利索-全球专利检索

  1. 登录
  2. 注册
发明公开
CN106991078A 一种高效的精度和耗时可调的TMS320C6678复数向量求模方法 审中-实审

一种高效的精度和耗时可调的TMS320C6678复数向量求模方法
摘要：
本发明涉及一种高效的精度和耗时可调的TMS320C6678复数向量求模方法，包括以下步骤：(a)根据具体系统精度要求选择合适的牛顿迭代次数T，初始化循环次数n＝0，记向量总长度为N，初始化输出y；(b)判断n＜N，如果n＜N则转至步骤(c)，否则转至(i)；(c)选取复数向量中第n个复数x(n)＝a+bi；(d)计算x(n)的模值平方S＝a2+b2；(e)通过TI提供的RSQRSP指令计算得到的近似值；(f)利用New‑Phapson公式对迭代T次得到P；(g)通过y(n)＝SP得到x(n)的模值y(n)；(h)n＝n+1，转至步骤(b)进行迭代；(i)输出结果y并停止。该方法能够根据系统对精度的要求，通过选择牛顿迭代次数T，可以得到满足系统精度要求的最少耗时的复数向量求模方法。
摘要（英）：

The invention relates to a highly-efficient accuracy and consumed time-adjustable TMS320C6678 complex number vector module calculating method comprising the steps of: (a), according to specific system accuracy requirements, selecting the proper number T of times of Newton iteration, initializing the number of times of circulation n=0, recording a total length of vectors as N, and initializing output y; (b), judging whether n+b<2> of x(n); (e), calculating an approximate value of 1/the square root of S according to an RSQRSP instruction provided by TI; (f), performing iteration on 1/the square root of S for T times by using a New-Phapson formula to obtain P; (g), obtaining the module value y(n) of x(n) via y(n)=SP; (h), n=n+1, turning to the step (b) for iteration; (1) outputting the result y and stopping. The method enables the number T of times of Newton iteration to be selected according to accuracy requirements of a system and thus can cost least time while meeting the accuracy requirements of the system.

0/0