mirror of https://github.com/mkerrisk/man-pages
137 lines
4.9 KiB
Plaintext
137 lines
4.9 KiB
Plaintext
.\" Copyright (c) 2001-2003 The Open Group, All Rights Reserved
|
|
.TH "FMA" P 2003 "IEEE/The Open Group" "POSIX Programmer's Manual"
|
|
.\" fma
|
|
.SH NAME
|
|
fma, fmaf, fmal \- floating-point multiply-add
|
|
.SH SYNOPSIS
|
|
.LP
|
|
\fB#include <math.h>
|
|
.br
|
|
.sp
|
|
double fma(double\fP \fIx\fP\fB, double\fP \fIy\fP\fB, double\fP \fIz\fP\fB);
|
|
.br
|
|
float fmaf(float\fP \fIx\fP\fB, float\fP \fIy\fP\fB, float\fP \fIz\fP\fB);
|
|
.br
|
|
long double fmal(long double\fP \fIx\fP\fB, long double\fP \fIy\fP\fB,
|
|
long double\fP \fIz\fP\fB);
|
|
.br
|
|
\fP
|
|
.SH DESCRIPTION
|
|
.LP
|
|
These functions shall compute (\fIx\fP\ *\ \fIy\fP)\ +\ \fIz\fP, rounded
|
|
as one ternary operation: they shall
|
|
compute the value (as if) to infinite precision and round once to
|
|
the result format, according to the rounding mode characterized
|
|
by the value of FLT_ROUNDS.
|
|
.LP
|
|
An application wishing to check for error situations should set \fIerrno\fP
|
|
to zero and call
|
|
\fIfeclearexcept\fP(FE_ALL_EXCEPT) before calling these functions.
|
|
On return, if \fIerrno\fP is non-zero or
|
|
\fIfetestexcept\fP(FE_INVALID | FE_DIVBYZERO | FE_OVERFLOW | FE_UNDERFLOW)
|
|
is non-zero, an error has occurred.
|
|
.SH RETURN VALUE
|
|
.LP
|
|
Upon successful completion, these functions shall return (\fIx\fP\ *\ \fIy\fP)\ +
|
|
\fIz\fP, rounded as one ternary
|
|
operation.
|
|
.LP
|
|
If
|
|
\fIx\fP or \fIy\fP are NaN, a NaN shall be returned.
|
|
.LP
|
|
If \fIx\fP multiplied by \fIy\fP is an exact infinity and \fIz\fP
|
|
is also an infinity but with the opposite sign, a domain
|
|
error shall occur, and either a NaN (if supported), or an implementation-defined
|
|
value shall be returned.
|
|
.LP
|
|
If one of \fIx\fP and \fIy\fP is infinite, the other is zero, and
|
|
\fIz\fP is not a NaN, a domain error shall occur, and
|
|
either a NaN (if supported), or an implementation-defined value shall
|
|
be returned.
|
|
.LP
|
|
If one of \fIx\fP and \fIy\fP is infinite, the other is zero, and
|
|
\fIz\fP is a NaN, a NaN shall be returned and a domain
|
|
error may occur.
|
|
.LP
|
|
If \fIx\fP* \fIy\fP is not 0*Inf nor Inf*0 and \fIz\fP is a NaN, a
|
|
NaN shall be returned.
|
|
.SH ERRORS
|
|
.LP
|
|
These functions shall fail if:
|
|
.TP 7
|
|
Domain\ Error
|
|
The value of \fIx\fP* \fIy\fP+ \fIz\fP is invalid, or the value \fIx\fP*
|
|
\fIy\fP is invalid and \fIz\fP is not a NaN.
|
|
.LP
|
|
If the integer expression (math_errhandling & MATH_ERRNO) is non-zero,
|
|
then \fIerrno\fP shall be set to [EDOM]. If the
|
|
integer expression (math_errhandling & MATH_ERREXCEPT) is non-zero,
|
|
then the invalid floating-point exception shall be raised.
|
|
.TP 7
|
|
Range\ Error
|
|
The result overflows.
|
|
.LP
|
|
If the integer expression (math_errhandling & MATH_ERRNO) is non-zero,
|
|
then \fIerrno\fP shall be set to [ERANGE]. If the
|
|
integer expression (math_errhandling & MATH_ERREXCEPT) is non-zero,
|
|
then the overflow floating-point exception shall be raised.
|
|
.sp
|
|
.LP
|
|
These functions may fail if:
|
|
.TP 7
|
|
Domain\ Error
|
|
The value \fIx\fP* \fIy\fP is invalid and \fIz\fP is a NaN.
|
|
.LP
|
|
If the integer expression (math_errhandling & MATH_ERRNO) is non-zero,
|
|
then \fIerrno\fP shall be set to [EDOM]. If the
|
|
integer expression (math_errhandling & MATH_ERREXCEPT) is non-zero,
|
|
then the invalid floating-point exception shall be raised.
|
|
.TP 7
|
|
Range\ Error
|
|
The result underflows.
|
|
.LP
|
|
If the integer expression (math_errhandling & MATH_ERRNO) is non-zero,
|
|
then \fIerrno\fP shall be set to [ERANGE]. If the
|
|
integer expression (math_errhandling & MATH_ERREXCEPT) is non-zero,
|
|
then the underflow floating-point exception shall be
|
|
raised.
|
|
.sp
|
|
.LP
|
|
\fIThe following sections are informative.\fP
|
|
.SH EXAMPLES
|
|
.LP
|
|
None.
|
|
.SH APPLICATION USAGE
|
|
.LP
|
|
On error, the expressions (math_errhandling & MATH_ERRNO) and (math_errhandling
|
|
& MATH_ERREXCEPT) are independent of
|
|
each other, but at least one of them must be non-zero.
|
|
.SH RATIONALE
|
|
.LP
|
|
In many cases, clever use of floating (\fIfused\fP) multiply-add leads
|
|
to much improved code; but its unexpected use by the
|
|
compiler can undermine carefully written code. The FP_CONTRACT macro
|
|
can be used to disallow use of floating multiply-add; and the
|
|
\fIfma\fP() function guarantees its use where desired. Many current
|
|
machines provide hardware floating multiply-add instructions;
|
|
software implementation can be used for others.
|
|
.SH FUTURE DIRECTIONS
|
|
.LP
|
|
None.
|
|
.SH SEE ALSO
|
|
.LP
|
|
\fIfeclearexcept\fP() , \fIfetestexcept\fP() , the Base
|
|
Definitions volume of IEEE\ Std\ 1003.1-2001, Section 4.18, Treatment
|
|
of
|
|
Error Conditions for Mathematical Functions, \fI<math.h>\fP
|
|
.SH COPYRIGHT
|
|
Portions of this text are reprinted and reproduced in electronic form
|
|
from IEEE Std 1003.1, 2003 Edition, Standard for Information Technology
|
|
-- Portable Operating System Interface (POSIX), The Open Group Base
|
|
Specifications Issue 6, Copyright (C) 2001-2003 by the Institute of
|
|
Electrical and Electronics Engineers, Inc and The Open Group. In the
|
|
event of any discrepancy between this version and the original IEEE and
|
|
The Open Group Standard, the original IEEE and The Open Group Standard
|
|
is the referee document. The original Standard can be obtained online at
|
|
http://www.opengroup.org/unix/online.html .
|