定点。浮点DSP区别(国外英语).docVIP

定点。浮点DSP区别(国外英语资料) Since the launch of the floating point digital signal processor (DSP) more than a decade ago, it has provided a more sophisticated alternative to real-time signal processing. However, fixed-point devices are still the mainstay of the industry today, and low cost is the main reason. Fixed-point DSP, which has a lower price per device, is an important advantage for mass mass-market applications. In comparison, floating point DSP can realize more quick and easy development, so the development cost is more important than the unit manufacturing cost for small-scale applications, but also the best choice. In recent years, high density integration and support improvements have made both DSP relatively close to both convenience and cost. For now, the choice of device types is increasingly dependent on whether the application data set requires more computing power in floating-point format. As a result, developers who design mass-production signal processing applications are now discovering more intrinsic value in floating-point formats. They look beyond the traditional fixed-point DSP development model and explore the design opportunities presented by floating point DSP. Different digital formats The fundamental difference between fixed-point and floating-point DSP is that they differ in their digital representation of the data. Fixed-point hardware strictly performs integer operations, while floating point DSP supports integer operations and real arithmetic, which is standardized by scientific notation. The word length is 16 bit fixed-point DSP implementation (rovide) 64K precision, with the symbol integer range of -215 to 215-1. In contrast, floating point DSP divides the data path into two parts: one is the number of tails that can be used as an integer or the number of real Numbers, and the second is the exponent. In a 32-bit floating point DSP that supports the industry standard single-precision operation, the tail number is 24, with an index of 8. B