论文部分内容阅读
针对计量测试技术等领域对数据运算精度高的要求,提出了一种基于定点DSP芯片的数据运算精度提高方法,该方法借用浮点运算底层操作的机理,根据运算数据的数值范围,通过引入规格化因子,动态调整运算数据格式,使得参与运算的数据控制在预知的数值范围之内,该方法可以在略微增加运算量的条件下,有效提高定点DSP运算结果的精度。
Aiming at the requirements of high precision of data computation in the fields of metrology and test technology, a data processing accuracy improvement method based on fixed-point DSP chip is proposed. This method borrows the mechanism of floating-point operation, and according to the numerical range of operation data, This method can adjust the arithmetic data format dynamically to control the data involved in the operation within the range of the predicted value. This method can effectively improve the accuracy of fixed-point DSP operation with a slight increase in the amount of computation.