| 摘要 |
PURPOSE:To perform calculation with high accuracy with a comparatively few number of iteration, by performing iteractive calculation by assembling an error correction equation in a convergence calculating equation. CONSTITUTION:The initial values (a0, b0, and c0) of the argument and two constants of an elementary function to be found are inputted to computing elements 100, 101, and 103. Those computing elements perform n-times of sequen tial iterative calculation based on relative equations (ai+1=ai+ or -bi.deltai), (bi+1=bi+ or -ai.deltai), and (ci+1=ci+ or -cRi). The deltai is an arbitrary variation, and the cRi is the constant corresponding to the number of iteration, and they are stored in the memory 103 in advance. In a process of iteration, the value of the bi or the ci is inputted to a controller 104, and whether the calculation for (n+1)-times of the computing element are addition or subtraction is con trolled by a control signal (a) corresponding to the elementary function to be found. After n-times of iteration, arithmetic calculation of (dn=bn-m.an.cn) or (dn=cn+bn/an) is performed by a computing element 106, then, an elementary function value dn can be outputted.
|
