![]() ![]() ![]() (Sorry for all the mathematical-ness I wanted to explain properly )Īlso, one must take the number of pixel pipelines, vertex shaders and ROPs there are in the GPU, or in the case of 8000 series, the number of unified shaders present in the core, and the clock speed of the unified shaders.Īnd I thought I'd add. =22400 MB/s memory bandwidth or 22.4GB/s memory bandwidth So, for a card like the 7600GT with memory clocked at 700MHz (1400MHz DDR) and a 128-bit memory bus: That is because GPUs are structured like your CPU, the difference being that CPU’s are built to be Jack of all Trades in terms of what they can do. Cores on a GPU are very similar to cores on a CPU. Where X is memory clock and Y is memory bus: Answer (1 of 6): Core clock refers to the speed of the cores on the Graphics processor. Memory bandwidth is calculated by multiplying the DDR speed of the memory by the the memory bus divided by eight. This works out in the ratio of roughly 5.6 / 1 respectfully. Now the overclocked edition sits at 4900 / 875 respectfully. The higher the core clock, shader clock (if applicable) and memory bandwidth (clock is not as important as bandwidth, although it affects bandwidth), the better the GPU will be. The radeon 5770 memory / clock speeds sit at 4800 / 850 respectfully. Memory bus and memory speed (which are the deciding factors in the more important memory bandwidth) are much more important than the amount of memory. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |