Over the past ten years, we have have seen GPU’s go from the fixed-function rendering pipeline (using OpenGL, or Direct3D programming API’s) to programmable shader pipelines (using GLSL, or HLSL) and in the recent years we have seen the introduction of general purpose programming on the GPU. In order to allow your program to scale well when more processors are added to the chip, we must take advantage of multi-threaded programming techniques. A program that runs in a single tread (single threaded, or sequential programs) will only benefit from a performance improvement when the single core of a multi-core processor is improved. This trend towards processor parallelization greatly impacts the kind of software programmers need to write in order to take advantage of multi-core CPU’s and GPU’s.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |