Abstract
Time constraints and cost are the major limitations for developing optimized DSP based solutions for the consumer market. Exploiting the DSP architecture and software development tools with various optimization techniques can lead to the desired performance goals. Developing a highly optimized with more features in the shortest possible time gives an edge over the competitors. Optimization helps to cut the cost (processing power and memory) of an existing product or helps to introduce more features and more functionality with the same cost. A cost effective solution requires employing optimization not only in its constituent components but also in total. This paper presents optimization techniques for execution speed and memory in general and specific to Texas Instruments’ TMS320C6x series of DSP processors. Texas Instruments’ optimizing C compiler does most of the desired optimization in C. At the heart of the compiler based optimization techniques is tuning the C code to give maximum information about the code to the compiler enabling the compiler to do maximum optimization. An assembly based solution should be used only when the compiler is unable to exploit the full resources of the C6x CPU.

Author Rupesh W Kumbhare

Download this paper