xref: /freebsd/share/man/man7/simd.7 (revision b5daf675efc746611c7cfcd1fa474b8905064c4b)
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 July 29, 2025
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 for 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 PPC64
54.It    bcmp              Ta    A       Ta        Ta    S1
55.It    bcopy             Ta    A       Ta    S   Ta    S      Ta    SV
56.It    bzero             Ta    A       Ta    S   Ta    S
57.It    div               Ta            Ta        Ta    S
58.It    index             Ta    A       Ta        Ta    S1
59.It    ldiv              Ta            Ta        Ta    S
60.It    lldiv             Ta            Ta        Ta    S
61.It    memchr            Ta    A       Ta        Ta    S1
62.It    memcmp            Ta    A       Ta    S   Ta    S1
63.It    memccpy           Ta    A       Ta        Ta    S1
64.It    memcpy            Ta    A       Ta    S   Ta    S      Ta    SV
65.It    memmove           Ta    A       Ta    S   Ta    S      Ta    SV
66.It    memrchr           Ta    A       Ta        Ta    S1
67.It    memset            Ta    A       Ta    S   Ta    S
68.It    rindex            Ta    A       Ta        Ta    S1
69.It    stpcpy            Ta    A       Ta        Ta    S1
70.It    stpncpy           Ta            Ta        Ta    S1
71.It    strcat            Ta    A       Ta        Ta    S1
72.It    strchr            Ta    A       Ta        Ta    S1
73.It    strchrnul         Ta    A       Ta        Ta    S1
74.It    strcmp            Ta    A       Ta    S   Ta    S1
75.It    strcpy            Ta    A       Ta        Ta    S1     Ta    S2
76.It    strcspn           Ta    S       Ta        Ta    S2
77.It    strlcat           Ta    A       Ta        Ta    S1
78.It    strlcpy           Ta    A       Ta        Ta    S1
79.It    strlen            Ta    A       Ta    S   Ta    S1
80.It    strncat           Ta    A       Ta        Ta    S1
81.It    strncmp           Ta    A       Ta    S   Ta    S1
82.It    strncpy           Ta            Ta        Ta    S1     Ta    S2
83.It    strnlen           Ta    A       Ta        Ta    S1
84.It    strrchr           Ta    A       Ta        Ta    S1
85.It    strpbrk           Ta    S       Ta        Ta    S2
86.It    strsep            Ta    S       Ta        Ta    S2
87.It    strspn            Ta    S       Ta        Ta    S2
88.It    timingsafe_bcmp   Ta    A       Ta        Ta    S1
89.It    timingsafe_memcmp Ta    S       Ta        Ta    S
90.El
91.Pp
92.Sy S Ns :\ scalar (non-SIMD),
93.Sy 1 Ns :\ amd64 baseline,
94.Sy 2 Ns :\ x86-64-v2
95or PowerPC\ 2.05,
96.Sy 3 Ns :\ x86-64-v3,
97.Sy 4 Ns :\ x86-64-v4,
98.Sy V Ns :\ PowerPC\ VSX,
99.Sy A Ns :\ Arm\ ASIMD (NEON).
100.
101.Sh ENVIRONMENT
102.Bl -tag
103.It Ev ARCHLEVEL
104On
105.Em amd64 ,
106controls the level of SIMD enhancements used.
107If this variable is set to an architecture level from the list below
108and that architecture level is supported by the processor, SIMD
109enhancements up to
110.Ev ARCHLEVEL
111are used.
112If
113.Ev ARCHLEVEL
114is unset, not recognised, or not supported by the processor, the highest
115level of SIMD enhancements supported by the processor is used.
116.Pp
117A suffix beginning with
118.Sq ":"
119or
120.Sq "+"
121in
122.Ev ARCHLEVEL
123is ignored and may be used for future extensions.
124The architecture level can be prefixed with a
125.Sq "!"
126character to force use of the requested architecture level, even if the
127processor does not advertise that it is supported.
128This usually causes applications to crash and should only be used for
129testing purposes or if architecture level detection yields incorrect
130results.
131.Pp
132The architecture levels follow the AMD64 SysV ABI supplement:
133.Bl -tag -width x86-64-v2
134.It Cm scalar
135scalar enhancements only (no SIMD)
136.It Cm baseline
137cmov, cx8, x87 FPU, fxsr, MMX, osfxsr, SSE, SSE2
138.It Cm x86-64-v2
139cx16, lahf/sahf, popcnt, SSE3, SSSE3, SSE4.1, SSE4.2
140.It Cm x86-64-v3
141AVX, AVX2, BMI1, BMI2, F16C, FMA, lzcnt, movbe, osxsave
142.It Cm x86-64-v4
143AVX-512F/BW/CD/DQ/VL
144.El
145.El
146.
147.Sh DIAGNOSTICS
148.Bl -diag
149.It "Illegal Instruction"
150Printed by
151.Xr sh 1
152if a command is terminated through delivery of a
153.Dv SIGILL
154signal, see
155.Xr signal 3 .
156.Pp
157Use of an unsupported architecture level was forced by setting
158.Ev ARCHLEVEL
159to a string beginning with a
160.Sq "!"
161character, causing a process to crash due to use of an unsupported
162instruction.
163Unset
164.Ev ARCHLEVEL ,
165remove the
166.Sq "!"
167prefix or select a supported architecture level.
168.Pp
169Message may also appear for unrelated reasons.
170.El
171.
172.Sh SEE ALSO
173.Xr string 3 ,
174.Xr arch 7
175.Rs
176.%A H. J. Lu
177.%A Michael Matz
178.%A Milind Girkar
179.%A Jan Hubi\[u010D]ka \" \(vc
180.%A Andreas Jaeger
181.%A Mark Mitchell
182.%B System V Application Binary Interface
183.%D May 23, 2023
184.%T AMD64 Architecture Processor Supplement
185.%O Version 1.0
186.Re
187.
188.Sh HISTORY
189Architecture-specific enhanced
190.Em libc
191functions were added starting
192with
193.Fx 2.0
194for
195.Cm i386 ,
196.Fx 6.0
197for
198.Cm arm ,
199.Fx 6.1
200for
201.Cm amd64 ,
202.Fx 11.0
203for
204.Cm aarch64 ,
205and
206.Fx 12.0
207for
208.Cm powerpc64 .
209SIMD-enhanced functions were first added with
210.Fx 13.0
211for
212.Cm powerpc64
213and with
214.Fx 14.1
215for
216.Cm amd64 .
217.Pp
218A
219.Nm
220manual page appeared in
221.Fx 14.1 .
222.
223.Sh AUTHOR
224.An Robert Clausecker Aq Mt fuz@FreeBSD.org
225.
226.Sh CAVEATS
227Other parts of
228.Fx
229such as cryptographic routines in the kernel or in
230OpenSSL may also use SIMD enhancements.
231These enhancements are not subject to the
232.Ev ARCHLEVEL
233variable and may have their own configuration
234mechanism.
235.
236.Sh BUGS
237Use of SIMD enhancements cannot be configured on powerpc64.
238