diff options
| -rw-r--r-- | lib/libm/man/math.3 | 36 |
1 files changed, 13 insertions, 23 deletions
diff --git a/lib/libm/man/math.3 b/lib/libm/man/math.3 index f7337be2f4e..d3338ab3deb 100644 --- a/lib/libm/man/math.3 +++ b/lib/libm/man/math.3 @@ -1,4 +1,4 @@ -.\" $OpenBSD: math.3,v 1.24 2008/07/29 22:12:29 martynas Exp $ +.\" $OpenBSD: math.3,v 1.25 2010/02/20 20:37:51 schwarze Exp $ .\" Copyright (c) 1985 Regents of the University of California. .\" All rights reserved. .\" @@ -28,12 +28,9 @@ .\" .\" from: @(#)math.3 6.10 (Berkeley) 5/6/91 .\" -.if n \ -.ds Si sig. -.if t \ -.ds Si significant -.Dd $Mdocdate: July 29 2008 $ +.Dd $Mdocdate: February 20 2010 $ .Dt MATH 3 +.Os .Sh NAME .Nm math .Nd introduction to mathematical library functions @@ -167,7 +164,7 @@ Properties of D_floating-point: .It Radix: Binary. .It Precision: -56 \*(Si bits, roughly 17 \*(Si decimal digits. +56 significant bits, roughly 17 significant decimal digits. If x and x' are consecutive positive D_floating-point numbers (they differ by 1 \fIulp\fR), then .Li 1.3e-17 \*(Lt 0.5**56 \*(Lt (x'-x)/x \*(Le 0.5**55 \*(Lt 2.8e-17. @@ -330,7 +327,7 @@ and the Motorola 68881 has all the functions in .Em libm on chip, and is faster and more accurate to boot; -it, Apple, the i8087, Z8070 and WE32106 all use 64 \*(Si bits. +it, Apple, the i8087, Z8070 and WE32106 all use 64 significant bits. The main virtue of .Bx 4.3 's .Em libm @@ -352,7 +349,7 @@ Double-Precision: .It Radix: Binary. .It Precision: -53 \*(Si bits, roughly equivalent to 16 \*(Si decimals. +53 significant bits, roughly equivalent to 16 significant decimals. .br If x and x' are consecutive positive Double-Precision numbers (they differ by 1 \fIulp\fR, then @@ -403,14 +400,10 @@ comparison, rather than mere (in)equality, signal Invalid Operation when \*(Na is involved. .El .It Rounding: -Every algebraic operation (+, -, \(**, /, -.if n \ -sqrt) -.if t \ -\(sr) +Every algebraic operation (+, -, \(**, /, sqrt) is rounded by default to within half a \fIulp\fR, and when the rounding error is exactly half a \fIulp\fR then -the rounded value's least \*(Si bit is zero. +the rounded value's least significant bit is zero. This kind of rounding is usually the best kind, sometimes provably so. For instance, for every @@ -496,10 +489,11 @@ because they would have been rounded off anyway. So gradual underflows are usually \fIprovably\fR ignorable. The same cannot be said of underflows flushed to 0. .El -.Bl -tag -width XXX +.Pp At the option of an implementor conforming to .St -ieee754 , other ways to cope with exceptions may be provided: +.Bl -tag -width XXX .It 4) ABORT. This mechanism classifies an exception in @@ -605,7 +599,7 @@ Single-Precision: .It Radix: Binary. .It Precision: -24 \*(Si bits, roughly equivalent to 7 \*(Si decimals. +24 significant bits, roughly equivalent to 7 significant decimals. .br If x and x' are consecutive positive Double-Precision numbers (they differ by 1 \fIulp\fR, then @@ -656,14 +650,10 @@ comparison, rather than mere (in)equality, signal Invalid Operation when \*(Na is involved. .El .It Rounding: -Every algebraic operation (+, -, \(**, /, -.if n \ -sqrt) -.if t \ -\(sr) +Every algebraic operation (+, -, \(**, /, sqrt) is rounded by default to within half a \fIulp\fR, and when the rounding error is exactly half a \fIulp\fR then -the rounded value's least \*(Si bit is zero. +the rounded value's least significant bit is zero. This kind of rounding is usually the best kind, sometimes provably so. For instance, for every |