1 '\" te 2 .\" Copyright (c) 2007, Sun Microsystems, Inc. All Rights Reserved. 3 .\" The contents of this file are subject to the terms of the Common Development and Distribution License (the "License"). You may not use this file except in compliance with the License. 4 .\" You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE or http://www.opensolaris.org/os/licensing. See the License for the specific language governing permissions and limitations under the License. 5 .\" When distributing Covered Code, include this CDDL HEADER in each file and include the License file at usr/src/OPENSOLARIS.LICENSE. If applicable, add the following below this CDDL HEADER, with the fields enclosed by brackets "[]" replaced with your own identifying information: Portions Copyright [yyyy] [name of copyright owner] 6 .TH vrsqrt_ 3MVEC "14 Dec 2007" "SunOS 5.11" "Vector Math Library Functions" 7 .SH NAME 8 vrsqrt_, vrsqrtf_ \- vector reciprocal square root functions 9 .SH SYNOPSIS 10 .LP 11 .nf 12 cc [ \fIflag\fR\&.\|.\|. ] \fIfile\fR\&.\|.\|. \fB-lmvec\fR [ \fIlibrary\fR\&.\|.\|. ] 13 14 \fBvoid\fR \fBvrsqrt_\fR(\fBint *\fR\fIn\fR, \fBdouble * restrict\fR \fIx\fR, \fBint *\fR\fIstridex\fR, 15 \fBdouble * restrict\fR \fIy\fR, \fBint *\fR\fIstridey\fR); 16 .fi 17 18 .LP 19 .nf 20 \fBvoid\fR \fBvrsqrtf_\fR(\fBint *\fR\fIn\fR, \fBfloat * restrict\fR \fIx\fR, \fBint *\fR\fIstridex\fR, 21 \fBfloat * restrict\fR \fIy\fR, \fBint *\fR\fIstridey\fR); 22 .fi 23 24 .SH DESCRIPTION 25 .sp 26 .LP 27 These functions evaluate the function \fBrsqrt\fR(\fIx\fR), defined by 28 \fBrsqrt\fR(\fIx\fR) = 1 / \fBsqrt\fR(\fIx\fR), for an entire vector of values 29 at once. The first parameter specifies the number of values to compute. 30 Subsequent parameters specify the argument and result vectors. Each vector is 31 described by a pointer to the first element and a stride, which is the 32 increment between successive elements. 33 .sp 34 .LP 35 Specifically, \fBvrsqrt_\fR(\fIn\fR, \fIx\fR, \fIsx\fR, \fIy\fR, \fIsy\fR) 36 computes \fIy\fR[\fIi\fR * *\fIsy\fR] = \fBrsqrt\fR(\fIx\fR[\fIi\fR * 37 *\fIsx\fR]) for each \fIi\fR = 0, 1, ..., *\fIn\fR - 1. The \fBvrsqrtf_()\fR 38 function performs the same computation for single precision data. 39 .sp 40 .LP 41 These functions are not guaranteed to deliver results that are identical to the 42 results of evaluating 1.0 / \fBsqrt\fR(\fIx\fR) given the same arguments. 43 Non-exceptional results, however, are accurate to within a unit in the last 44 place. 45 .SH USAGE 46 .sp 47 .LP 48 The element count *\fIn\fR must be greater than zero. The strides for the 49 argument and result arrays can be arbitrary integers, but the arrays themselves 50 must not be the same or overlap. A zero stride effectively collapses an entire 51 vector into a single element. A negative stride causes a vector to be accessed 52 in descending memory order, but note that the corresponding pointer must still 53 point to the first element of the vector to be used; if the stride is negative, 54 this will be the highest-addressed element in memory. This convention differs 55 from the Level 1 BLAS, in which array parameters always refer to the 56 lowest-addressed element in memory even when negative increments are used. 57 .sp 58 .LP 59 These functions assume that the default round-to-nearest rounding direction 60 mode is in effect. On x86, these functions also assume that the default 61 round-to-64-bit rounding precision mode is in effect. The result of calling a 62 vector function with a non-default rounding mode in effect is undefined. 63 .sp 64 .LP 65 These functions handle special cases and exceptions in the spirit of IEEE 754. 66 In particular, 67 .RS +4 68 .TP 69 .ie t \(bu 70 .el o 71 if \fIx\fR < 0, \fBrsqrt\fR(\fIx\fR) is NaN, and an invalid operation exception 72 is raised, 73 .RE 74 .RS +4 75 .TP 76 .ie t \(bu 77 .el o 78 \fBrsqrt\fR(NaN) is NaN, 79 .RE 80 .RS +4 81 .TP 82 .ie t \(bu 83 .el o 84 \fBrsqrt\fR(+Inf) is +0, 85 .RE 86 .RS +4 87 .TP 88 .ie t \(bu 89 .el o 90 \fBrsqrt\fR(\(+-0) is \(+-Inf, and a division-by-zero exception is raised. 91 .RE 92 .sp 93 .LP 94 An application wanting to check for exceptions should call 95 \fBfeclearexcept\fR(\fBFE_ALL_EXCEPT\fR) before calling these functions. On 96 return, if \fBfetestexcept\fR(\fBFE_INVALID\fR | \fBFE_DIVBYZERO\fR | 97 \fBFE_OVERFLOW\fR | \fBFE_UNDERFLOW\fR) is non-zero, an exception has been 98 raised. The application can then examine the result or argument vectors for 99 exceptional values. Some vector functions can raise the inexact exception even 100 if all elements of the argument array are such that the numerical results are 101 exact. 102 .SH ATTRIBUTES 103 .sp 104 .LP 105 See \fBattributes\fR(5) for descriptions of the following attributes: 106 .sp 107 108 .sp 109 .TS 110 tab() box; 111 cw(2.75i) |cw(2.75i) 112 lw(2.75i) |lw(2.75i) 113 . 114 ATTRIBUTE TYPEATTRIBUTE VALUE 115 _ 116 Interface StabilityCommitted 117 _ 118 MT-LevelMT-Safe 119 .TE 120 121 .SH SEE ALSO 122 .sp 123 .LP 124 \fBsqrt\fR(3M), \fBfeclearexcept\fR(3M), \fBfetestexcept\fR(3M), 125 \fBattributes\fR(5)