From e8d23916ace140efd3f62d480f0f5cd4481cb828 Mon Sep 17 00:00:00 2001 From: batt Date: Wed, 23 Feb 2011 18:20:08 +0000 Subject: [PATCH] Add new macro: INT_MULT. See commit for details. git-svn-id: https://src.develer.com/svnoss/bertos/trunk@4725 38d2e660-2303-0410-9eaa-f027e97ec537 --- bertos/cfg/macros.h | 36 ++++++++++++++++++++++++++++++++++++ 1 file changed, 36 insertions(+) diff --git a/bertos/cfg/macros.h b/bertos/cfg/macros.h index 17f9c676..073ad55f 100644 --- a/bertos/cfg/macros.h +++ b/bertos/cfg/macros.h @@ -196,6 +196,42 @@ */ #define DIV_ROUNDUP(dividend, divisor) (((dividend) + (divisor) - 1) / (divisor)) + +/** + * Perform a multiply between the integer \a a and the float constant \a f. + * + * This macro can be used in order to avoid floating point arithmetics + * in expressions like this: + * \code + * int a, b; + * a = b * 0.5579652750; + * \endcode + * + * This macro rounds the floating point constant to a fraction, + * usign (2 ^ prec) as the denominator. + * For instance, with prec = 8, the constant 0.5579652750 will be rounded to: + * (143 / 256) = 0.55859375 + * So, the former code will be transformed to: + * \code + * a = b * 143 / 256; + * \endcode + * + * Since the denominator is a power of 2, we rely on the compiler to optimize + * this to a right shift. + * So, when you have to multiply an integer by a float constant, this macro + * will not use the floating point arithmentics. + * The operation will be converted to a mul + shift, with a huge performance boost. + * + * \note \a f MUST be a constant in order gain performance benefits. + * + * \param a integer you want to multiply + * \param f floating point constant which you want to multply with \a a + * \param prec conversion precision, ranges from 1 to the number of bits in a long. + * The higher, the better the approximation of the float constant will be. + */ +#define INT_MULT(a, f, prec) (((a) * (long)((f) * (1 << (prec)) + 0.5)) >> (prec)) + + /** Round up \a x to an even multiple of the 2's power \a pad. */ #define ROUND_UP2(x, pad) (((x) + ((pad) - 1)) & ~((pad) - 1)) -- 2.25.1