ATI’s HyperMemory Technology
At its most basic level, both HyperMemory and TurboCache are attempting to do the same thing. ATI’s version, called HyperMemory, utilizes a smaller on-board frame buffer and the fast PCI Express bus to take advantage of system memory for graphics rendering purposes.
This high-level overview (above) gives you the general idea of what both companies are attempting in the low price budget cards.
Using the north bridge of the core logic chipset that usually controls the memory (except for the case of AMD’s Athlon 64 processors), ATI’s X300 SE HyperMemory GPU can directly address system memory for frame buffer usage, just as the TurboCache technology from NVIDIA allows the 6200 TC cards to do the same.
But just because the two technologies are similar in their intentions, does not mean they are similar in their results. The real issue in our technical comparisons of these two technologies is that it takes more than just hardware to make all of this work. Since there is no industry standard for sharing system memory with a GPU over the PCI Express bus, there is actually more software intricacies that must be handled by the video driver than you might think.
How does the operating system know which memory is roped off for the GPU? When can the OS force the GPU to give it back? What happens to the data that was stored there? When can memory be safely allocated and deallocated from main memory? What data should be stored locally and what should be stored on the system memory to get better performance? All of these questions and more complex ones need to be answered by the driver engineers at both ATI and NVIDIA and that will ultimately determine the overall performance of these video cards.
