The Methods to Realize Float Operation in the ADSP2106X Systems

摘要

Adopting the the processors that have high arithmetic speed and support hardware float operation, it is a good choice to design microcomputer application systems that can satisfy the demands of high arithmetic speed, large dynamic data scope and high precision. ADSP2106X is a kind of microprocessor with these characteristics. This paper combines with the structure and the function characteristics of ADSP2106X, some questions needed to solve when realizing float operations in application systems and the corresponding methods such as the float formats, the change each other between fixed-point and the floating-point data format, the change each other between the different float formats, the storage of floating number, floating number transmission, floating number arithmetic, floating-point operation abnormity handling and so on are discussed in detail, this has valuable reference to design computer application system with high performance for floating-point processing function..