ASUS hardware vendors and game vendors, more than 100 games of 60 gaming companies have supported Ageia hardware acceleration, laying the groundwork for today's development.
Advantages and disadvantages of physical accelerator cards
The advantage of the physical accelerator card is that the CPU can be freed up so that physical operations are completely referred to the physical accelerator card. In the face of the future of a large number of physical, CPU capacity is very limited, such as object collisions, particle collisions, liquid sputtering, dust operations and even simulated high-intensity weather change operations, only with the help of physical accelerator card can greatly improve the performance of the game. But the shortcomings of the physical accelerator card is also obvious, first use a single function of the physical accelerator card to waste a PCI or PCI-E slot, in addition to its single function, so that its sales price is not accepted by the vast number of consumers. That's exactly what I saw. Nvidia and ATI use the current mainstream graphics card as a physical accelerator card is easier for consumers to identify, the first increase of the physical accelerator card is a display and processing functions of the graphics card, and the use of ordinary graphics card can be, after all, the current large number of physical real-time operation has not been popular in each 3D game.
Graphics manufacturers release physical acceleration technology
The physical acceleration technology of AMD's graphics card is named as asymmetric physical processing technology. It requires at least two ATI graphics card support, one of the ATI graphics card as a physical processor, the other is used as a GPU, and also supports two different models of ATI graphics card work together, so that users can have full upgrade flexibility. Another kind of tri-card interconnection technology, called "Triple play Operation", where two ATI graphics cards work in the crossfire mode while the other a relatively weak ATI graphics card will be a dedicated physical processor. Only AMD's own chipsets now support asymmetric physical acceleration technologies, including AMD 580X, Aticrossfire Xpress 3200 and RD790. The Intel 975X and P965 chipsets that support Crossfire are still not supported. Nvidia's physical acceleration scheme is very similar to ATI, and is based on the Havok FX engine to achieve GPU physical acceleration. At the same time, two methods can be used to achieve physical processing, the first is SLI mode, the second one is three graphics interconnection combination.
What conditions can support the physical accelerator card
The physical acceleration technology of Nvidia and ATI is completely designed according to the API of the Havok FX, so only the games support the physical operations of Havok FX, and their physical acceleration technology can be used in this game, but Microsoft's Direct X is needed to support it. There are already a variety of games to support the Ageia Novodex API, but these games are not supported by Nvidia and ATI physical acceleration technology.
How to improve this situation?
This situation is expected to be improved after the introduction of DX10, with the support of the Microsoft API, no matter which physical engine has the space to play, I believe there will be more games to add more physical operations. In the future, the DX10 GPU will be a more general-purpose processor, that is, a lot of 3D processing to the CPU will be transferred to the GPU, because the powerful GPU is more suitable for these tasks than the CPU, which will greatly improve the performance of the game.
Summary: Today's dual-card interconnection has been developed to three cards interconnection, by the GPU responsible for rendering, physical acceleration card lifting speed, CPU management game AI settings, nvidia and ATI for us to outline a beautiful future of the game prospects, their physical acceleration program is obviously more civilian, more simple and practical, The problem ahead is the support of the game and the extent to which Microsoft DirectX supports it. And a long time will be combined, and long will be part of the hardware development will continue to deduce, and consumers only need to enjoy the high technology to bring pleasure.