| <html lang="en"> |
| <head> |
| <title>Rounding - The GNU C Library</title> |
| <meta http-equiv="Content-Type" content="text/html"> |
| <meta name="description" content="The GNU C Library"> |
| <meta name="generator" content="makeinfo 4.13"> |
| <link title="Top" rel="start" href="index.html#Top"> |
| <link rel="up" href="Arithmetic.html#Arithmetic" title="Arithmetic"> |
| <link rel="prev" href="Floating-Point-Errors.html#Floating-Point-Errors" title="Floating Point Errors"> |
| <link rel="next" href="Control-Functions.html#Control-Functions" title="Control Functions"> |
| <link href="http://www.gnu.org/software/texinfo/" rel="generator-home" title="Texinfo Homepage"> |
| <!-- |
| This file documents the GNU C library. |
| |
| This is Edition 0.12, last updated 2007-10-27, |
| of `The GNU C Library Reference Manual', for version |
| 2.8 (Sourcery G++ Lite 2011.03-41). |
| |
| Copyright (C) 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2001, 2002, |
| 2003, 2007, 2008, 2010 Free Software Foundation, Inc. |
| |
| Permission is granted to copy, distribute and/or modify this document |
| under the terms of the GNU Free Documentation License, Version 1.3 or |
| any later version published by the Free Software Foundation; with the |
| Invariant Sections being ``Free Software Needs Free Documentation'' |
| and ``GNU Lesser General Public License'', the Front-Cover texts being |
| ``A GNU Manual'', and with the Back-Cover Texts as in (a) below. A |
| copy of the license is included in the section entitled "GNU Free |
| Documentation License". |
| |
| (a) The FSF's Back-Cover Text is: ``You have the freedom to |
| copy and modify this GNU manual. Buying copies from the FSF |
| supports it in developing GNU and promoting software freedom.''--> |
| <meta http-equiv="Content-Style-Type" content="text/css"> |
| <style type="text/css"><!-- |
| pre.display { font-family:inherit } |
| pre.format { font-family:inherit } |
| pre.smalldisplay { font-family:inherit; font-size:smaller } |
| pre.smallformat { font-family:inherit; font-size:smaller } |
| pre.smallexample { font-size:smaller } |
| pre.smalllisp { font-size:smaller } |
| span.sc { font-variant:small-caps } |
| span.roman { font-family:serif; font-weight:normal; } |
| span.sansserif { font-family:sans-serif; font-weight:normal; } |
| --></style> |
| <link rel="stylesheet" type="text/css" href="../cs.css"> |
| </head> |
| <body> |
| <div class="node"> |
| <a name="Rounding"></a> |
| <p> |
| Next: <a rel="next" accesskey="n" href="Control-Functions.html#Control-Functions">Control Functions</a>, |
| Previous: <a rel="previous" accesskey="p" href="Floating-Point-Errors.html#Floating-Point-Errors">Floating Point Errors</a>, |
| Up: <a rel="up" accesskey="u" href="Arithmetic.html#Arithmetic">Arithmetic</a> |
| <hr> |
| </div> |
| |
| <h3 class="section">20.6 Rounding Modes</h3> |
| |
| <p>Floating-point calculations are carried out internally with extra |
| precision, and then rounded to fit into the destination type. This |
| ensures that results are as precise as the input data. IEEE 754<!-- /@w --> |
| defines four possible rounding modes: |
| |
| <dl> |
| <dt>Round to nearest.<dd>This is the default mode. It should be used unless there is a specific |
| need for one of the others. In this mode results are rounded to the |
| nearest representable value. If the result is midway between two |
| representable values, the even representable is chosen. <dfn>Even</dfn> here |
| means the lowest-order bit is zero. This rounding mode prevents |
| statistical bias and guarantees numeric stability: round-off errors in a |
| lengthy calculation will remain smaller than half of <code>FLT_EPSILON</code>. |
| |
| <!-- @item Round toward @math{+@infinity{}} --> |
| <br><dt>Round toward plus Infinity.<dd>All results are rounded to the smallest representable value |
| which is greater than the result. |
| |
| <!-- @item Round toward @math{-@infinity{}} --> |
| <br><dt>Round toward minus Infinity.<dd>All results are rounded to the largest representable value which is less |
| than the result. |
| |
| <br><dt>Round toward zero.<dd>All results are rounded to the largest representable value whose |
| magnitude is less than that of the result. In other words, if the |
| result is negative it is rounded up; if it is positive, it is rounded |
| down. |
| </dl> |
| |
| <p class="noindent"><samp><span class="file">fenv.h</span></samp> defines constants which you can use to refer to the |
| various rounding modes. Each one will be defined if and only if the FPU |
| supports the corresponding rounding mode. |
| |
| <dl> |
| <!-- fenv.h --> |
| <!-- ISO --> |
| <a name="index-FE_005fTONEAREST-2389"></a><dt><code>FE_TONEAREST</code><dd>Round to nearest. |
| |
| <!-- fenv.h --> |
| <!-- ISO --> |
| <p><a name="index-FE_005fUPWARD-2390"></a><br><dt><code>FE_UPWARD</code><dd>Round toward +&infin;. |
| |
| <!-- fenv.h --> |
| <!-- ISO --> |
| <p><a name="index-FE_005fDOWNWARD-2391"></a><br><dt><code>FE_DOWNWARD</code><dd>Round toward -&infin;. |
| |
| <!-- fenv.h --> |
| <!-- ISO --> |
| <p><a name="index-FE_005fTOWARDZERO-2392"></a><br><dt><code>FE_TOWARDZERO</code><dd>Round toward zero. |
| </dl> |
| |
| <p>Underflow is an unusual case. Normally, IEEE 754<!-- /@w --> floating point |
| numbers are always normalized (see <a href="Floating-Point-Concepts.html#Floating-Point-Concepts">Floating Point Concepts</a>). |
| Numbers smaller than 2^r (where r is the minimum exponent, |
| <code>FLT_MIN_RADIX-1</code> for <var>float</var>) cannot be represented as |
| normalized numbers. Rounding all such numbers to zero or 2^r |
| would cause some algorithms to fail at 0. Therefore, they are left in |
| denormalized form. That produces loss of precision, since some bits of |
| the mantissa are stolen to indicate the decimal point. |
| |
| <p>If a result is too small to be represented as a denormalized number, it |
| is rounded to zero. However, the sign of the result is preserved; if |
| the calculation was negative, the result is <dfn>negative zero</dfn>. |
| Negative zero can also result from some operations on infinity, such as |
| 4/-&infin;. Negative zero behaves identically to zero except |
| when the <code>copysign</code> or <code>signbit</code> functions are used to check |
| the sign bit directly. |
| |
| <p>At any time one of the above four rounding modes is selected. You can |
| find out which one with this function: |
| |
| <!-- fenv.h --> |
| <!-- ISO --> |
| <div class="defun"> |
| — Function: int <b>fegetround</b> (<var>void</var>)<var><a name="index-fegetround-2393"></a></var><br> |
| <blockquote><p>Returns the currently selected rounding mode, represented by one of the |
| values of the defined rounding mode macros. |
| </p></blockquote></div> |
| |
| <p class="noindent">To change the rounding mode, use this function: |
| |
| <!-- fenv.h --> |
| <!-- ISO --> |
| <div class="defun"> |
| — Function: int <b>fesetround</b> (<var>int round</var>)<var><a name="index-fesetround-2394"></a></var><br> |
| <blockquote><p>Changes the currently selected rounding mode to <var>round</var>. If |
| <var>round</var> does not correspond to one of the supported rounding modes |
| nothing is changed. <code>fesetround</code> returns zero if it changed the |
| rounding mode, a nonzero value if the mode is not supported. |
| </p></blockquote></div> |
| |
| <p>You should avoid changing the rounding mode if possible. It can be an |
| expensive operation; also, some hardware requires you to compile your |
| program differently for it to work. The resulting code may run slower. |
| See your compiler documentation for details. |
| <!-- This section used to claim that functions existed to round one number --> |
| <!-- in a specific fashion. I can't find any functions in the library --> |
| <!-- that do that. -zw --> |
| |
| </body></html> |
| |