URL
https://opencores.org/ocsvn/openrisc_me/openrisc_me/trunk
Subversion Repositories openrisc_me
[/] [openrisc/] [trunk/] [gnu-src/] [newlib-1.17.0/] [newlib/] [libm/] [mathfp/] [sf_sqrt.c] - Rev 158
Compare with Previous | Blame | View Log
/* @(#)z_sqrtf.c 1.0 98/08/13 */ /***************************************************************** * The following routines are coded directly from the algorithms * and coefficients given in "Software Manual for the Elementary * Functions" by William J. Cody, Jr. and William Waite, Prentice * Hall, 1980. *****************************************************************/ /****************************************************************** * Square Root * * Input: * x - floating point value * * Output: * square-root of x * * Description: * This routine performs floating point square root. * * The initial approximation is computed as * y0 = 0.41731 + 0.59016 * f * where f is a fraction such that x = f * 2^exp. * * Three Newton iterations in the form of Heron's formula * are then performed to obtain the final value: * y[i] = (y[i-1] + f / y[i-1]) / 2, i = 1, 2, 3. * *****************************************************************/ #include "fdlibm.h" #include "zmath.h" float _DEFUN (sqrtf, (float), float x) { float f, y; int exp, i, odd; /* Check for special values. */ switch (numtestf (x)) { case NAN: errno = EDOM; return (x); case INF: if (isposf (x)) { errno = EDOM; return (z_notanum_f.f); } else { errno = ERANGE; return (z_infinity_f.f); } } /* Initial checks are performed here. */ if (x == 0.0) return (0.0); if (x < 0) { errno = EDOM; return (z_notanum_f.f); } /* Find the exponent and mantissa for the form x = f * 2^exp. */ f = frexpf (x, &exp); odd = exp & 1; /* Get the initial approximation. */ y = 0.41731 + 0.59016 * f; f *= 0.5; /* Calculate the remaining iterations. */ for (i = 0; i < 2; ++i) y = y * 0.5 + f / y; /* Calculate the final value. */ if (odd) { y *= __SQRT_HALF; exp++; } exp >>= 1; y = ldexpf (y, exp); return (y); } #ifdef _DOUBLE_IS_32BITS double sqrt (double x) { return (double) sqrtf ((float) x); } #endif /* _DOUBLE_IS_32BITS */