xref: /freebsd/share/man/man7/simd.7 (revision a559ef1ac85947df1bff82b867bbfc07a8085092)
1.\" Copyright (c) 2023 The FreeBSD Foundation
2.
3.\" This documentation was written by Robert Clausecker <fuz@FreeBSD.org>
4.\" under sponsorship from the FreeBSD Foundation.
5.
6.\" Redistribution and use in source and binary forms, with or without
7.\" modification, are permitted provided that the following conditions
8.\" are met:
9.\" 1. Redistributions of source code must retain the above copyright
10.\"    notice, this list of conditions and the following disclaimer.
11.\" 2. Redistributions in binary form must reproduce the above copyright
12.\"    notice, this list of conditions and the following disclaimer in the
13.\"    documentation and/or other materials provided with the distribution.
14.
15.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ''AS IS'' AND
16.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
17.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
18.\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
19.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
20.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
21.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
22.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
23.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
24.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
25.\" SUCH DAMAGE
26.
27.Dd August 23, 2023
28.Dt SIMD 7
29.Os
30.Sh NAME
31.Nm simd
32.Nd SIMD enhancements
33.
34.Sh DESCRIPTION
35On some architectures, the
36.Fx
37.Em libc
38provides enhanced implementations of commonly used functions, replacing
39the architecture-independent implementations used otherwise.
40Depending on architecture and function, an enhanced
41implementation of a function may either always be used or the
42.Em libc
43detects at runtime which SIMD instruction set extensions are
44supported and picks the most suitable implementation automatically.
45On
46.Cm amd64 ,
47the environment variable
48.Ev ARCHLEVEL
49can be used to override this mechanism.
50.Pp
51Enhanced functions are present in the following architectures:
52.Bl -column FUNCTION__ aarch64_ arm_ amd64_ i386_ ppc64_ -offset indent
53.It Em FUNCTION  Ta Em AARCH64 Ta Em ARM Ta Em AMD64  Ta Em I386 Ta Em PPC64
54.It    bcmp      Ta            Ta        Ta    S1     Ta    S
55.It    bcopy     Ta            Ta    S   Ta    S      Ta    S    Ta    SV
56.It    bzero     Ta            Ta    S   Ta    S      Ta    S
57.It    div       Ta            Ta        Ta    S      Ta    S
58.It    index     Ta    S       Ta        Ta    S1
59.It    ldiv      Ta            Ta        Ta    S      Ta    S
60.It    lldiv     Ta            Ta        Ta    S
61.It    memcmp    Ta            Ta    S   Ta    S1     Ta    S
62.It    memcpy    Ta    S       Ta    S   Ta    S      Ta    S    Ta    SV
63.It    memmove   Ta    S       Ta    S   Ta    S      Ta    S    Ta    SV
64.It    memset    Ta            Ta    S   Ta    S      Ta    S
65.It    rindex    Ta    S
66.It    stpcpy    Ta            Ta        Ta    S1
67.It    strcat    Ta            Ta        Ta    S      Ta    S
68.It    strchr    Ta    S       Ta        Ta    S1     Ta    S
69.It    strchrnul Ta            Ta        Ta    S1
70.It    strcmp    Ta            Ta    S   Ta    S      Ta    S
71.It    strcpy    Ta            Ta        Ta    S1     Ta    S    Ta    S2
72.It    strcspn   Ta            Ta        Ta    S2
73.It    strlen    Ta            Ta    S   Ta    S1
74.It    strncmp   Ta            Ta    S   Ta           Ta    S
75.It    strncpy   Ta            Ta        Ta           Ta         Ta    S2
76.It    strrchr   Ta    S       Ta        Ta           Ta    S
77.It    strspn    Ta            Ta        Ta    S2
78.It    swab      Ta            Ta        Ta           Ta    S
79.It    wcschr    Ta            Ta        Ta           Ta    S
80.It    wcscmp    Ta            Ta        Ta           Ta    S
81.It    wcslen    Ta            Ta        Ta           Ta    S
82.It    wmemchr   Ta            Ta        Ta           Ta    S
83.El
84.Pp
85.Sy S Ns :\ scalar (non-SIMD),
86.Sy 1 Ns :\ amd64 baseline,
87.Sy 2 Ns :\ x86-64-v2
88or PowerPC\ 2.05,
89.Sy 3 Ns :\ x86-64-v3,
90.Sy 4 Ns :\ x86-64-v4,
91.Sy V Ns :\ PowerPC\ VSX.
92.
93.Sh ENVIRONMENT
94.Bl -tag
95.It Ev ARCHLEVEL
96On
97.Em amd64 ,
98controls the level of SIMD enhancements used.
99If this variable is set to an architecture level from the list below
100and that architecture level is supported by the processor, SIMD
101enhancements up to
102.Ev ARCHLEVEL
103are used.
104If
105.Ev ARCHLEVEL
106is unset, not recognised, or not supported by the processor, the highest
107level of SIMD enhancements supported by the processor is used.
108.Pp
109A suffix beginning with
110.Sq ":"
111or
112.Sq "+"
113in
114.Ev ARCHLEVEL
115is ignored and may be used for future extensions.
116The architecture level can be prefixed with a
117.Sq "!"
118character to force use of the requested architecture level, even if the
119processor does not advertise that it is supported.
120This usually causes applications to crash and should only be used for
121testing purposes or if architecture level detection yields incorrect
122results.
123.Pp
124The architecture levels follow the AMD64 SysV ABI supplement:
125.Bl -tag -width x86-64-v2
126.It Cm scalar
127scalar enhancements only (no SIMD)
128.It Cm baseline
129cmov, cx8, x87 FPU, fxsr, MMX, osfxsr, SSE, SSE2
130.It Cm x86-64-v2
131cx16, lahf/sahf, popcnt, SSE3, SSSE3, SSE4.1, SSE4.2
132.It Cm x86-64-v3
133AVX, AVX2, BMI1, BMI2, F16C, FMA, lzcnt, movbe, osxsave
134.It Cm x86-64-v4
135AVX-512F/BW/CD/DQ/VL
136.El
137.El
138.
139.Sh DIAGNOSTICS
140.Bl -diag
141.It "Illegal Instruction"
142Printed by
143.Xr sh 1
144if a command is terminated through delivery of a
145.Dv SIGILL
146signal, see
147.Xr signal 3 .
148.Pp
149Use of an unsupported architecture level was forced by setting
150.Ev ARCHLEVEL
151to a string beginning with a
152.Sq "!"
153character, causing a process to crash due to use of an unsupported
154instruction.
155Unset
156.Ev ARCHLEVEL ,
157remove the
158.Sq "!"
159prefix or select a supported architecture level.
160.Pp
161Message may also appear for unrelated reasons.
162.El
163.
164.Sh SEE ALSO
165.Xr string 3 ,
166.Xr arch 7
167.Rs
168.%A H. J. Lu
169.%A Michael Matz
170.%A Milind Girkar
171.%A Jan Hubi\[u010D]ka \" \(vc
172.%A Andreas Jaeger
173.%A Mark Mitchell
174.%B System V Application Binary Interface
175.%D May 23, 2023
176.%T AMD64 Architecture Processor Supplement
177.%O Version 1.0
178.Re
179.
180.Sh HISTORY
181Architecture-specific enhanced
182.Em libc
183functions were added starting
184with
185.Fx 2.0
186for
187.Cm i386 ,
188.Fx 6.0
189for
190.Cm arm ,
191.Fx 6.1
192for
193.Cm amd64 ,
194.Fx 11.0
195for
196.Cm aarch64 ,
197and
198.Fx 12.0
199for
200.Cm powerpc64 .
201SIMD-enhanced functions were first added with
202.Fx 13.0
203for
204.Cm powerpc64
205and with
206.Fx 14.0
207for
208.Cm amd64 .
209.Pp
210A
211.Nm
212manual page appeared in
213.Fx 14.0 .
214.
215.Sh AUTHOR
216.An Robert Clausecker Aq Mt fuz@FreeBSD.org
217.
218.Sh CAVEATS
219Other parts of
220.Fx
221such as cryptographic routines in the kernel or in
222OpenSSL may also use SIMD enhancements.
223These enhancements are not subject to the
224.Ev ARCHLEVEL
225variable and may have their own configuration
226mechanism.
227.
228.Sh BUGS
229Use of SIMD enhancements cannot be configured on powerpc64.
230