You cannot select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
854 lines
61 KiB
HTML
854 lines
61 KiB
HTML
<html lang="en">
|
|
<head>
|
|
<title>x86 Options - Using the GNU Compiler Collection (GCC)</title>
|
|
<meta http-equiv="Content-Type" content="text/html">
|
|
<meta name="description" content="Using the GNU Compiler Collection (GCC)">
|
|
<meta name="generator" content="makeinfo 4.13">
|
|
<link title="Top" rel="start" href="index.html#Top">
|
|
<link rel="up" href="Submodel-Options.html#Submodel-Options" title="Submodel Options">
|
|
<link rel="prev" href="VxWorks-Options.html#VxWorks-Options" title="VxWorks Options">
|
|
<link rel="next" href="x86-Windows-Options.html#x86-Windows-Options" title="x86 Windows Options">
|
|
<link href="http://www.gnu.org/software/texinfo/" rel="generator-home" title="Texinfo Homepage">
|
|
<!--
|
|
Copyright (C) 1988-2015 Free Software Foundation, Inc.
|
|
|
|
Permission is granted to copy, distribute and/or modify this document
|
|
under the terms of the GNU Free Documentation License, Version 1.3 or
|
|
any later version published by the Free Software Foundation; with the
|
|
Invariant Sections being ``Funding Free Software'', the Front-Cover
|
|
Texts being (a) (see below), and with the Back-Cover Texts being (b)
|
|
(see below). A copy of the license is included in the section entitled
|
|
``GNU Free Documentation License''.
|
|
|
|
(a) The FSF's Front-Cover Text is:
|
|
|
|
A GNU Manual
|
|
|
|
(b) The FSF's Back-Cover Text is:
|
|
|
|
You have freedom to copy and modify this GNU Manual, like GNU
|
|
software. Copies published by the Free Software Foundation raise
|
|
funds for GNU development.-->
|
|
<meta http-equiv="Content-Style-Type" content="text/css">
|
|
<style type="text/css"><!--
|
|
pre.display { font-family:inherit }
|
|
pre.format { font-family:inherit }
|
|
pre.smalldisplay { font-family:inherit; font-size:smaller }
|
|
pre.smallformat { font-family:inherit; font-size:smaller }
|
|
pre.smallexample { font-size:smaller }
|
|
pre.smalllisp { font-size:smaller }
|
|
span.sc { font-variant:small-caps }
|
|
span.roman { font-family:serif; font-weight:normal; }
|
|
span.sansserif { font-family:sans-serif; font-weight:normal; }
|
|
--></style>
|
|
</head>
|
|
<body>
|
|
<div class="node">
|
|
<a name="x86-Options"></a>
|
|
<p>
|
|
Next: <a rel="next" accesskey="n" href="x86-Windows-Options.html#x86-Windows-Options">x86 Windows Options</a>,
|
|
Previous: <a rel="previous" accesskey="p" href="VxWorks-Options.html#VxWorks-Options">VxWorks Options</a>,
|
|
Up: <a rel="up" accesskey="u" href="Submodel-Options.html#Submodel-Options">Submodel Options</a>
|
|
<hr>
|
|
</div>
|
|
|
|
<h4 class="subsection">3.17.53 x86 Options</h4>
|
|
|
|
<p><a name="index-x86-Options-2655"></a>
|
|
These ‘<samp><span class="samp">-m</span></samp>’ options are defined for the x86 family of computers.
|
|
|
|
<dl>
|
|
<dt><code>-march=</code><var>cpu-type</var><dd><a name="index-march-2656"></a>Generate instructions for the machine type <var>cpu-type</var>. In contrast to
|
|
<samp><span class="option">-mtune=</span><var>cpu-type</var></samp>, which merely tunes the generated code
|
|
for the specified <var>cpu-type</var>, <samp><span class="option">-march=</span><var>cpu-type</var></samp> allows GCC
|
|
to generate code that may not run at all on processors other than the one
|
|
indicated. Specifying <samp><span class="option">-march=</span><var>cpu-type</var></samp> implies
|
|
<samp><span class="option">-mtune=</span><var>cpu-type</var></samp>.
|
|
|
|
<p>The choices for <var>cpu-type</var> are:
|
|
|
|
<dl>
|
|
<dt>‘<samp><span class="samp">native</span></samp>’<dd>This selects the CPU to generate code for at compilation time by determining
|
|
the processor type of the compiling machine. Using <samp><span class="option">-march=native</span></samp>
|
|
enables all instruction subsets supported by the local machine (hence
|
|
the result might not run on different machines). Using <samp><span class="option">-mtune=native</span></samp>
|
|
produces code optimized for the local machine under the constraints
|
|
of the selected instruction set.
|
|
|
|
<br><dt>‘<samp><span class="samp">i386</span></samp>’<dd>Original Intel i386 CPU.
|
|
|
|
<br><dt>‘<samp><span class="samp">i486</span></samp>’<dd>Intel i486 CPU. (No scheduling is implemented for this chip.)
|
|
|
|
<br><dt>‘<samp><span class="samp">i586</span></samp>’<dt>‘<samp><span class="samp">pentium</span></samp>’<dd>Intel Pentium CPU with no MMX support.
|
|
|
|
<br><dt>‘<samp><span class="samp">pentium-mmx</span></samp>’<dd>Intel Pentium MMX CPU, based on Pentium core with MMX instruction set support.
|
|
|
|
<br><dt>‘<samp><span class="samp">pentiumpro</span></samp>’<dd>Intel Pentium Pro CPU.
|
|
|
|
<br><dt>‘<samp><span class="samp">i686</span></samp>’<dd>When used with <samp><span class="option">-march</span></samp>, the Pentium Pro
|
|
instruction set is used, so the code runs on all i686 family chips.
|
|
When used with <samp><span class="option">-mtune</span></samp>, it has the same meaning as ‘<samp><span class="samp">generic</span></samp>’.
|
|
|
|
<br><dt>‘<samp><span class="samp">pentium2</span></samp>’<dd>Intel Pentium II CPU, based on Pentium Pro core with MMX instruction set
|
|
support.
|
|
|
|
<br><dt>‘<samp><span class="samp">pentium3</span></samp>’<dt>‘<samp><span class="samp">pentium3m</span></samp>’<dd>Intel Pentium III CPU, based on Pentium Pro core with MMX and SSE instruction
|
|
set support.
|
|
|
|
<br><dt>‘<samp><span class="samp">pentium-m</span></samp>’<dd>Intel Pentium M; low-power version of Intel Pentium III CPU
|
|
with MMX, SSE and SSE2 instruction set support. Used by Centrino notebooks.
|
|
|
|
<br><dt>‘<samp><span class="samp">pentium4</span></samp>’<dt>‘<samp><span class="samp">pentium4m</span></samp>’<dd>Intel Pentium 4 CPU with MMX, SSE and SSE2 instruction set support.
|
|
|
|
<br><dt>‘<samp><span class="samp">prescott</span></samp>’<dd>Improved version of Intel Pentium 4 CPU with MMX, SSE, SSE2 and SSE3 instruction
|
|
set support.
|
|
|
|
<br><dt>‘<samp><span class="samp">nocona</span></samp>’<dd>Improved version of Intel Pentium 4 CPU with 64-bit extensions, MMX, SSE,
|
|
SSE2 and SSE3 instruction set support.
|
|
|
|
<br><dt>‘<samp><span class="samp">core2</span></samp>’<dd>Intel Core 2 CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3 and SSSE3
|
|
instruction set support.
|
|
|
|
<br><dt>‘<samp><span class="samp">nehalem</span></samp>’<dd>Intel Nehalem CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3, SSSE3,
|
|
SSE4.1, SSE4.2 and POPCNT instruction set support.
|
|
|
|
<br><dt>‘<samp><span class="samp">westmere</span></samp>’<dd>Intel Westmere CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3, SSSE3,
|
|
SSE4.1, SSE4.2, POPCNT, AES and PCLMUL instruction set support.
|
|
|
|
<br><dt>‘<samp><span class="samp">sandybridge</span></samp>’<dd>Intel Sandy Bridge CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3, SSSE3,
|
|
SSE4.1, SSE4.2, POPCNT, AVX, AES and PCLMUL instruction set support.
|
|
|
|
<br><dt>‘<samp><span class="samp">ivybridge</span></samp>’<dd>Intel Ivy Bridge CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3, SSSE3,
|
|
SSE4.1, SSE4.2, POPCNT, AVX, AES, PCLMUL, FSGSBASE, RDRND and F16C
|
|
instruction set support.
|
|
|
|
<br><dt>‘<samp><span class="samp">haswell</span></samp>’<dd>Intel Haswell CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3, SSSE3,
|
|
SSE4.1, SSE4.2, POPCNT, AVX, AVX2, AES, PCLMUL, FSGSBASE, RDRND, FMA,
|
|
BMI, BMI2 and F16C instruction set support.
|
|
|
|
<br><dt>‘<samp><span class="samp">broadwell</span></samp>’<dd>Intel Broadwell CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3, SSSE3,
|
|
SSE4.1, SSE4.2, POPCNT, AVX, AVX2, AES, PCLMUL, FSGSBASE, RDRND, FMA,
|
|
BMI, BMI2, F16C, RDSEED, ADCX and PREFETCHW instruction set support.
|
|
|
|
<br><dt>‘<samp><span class="samp">bonnell</span></samp>’<dd>Intel Bonnell CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3 and SSSE3
|
|
instruction set support.
|
|
|
|
<br><dt>‘<samp><span class="samp">silvermont</span></samp>’<dd>Intel Silvermont CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3, SSSE3,
|
|
SSE4.1, SSE4.2, POPCNT, AES, PCLMUL and RDRND instruction set support.
|
|
|
|
<br><dt>‘<samp><span class="samp">knl</span></samp>’<dd>Intel Knight's Landing CPU with 64-bit extensions, MOVBE, MMX, SSE, SSE2, SSE3,
|
|
SSSE3, SSE4.1, SSE4.2, POPCNT, AVX, AVX2, AES, PCLMUL, FSGSBASE, RDRND, FMA,
|
|
BMI, BMI2, F16C, RDSEED, ADCX, PREFETCHW, AVX512F, AVX512PF, AVX512ER and
|
|
AVX512CD instruction set support.
|
|
|
|
<br><dt>‘<samp><span class="samp">k6</span></samp>’<dd>AMD K6 CPU with MMX instruction set support.
|
|
|
|
<br><dt>‘<samp><span class="samp">k6-2</span></samp>’<dt>‘<samp><span class="samp">k6-3</span></samp>’<dd>Improved versions of AMD K6 CPU with MMX and 3DNow! instruction set support.
|
|
|
|
<br><dt>‘<samp><span class="samp">athlon</span></samp>’<dt>‘<samp><span class="samp">athlon-tbird</span></samp>’<dd>AMD Athlon CPU with MMX, 3dNOW!, enhanced 3DNow! and SSE prefetch instructions
|
|
support.
|
|
|
|
<br><dt>‘<samp><span class="samp">athlon-4</span></samp>’<dt>‘<samp><span class="samp">athlon-xp</span></samp>’<dt>‘<samp><span class="samp">athlon-mp</span></samp>’<dd>Improved AMD Athlon CPU with MMX, 3DNow!, enhanced 3DNow! and full SSE
|
|
instruction set support.
|
|
|
|
<br><dt>‘<samp><span class="samp">k8</span></samp>’<dt>‘<samp><span class="samp">opteron</span></samp>’<dt>‘<samp><span class="samp">athlon64</span></samp>’<dt>‘<samp><span class="samp">athlon-fx</span></samp>’<dd>Processors based on the AMD K8 core with x86-64 instruction set support,
|
|
including the AMD Opteron, Athlon 64, and Athlon 64 FX processors.
|
|
(This supersets MMX, SSE, SSE2, 3DNow!, enhanced 3DNow! and 64-bit
|
|
instruction set extensions.)
|
|
|
|
<br><dt>‘<samp><span class="samp">k8-sse3</span></samp>’<dt>‘<samp><span class="samp">opteron-sse3</span></samp>’<dt>‘<samp><span class="samp">athlon64-sse3</span></samp>’<dd>Improved versions of AMD K8 cores with SSE3 instruction set support.
|
|
|
|
<br><dt>‘<samp><span class="samp">amdfam10</span></samp>’<dt>‘<samp><span class="samp">barcelona</span></samp>’<dd>CPUs based on AMD Family 10h cores with x86-64 instruction set support. (This
|
|
supersets MMX, SSE, SSE2, SSE3, SSE4A, 3DNow!, enhanced 3DNow!, ABM and 64-bit
|
|
instruction set extensions.)
|
|
|
|
<br><dt>‘<samp><span class="samp">bdver1</span></samp>’<dd>CPUs based on AMD Family 15h cores with x86-64 instruction set support. (This
|
|
supersets FMA4, AVX, XOP, LWP, AES, PCL_MUL, CX16, MMX, SSE, SSE2, SSE3, SSE4A,
|
|
SSSE3, SSE4.1, SSE4.2, ABM and 64-bit instruction set extensions.)
|
|
<br><dt>‘<samp><span class="samp">bdver2</span></samp>’<dd>AMD Family 15h core based CPUs with x86-64 instruction set support. (This
|
|
supersets BMI, TBM, F16C, FMA, FMA4, AVX, XOP, LWP, AES, PCL_MUL, CX16, MMX,
|
|
SSE, SSE2, SSE3, SSE4A, SSSE3, SSE4.1, SSE4.2, ABM and 64-bit instruction set
|
|
extensions.)
|
|
<br><dt>‘<samp><span class="samp">bdver3</span></samp>’<dd>AMD Family 15h core based CPUs with x86-64 instruction set support. (This
|
|
supersets BMI, TBM, F16C, FMA, FMA4, FSGSBASE, AVX, XOP, LWP, AES,
|
|
PCL_MUL, CX16, MMX, SSE, SSE2, SSE3, SSE4A, SSSE3, SSE4.1, SSE4.2, ABM and
|
|
64-bit instruction set extensions.
|
|
<br><dt>‘<samp><span class="samp">bdver4</span></samp>’<dd>AMD Family 15h core based CPUs with x86-64 instruction set support. (This
|
|
supersets BMI, BMI2, TBM, F16C, FMA, FMA4, FSGSBASE, AVX, AVX2, XOP, LWP,
|
|
AES, PCL_MUL, CX16, MOVBE, MMX, SSE, SSE2, SSE3, SSE4A, SSSE3, SSE4.1,
|
|
SSE4.2, ABM and 64-bit instruction set extensions.
|
|
|
|
<br><dt>‘<samp><span class="samp">btver1</span></samp>’<dd>CPUs based on AMD Family 14h cores with x86-64 instruction set support. (This
|
|
supersets MMX, SSE, SSE2, SSE3, SSSE3, SSE4A, CX16, ABM and 64-bit
|
|
instruction set extensions.)
|
|
|
|
<br><dt>‘<samp><span class="samp">btver2</span></samp>’<dd>CPUs based on AMD Family 16h cores with x86-64 instruction set support. This
|
|
includes MOVBE, F16C, BMI, AVX, PCL_MUL, AES, SSE4.2, SSE4.1, CX16, ABM,
|
|
SSE4A, SSSE3, SSE3, SSE2, SSE, MMX and 64-bit instruction set extensions.
|
|
|
|
<br><dt>‘<samp><span class="samp">winchip-c6</span></samp>’<dd>IDT WinChip C6 CPU, dealt in same way as i486 with additional MMX instruction
|
|
set support.
|
|
|
|
<br><dt>‘<samp><span class="samp">winchip2</span></samp>’<dd>IDT WinChip 2 CPU, dealt in same way as i486 with additional MMX and 3DNow!
|
|
instruction set support.
|
|
|
|
<br><dt>‘<samp><span class="samp">c3</span></samp>’<dd>VIA C3 CPU with MMX and 3DNow! instruction set support. (No scheduling is
|
|
implemented for this chip.)
|
|
|
|
<br><dt>‘<samp><span class="samp">c3-2</span></samp>’<dd>VIA C3-2 (Nehemiah/C5XL) CPU with MMX and SSE instruction set support.
|
|
(No scheduling is
|
|
implemented for this chip.)
|
|
|
|
<br><dt>‘<samp><span class="samp">geode</span></samp>’<dd>AMD Geode embedded processor with MMX and 3DNow! instruction set support.
|
|
</dl>
|
|
|
|
<br><dt><code>-mtune=</code><var>cpu-type</var><dd><a name="index-mtune-2657"></a>Tune to <var>cpu-type</var> everything applicable about the generated code, except
|
|
for the ABI and the set of available instructions.
|
|
While picking a specific <var>cpu-type</var> schedules things appropriately
|
|
for that particular chip, the compiler does not generate any code that
|
|
cannot run on the default machine type unless you use a
|
|
<samp><span class="option">-march=</span><var>cpu-type</var></samp> option.
|
|
For example, if GCC is configured for i686-pc-linux-gnu
|
|
then <samp><span class="option">-mtune=pentium4</span></samp> generates code that is tuned for Pentium 4
|
|
but still runs on i686 machines.
|
|
|
|
<p>The choices for <var>cpu-type</var> are the same as for <samp><span class="option">-march</span></samp>.
|
|
In addition, <samp><span class="option">-mtune</span></samp> supports 2 extra choices for <var>cpu-type</var>:
|
|
|
|
<dl>
|
|
<dt>‘<samp><span class="samp">generic</span></samp>’<dd>Produce code optimized for the most common IA32/AMD64/EM64T processors.
|
|
If you know the CPU on which your code will run, then you should use
|
|
the corresponding <samp><span class="option">-mtune</span></samp> or <samp><span class="option">-march</span></samp> option instead of
|
|
<samp><span class="option">-mtune=generic</span></samp>. But, if you do not know exactly what CPU users
|
|
of your application will have, then you should use this option.
|
|
|
|
<p>As new processors are deployed in the marketplace, the behavior of this
|
|
option will change. Therefore, if you upgrade to a newer version of
|
|
GCC, code generation controlled by this option will change to reflect
|
|
the processors
|
|
that are most common at the time that version of GCC is released.
|
|
|
|
<p>There is no <samp><span class="option">-march=generic</span></samp> option because <samp><span class="option">-march</span></samp>
|
|
indicates the instruction set the compiler can use, and there is no
|
|
generic instruction set applicable to all processors. In contrast,
|
|
<samp><span class="option">-mtune</span></samp> indicates the processor (or, in this case, collection of
|
|
processors) for which the code is optimized.
|
|
|
|
<br><dt>‘<samp><span class="samp">intel</span></samp>’<dd>Produce code optimized for the most current Intel processors, which are
|
|
Haswell and Silvermont for this version of GCC. If you know the CPU
|
|
on which your code will run, then you should use the corresponding
|
|
<samp><span class="option">-mtune</span></samp> or <samp><span class="option">-march</span></samp> option instead of <samp><span class="option">-mtune=intel</span></samp>.
|
|
But, if you want your application performs better on both Haswell and
|
|
Silvermont, then you should use this option.
|
|
|
|
<p>As new Intel processors are deployed in the marketplace, the behavior of
|
|
this option will change. Therefore, if you upgrade to a newer version of
|
|
GCC, code generation controlled by this option will change to reflect
|
|
the most current Intel processors at the time that version of GCC is
|
|
released.
|
|
|
|
<p>There is no <samp><span class="option">-march=intel</span></samp> option because <samp><span class="option">-march</span></samp> indicates
|
|
the instruction set the compiler can use, and there is no common
|
|
instruction set applicable to all processors. In contrast,
|
|
<samp><span class="option">-mtune</span></samp> indicates the processor (or, in this case, collection of
|
|
processors) for which the code is optimized.
|
|
</dl>
|
|
|
|
<br><dt><code>-mcpu=</code><var>cpu-type</var><dd><a name="index-mcpu-2658"></a>A deprecated synonym for <samp><span class="option">-mtune</span></samp>.
|
|
|
|
<br><dt><code>-mfpmath=</code><var>unit</var><dd><a name="index-mfpmath-2659"></a>Generate floating-point arithmetic for selected unit <var>unit</var>. The choices
|
|
for <var>unit</var> are:
|
|
|
|
<dl>
|
|
<dt>‘<samp><span class="samp">387</span></samp>’<dd>Use the standard 387 floating-point coprocessor present on the majority of chips and
|
|
emulated otherwise. Code compiled with this option runs almost everywhere.
|
|
The temporary results are computed in 80-bit precision instead of the precision
|
|
specified by the type, resulting in slightly different results compared to most
|
|
of other chips. See <samp><span class="option">-ffloat-store</span></samp> for more detailed description.
|
|
|
|
<p>This is the default choice for x86-32 targets.
|
|
|
|
<br><dt>‘<samp><span class="samp">sse</span></samp>’<dd>Use scalar floating-point instructions present in the SSE instruction set.
|
|
This instruction set is supported by Pentium III and newer chips,
|
|
and in the AMD line
|
|
by Athlon-4, Athlon XP and Athlon MP chips. The earlier version of the SSE
|
|
instruction set supports only single-precision arithmetic, thus the double and
|
|
extended-precision arithmetic are still done using 387. A later version, present
|
|
only in Pentium 4 and AMD x86-64 chips, supports double-precision
|
|
arithmetic too.
|
|
|
|
<p>For the x86-32 compiler, you must use <samp><span class="option">-march=</span><var>cpu-type</var></samp>, <samp><span class="option">-msse</span></samp>
|
|
or <samp><span class="option">-msse2</span></samp> switches to enable SSE extensions and make this option
|
|
effective. For the x86-64 compiler, these extensions are enabled by default.
|
|
|
|
<p>The resulting code should be considerably faster in the majority of cases and avoid
|
|
the numerical instability problems of 387 code, but may break some existing
|
|
code that expects temporaries to be 80 bits.
|
|
|
|
<p>This is the default choice for the x86-64 compiler.
|
|
|
|
<br><dt>‘<samp><span class="samp">sse,387</span></samp>’<dt>‘<samp><span class="samp">sse+387</span></samp>’<dt>‘<samp><span class="samp">both</span></samp>’<dd>Attempt to utilize both instruction sets at once. This effectively doubles the
|
|
amount of available registers, and on chips with separate execution units for
|
|
387 and SSE the execution resources too. Use this option with care, as it is
|
|
still experimental, because the GCC register allocator does not model separate
|
|
functional units well, resulting in unstable performance.
|
|
</dl>
|
|
|
|
<br><dt><code>-masm=</code><var>dialect</var><dd><a name="index-masm_003d_0040var_007bdialect_007d-2660"></a>Output assembly instructions using selected <var>dialect</var>. Also affects
|
|
which dialect is used for basic <code>asm</code> (see <a href="Basic-Asm.html#Basic-Asm">Basic Asm</a>) and
|
|
extended <code>asm</code> (see <a href="Extended-Asm.html#Extended-Asm">Extended Asm</a>). Supported choices (in dialect
|
|
order) are ‘<samp><span class="samp">att</span></samp>’ or ‘<samp><span class="samp">intel</span></samp>’. The default is ‘<samp><span class="samp">att</span></samp>’. Darwin does
|
|
not support ‘<samp><span class="samp">intel</span></samp>’.
|
|
|
|
<br><dt><code>-mieee-fp</code><dt><code>-mno-ieee-fp</code><dd><a name="index-mieee_002dfp-2661"></a><a name="index-mno_002dieee_002dfp-2662"></a>Control whether or not the compiler uses IEEE floating-point
|
|
comparisons. These correctly handle the case where the result of a
|
|
comparison is unordered.
|
|
|
|
<br><dt><code>-msoft-float</code><dd><a name="index-msoft_002dfloat-2663"></a>Generate output containing library calls for floating point.
|
|
|
|
<p><strong>Warning:</strong> the requisite libraries are not part of GCC.
|
|
Normally the facilities of the machine's usual C compiler are used, but
|
|
this can't be done directly in cross-compilation. You must make your
|
|
own arrangements to provide suitable library functions for
|
|
cross-compilation.
|
|
|
|
<p>On machines where a function returns floating-point results in the 80387
|
|
register stack, some floating-point opcodes may be emitted even if
|
|
<samp><span class="option">-msoft-float</span></samp> is used.
|
|
|
|
<br><dt><code>-mno-fp-ret-in-387</code><dd><a name="index-mno_002dfp_002dret_002din_002d387-2664"></a>Do not use the FPU registers for return values of functions.
|
|
|
|
<p>The usual calling convention has functions return values of types
|
|
<code>float</code> and <code>double</code> in an FPU register, even if there
|
|
is no FPU. The idea is that the operating system should emulate
|
|
an FPU.
|
|
|
|
<p>The option <samp><span class="option">-mno-fp-ret-in-387</span></samp> causes such values to be returned
|
|
in ordinary CPU registers instead.
|
|
|
|
<br><dt><code>-mno-fancy-math-387</code><dd><a name="index-mno_002dfancy_002dmath_002d387-2665"></a>Some 387 emulators do not support the <code>sin</code>, <code>cos</code> and
|
|
<code>sqrt</code> instructions for the 387. Specify this option to avoid
|
|
generating those instructions. This option is the default on
|
|
OpenBSD and NetBSD. This option is overridden when <samp><span class="option">-march</span></samp>
|
|
indicates that the target CPU always has an FPU and so the
|
|
instruction does not need emulation. These
|
|
instructions are not generated unless you also use the
|
|
<samp><span class="option">-funsafe-math-optimizations</span></samp> switch.
|
|
|
|
<br><dt><code>-malign-double</code><dt><code>-mno-align-double</code><dd><a name="index-malign_002ddouble-2666"></a><a name="index-mno_002dalign_002ddouble-2667"></a>Control whether GCC aligns <code>double</code>, <code>long double</code>, and
|
|
<code>long long</code> variables on a two-word boundary or a one-word
|
|
boundary. Aligning <code>double</code> variables on a two-word boundary
|
|
produces code that runs somewhat faster on a Pentium at the
|
|
expense of more memory.
|
|
|
|
<p>On x86-64, <samp><span class="option">-malign-double</span></samp> is enabled by default.
|
|
|
|
<p><strong>Warning:</strong> if you use the <samp><span class="option">-malign-double</span></samp> switch,
|
|
structures containing the above types are aligned differently than
|
|
the published application binary interface specifications for the x86-32
|
|
and are not binary compatible with structures in code compiled
|
|
without that switch.
|
|
|
|
<br><dt><code>-m96bit-long-double</code><dt><code>-m128bit-long-double</code><dd><a name="index-m96bit_002dlong_002ddouble-2668"></a><a name="index-m128bit_002dlong_002ddouble-2669"></a>These switches control the size of <code>long double</code> type. The x86-32
|
|
application binary interface specifies the size to be 96 bits,
|
|
so <samp><span class="option">-m96bit-long-double</span></samp> is the default in 32-bit mode.
|
|
|
|
<p>Modern architectures (Pentium and newer) prefer <code>long double</code>
|
|
to be aligned to an 8- or 16-byte boundary. In arrays or structures
|
|
conforming to the ABI, this is not possible. So specifying
|
|
<samp><span class="option">-m128bit-long-double</span></samp> aligns <code>long double</code>
|
|
to a 16-byte boundary by padding the <code>long double</code> with an additional
|
|
32-bit zero.
|
|
|
|
<p>In the x86-64 compiler, <samp><span class="option">-m128bit-long-double</span></samp> is the default choice as
|
|
its ABI specifies that <code>long double</code> is aligned on 16-byte boundary.
|
|
|
|
<p>Notice that neither of these options enable any extra precision over the x87
|
|
standard of 80 bits for a <code>long double</code>.
|
|
|
|
<p><strong>Warning:</strong> if you override the default value for your target ABI, this
|
|
changes the size of
|
|
structures and arrays containing <code>long double</code> variables,
|
|
as well as modifying the function calling convention for functions taking
|
|
<code>long double</code>. Hence they are not binary-compatible
|
|
with code compiled without that switch.
|
|
|
|
<br><dt><code>-mlong-double-64</code><dt><code>-mlong-double-80</code><dt><code>-mlong-double-128</code><dd><a name="index-mlong_002ddouble_002d64-2670"></a><a name="index-mlong_002ddouble_002d80-2671"></a><a name="index-mlong_002ddouble_002d128-2672"></a>These switches control the size of <code>long double</code> type. A size
|
|
of 64 bits makes the <code>long double</code> type equivalent to the <code>double</code>
|
|
type. This is the default for 32-bit Bionic C library. A size
|
|
of 128 bits makes the <code>long double</code> type equivalent to the
|
|
<code>__float128</code> type. This is the default for 64-bit Bionic C library.
|
|
|
|
<p><strong>Warning:</strong> if you override the default value for your target ABI, this
|
|
changes the size of
|
|
structures and arrays containing <code>long double</code> variables,
|
|
as well as modifying the function calling convention for functions taking
|
|
<code>long double</code>. Hence they are not binary-compatible
|
|
with code compiled without that switch.
|
|
|
|
<br><dt><code>-malign-data=</code><var>type</var><dd><a name="index-malign_002ddata-2673"></a>Control how GCC aligns variables. Supported values for <var>type</var> are
|
|
‘<samp><span class="samp">compat</span></samp>’ uses increased alignment value compatible uses GCC 4.8
|
|
and earlier, ‘<samp><span class="samp">abi</span></samp>’ uses alignment value as specified by the
|
|
psABI, and ‘<samp><span class="samp">cacheline</span></samp>’ uses increased alignment value to match
|
|
the cache line size. ‘<samp><span class="samp">compat</span></samp>’ is the default.
|
|
|
|
<br><dt><code>-mlarge-data-threshold=</code><var>threshold</var><dd><a name="index-mlarge_002ddata_002dthreshold-2674"></a>When <samp><span class="option">-mcmodel=medium</span></samp> is specified, data objects larger than
|
|
<var>threshold</var> are placed in the large data section. This value must be the
|
|
same across all objects linked into the binary, and defaults to 65535.
|
|
|
|
<br><dt><code>-mrtd</code><dd><a name="index-mrtd-2675"></a>Use a different function-calling convention, in which functions that
|
|
take a fixed number of arguments return with the <code>ret </code><var>num</var>
|
|
instruction, which pops their arguments while returning. This saves one
|
|
instruction in the caller since there is no need to pop the arguments
|
|
there.
|
|
|
|
<p>You can specify that an individual function is called with this calling
|
|
sequence with the function attribute <code>stdcall</code>. You can also
|
|
override the <samp><span class="option">-mrtd</span></samp> option by using the function attribute
|
|
<code>cdecl</code>. See <a href="Function-Attributes.html#Function-Attributes">Function Attributes</a>.
|
|
|
|
<p><strong>Warning:</strong> this calling convention is incompatible with the one
|
|
normally used on Unix, so you cannot use it if you need to call
|
|
libraries compiled with the Unix compiler.
|
|
|
|
<p>Also, you must provide function prototypes for all functions that
|
|
take variable numbers of arguments (including <code>printf</code>);
|
|
otherwise incorrect code is generated for calls to those
|
|
functions.
|
|
|
|
<p>In addition, seriously incorrect code results if you call a
|
|
function with too many arguments. (Normally, extra arguments are
|
|
harmlessly ignored.)
|
|
|
|
<br><dt><code>-mregparm=</code><var>num</var><dd><a name="index-mregparm-2676"></a>Control how many registers are used to pass integer arguments. By
|
|
default, no registers are used to pass arguments, and at most 3
|
|
registers can be used. You can control this behavior for a specific
|
|
function by using the function attribute <code>regparm</code>.
|
|
See <a href="Function-Attributes.html#Function-Attributes">Function Attributes</a>.
|
|
|
|
<p><strong>Warning:</strong> if you use this switch, and
|
|
<var>num</var> is nonzero, then you must build all modules with the same
|
|
value, including any libraries. This includes the system libraries and
|
|
startup modules.
|
|
|
|
<br><dt><code>-msseregparm</code><dd><a name="index-msseregparm-2677"></a>Use SSE register passing conventions for float and double arguments
|
|
and return values. You can control this behavior for a specific
|
|
function by using the function attribute <code>sseregparm</code>.
|
|
See <a href="Function-Attributes.html#Function-Attributes">Function Attributes</a>.
|
|
|
|
<p><strong>Warning:</strong> if you use this switch then you must build all
|
|
modules with the same value, including any libraries. This includes
|
|
the system libraries and startup modules.
|
|
|
|
<br><dt><code>-mvect8-ret-in-mem</code><dd><a name="index-mvect8_002dret_002din_002dmem-2678"></a>Return 8-byte vectors in memory instead of MMX registers. This is the
|
|
default on Solaris 8 and 9 and VxWorks to match the ABI of the Sun
|
|
Studio compilers until version 12. Later compiler versions (starting
|
|
with Studio 12 Update 1) follow the ABI used by other x86 targets, which
|
|
is the default on Solaris 10 and later. <em>Only</em> use this option if
|
|
you need to remain compatible with existing code produced by those
|
|
previous compiler versions or older versions of GCC.
|
|
|
|
<br><dt><code>-mpc32</code><dt><code>-mpc64</code><dt><code>-mpc80</code><dd><a name="index-mpc32-2679"></a><a name="index-mpc64-2680"></a><a name="index-mpc80-2681"></a>
|
|
Set 80387 floating-point precision to 32, 64 or 80 bits. When <samp><span class="option">-mpc32</span></samp>
|
|
is specified, the significands of results of floating-point operations are
|
|
rounded to 24 bits (single precision); <samp><span class="option">-mpc64</span></samp> rounds the
|
|
significands of results of floating-point operations to 53 bits (double
|
|
precision) and <samp><span class="option">-mpc80</span></samp> rounds the significands of results of
|
|
floating-point operations to 64 bits (extended double precision), which is
|
|
the default. When this option is used, floating-point operations in higher
|
|
precisions are not available to the programmer without setting the FPU
|
|
control word explicitly.
|
|
|
|
<p>Setting the rounding of floating-point operations to less than the default
|
|
80 bits can speed some programs by 2% or more. Note that some mathematical
|
|
libraries assume that extended-precision (80-bit) floating-point operations
|
|
are enabled by default; routines in such libraries could suffer significant
|
|
loss of accuracy, typically through so-called “catastrophic cancellation”,
|
|
when this option is used to set the precision to less than extended precision.
|
|
|
|
<br><dt><code>-mstackrealign</code><dd><a name="index-mstackrealign-2682"></a>Realign the stack at entry. On the x86, the <samp><span class="option">-mstackrealign</span></samp>
|
|
option generates an alternate prologue and epilogue that realigns the
|
|
run-time stack if necessary. This supports mixing legacy codes that keep
|
|
4-byte stack alignment with modern codes that keep 16-byte stack alignment for
|
|
SSE compatibility. See also the attribute <code>force_align_arg_pointer</code>,
|
|
applicable to individual functions.
|
|
|
|
<br><dt><code>-mpreferred-stack-boundary=</code><var>num</var><dd><a name="index-mpreferred_002dstack_002dboundary-2683"></a>Attempt to keep the stack boundary aligned to a 2 raised to <var>num</var>
|
|
byte boundary. If <samp><span class="option">-mpreferred-stack-boundary</span></samp> is not specified,
|
|
the default is 4 (16 bytes or 128 bits).
|
|
|
|
<p><strong>Warning:</strong> When generating code for the x86-64 architecture with
|
|
SSE extensions disabled, <samp><span class="option">-mpreferred-stack-boundary=3</span></samp> can be
|
|
used to keep the stack boundary aligned to 8 byte boundary. Since
|
|
x86-64 ABI require 16 byte stack alignment, this is ABI incompatible and
|
|
intended to be used in controlled environment where stack space is
|
|
important limitation. This option leads to wrong code when functions
|
|
compiled with 16 byte stack alignment (such as functions from a standard
|
|
library) are called with misaligned stack. In this case, SSE
|
|
instructions may lead to misaligned memory access traps. In addition,
|
|
variable arguments are handled incorrectly for 16 byte aligned
|
|
objects (including x87 long double and __int128), leading to wrong
|
|
results. You must build all modules with
|
|
<samp><span class="option">-mpreferred-stack-boundary=3</span></samp>, including any libraries. This
|
|
includes the system libraries and startup modules.
|
|
|
|
<br><dt><code>-mincoming-stack-boundary=</code><var>num</var><dd><a name="index-mincoming_002dstack_002dboundary-2684"></a>Assume the incoming stack is aligned to a 2 raised to <var>num</var> byte
|
|
boundary. If <samp><span class="option">-mincoming-stack-boundary</span></samp> is not specified,
|
|
the one specified by <samp><span class="option">-mpreferred-stack-boundary</span></samp> is used.
|
|
|
|
<p>On Pentium and Pentium Pro, <code>double</code> and <code>long double</code> values
|
|
should be aligned to an 8-byte boundary (see <samp><span class="option">-malign-double</span></samp>) or
|
|
suffer significant run time performance penalties. On Pentium III, the
|
|
Streaming SIMD Extension (SSE) data type <code>__m128</code> may not work
|
|
properly if it is not 16-byte aligned.
|
|
|
|
<p>To ensure proper alignment of this values on the stack, the stack boundary
|
|
must be as aligned as that required by any value stored on the stack.
|
|
Further, every function must be generated such that it keeps the stack
|
|
aligned. Thus calling a function compiled with a higher preferred
|
|
stack boundary from a function compiled with a lower preferred stack
|
|
boundary most likely misaligns the stack. It is recommended that
|
|
libraries that use callbacks always use the default setting.
|
|
|
|
<p>This extra alignment does consume extra stack space, and generally
|
|
increases code size. Code that is sensitive to stack space usage, such
|
|
as embedded systems and operating system kernels, may want to reduce the
|
|
preferred alignment to <samp><span class="option">-mpreferred-stack-boundary=2</span></samp>.
|
|
|
|
<br><dt><code>-mmmx</code><dd><a name="index-mmmx-2685"></a><dt><code>-msse</code><dd><a name="index-msse-2686"></a><dt><code>-msse2</code><dt><code>-msse3</code><dt><code>-mssse3</code><dt><code>-msse4</code><dt><code>-msse4a</code><dt><code>-msse4.1</code><dt><code>-msse4.2</code><dt><code>-mavx</code><dd><a name="index-mavx-2687"></a><dt><code>-mavx2</code><dt><code>-mavx512f</code><dt><code>-mavx512pf</code><dt><code>-mavx512er</code><dt><code>-mavx512cd</code><dt><code>-msha</code><dd><a name="index-msha-2688"></a><dt><code>-maes</code><dd><a name="index-maes-2689"></a><dt><code>-mpclmul</code><dd><a name="index-mpclmul-2690"></a><dt><code>-mclfushopt</code><dd><a name="index-mclfushopt-2691"></a><dt><code>-mfsgsbase</code><dd><a name="index-mfsgsbase-2692"></a><dt><code>-mrdrnd</code><dd><a name="index-mrdrnd-2693"></a><dt><code>-mf16c</code><dd><a name="index-mf16c-2694"></a><dt><code>-mfma</code><dd><a name="index-mfma-2695"></a><dt><code>-mfma4</code><dt><code>-mno-fma4</code><dt><code>-mprefetchwt1</code><dd><a name="index-mprefetchwt1-2696"></a><dt><code>-mxop</code><dd><a name="index-mxop-2697"></a><dt><code>-mlwp</code><dd><a name="index-mlwp-2698"></a><dt><code>-m3dnow</code><dd><a name="index-m3dnow-2699"></a><dt><code>-mpopcnt</code><dd><a name="index-mpopcnt-2700"></a><dt><code>-mabm</code><dd><a name="index-mabm-2701"></a><dt><code>-mbmi</code><dd><a name="index-mbmi-2702"></a><dt><code>-mbmi2</code><dt><code>-mlzcnt</code><dd><a name="index-mlzcnt-2703"></a><dt><code>-mfxsr</code><dd><a name="index-mfxsr-2704"></a><dt><code>-mxsave</code><dd><a name="index-mxsave-2705"></a><dt><code>-mxsaveopt</code><dd><a name="index-mxsaveopt-2706"></a><dt><code>-mxsavec</code><dd><a name="index-mxsavec-2707"></a><dt><code>-mxsaves</code><dd><a name="index-mxsaves-2708"></a><dt><code>-mrtm</code><dd><a name="index-mrtm-2709"></a><dt><code>-mtbm</code><dd><a name="index-mtbm-2710"></a><dt><code>-mmpx</code><dd><a name="index-mmpx-2711"></a><dt><code>-mmwaitx</code><dd><a name="index-mmwaitx-2712"></a>These switches enable the use of instructions in the MMX, SSE,
|
|
SSE2, SSE3, SSSE3, SSE4.1, AVX, AVX2, AVX512F, AVX512PF, AVX512ER, AVX512CD,
|
|
SHA, AES, PCLMUL, FSGSBASE, RDRND, F16C, FMA, SSE4A, FMA4, XOP, LWP, ABM,
|
|
BMI, BMI2, FXSR, XSAVE, XSAVEOPT, LZCNT, RTM, MPX, MWAITX or 3DNow!
|
|
extended instruction sets. Each has a corresponding <samp><span class="option">-mno-</span></samp> option
|
|
to disable use of these instructions.
|
|
|
|
<p>These extensions are also available as built-in functions: see
|
|
<a href="x86-Built_002din-Functions.html#x86-Built_002din-Functions">x86 Built-in Functions</a>, for details of the functions enabled and
|
|
disabled by these switches.
|
|
|
|
<p>To generate SSE/SSE2 instructions automatically from floating-point
|
|
code (as opposed to 387 instructions), see <samp><span class="option">-mfpmath=sse</span></samp>.
|
|
|
|
<p>GCC depresses SSEx instructions when <samp><span class="option">-mavx</span></samp> is used. Instead, it
|
|
generates new AVX instructions or AVX equivalence for all SSEx instructions
|
|
when needed.
|
|
|
|
<p>These options enable GCC to use these extended instructions in
|
|
generated code, even without <samp><span class="option">-mfpmath=sse</span></samp>. Applications that
|
|
perform run-time CPU detection must compile separate files for each
|
|
supported architecture, using the appropriate flags. In particular,
|
|
the file containing the CPU detection code should be compiled without
|
|
these options.
|
|
|
|
<br><dt><code>-mdump-tune-features</code><dd><a name="index-mdump_002dtune_002dfeatures-2713"></a>This option instructs GCC to dump the names of the x86 performance
|
|
tuning features and default settings. The names can be used in
|
|
<samp><span class="option">-mtune-ctrl=</span><var>feature-list</var></samp>.
|
|
|
|
<br><dt><code>-mtune-ctrl=</code><var>feature-list</var><dd><a name="index-mtune_002dctrl_003d_0040var_007bfeature_002dlist_007d-2714"></a>This option is used to do fine grain control of x86 code generation features.
|
|
<var>feature-list</var> is a comma separated list of <var>feature</var> names. See also
|
|
<samp><span class="option">-mdump-tune-features</span></samp>. When specified, the <var>feature</var> is turned
|
|
on if it is not preceded with ‘<samp><span class="samp">^</span></samp>’, otherwise, it is turned off.
|
|
<samp><span class="option">-mtune-ctrl=</span><var>feature-list</var></samp> is intended to be used by GCC
|
|
developers. Using it may lead to code paths not covered by testing and can
|
|
potentially result in compiler ICEs or runtime errors.
|
|
|
|
<br><dt><code>-mno-default</code><dd><a name="index-mno_002ddefault-2715"></a>This option instructs GCC to turn off all tunable features. See also
|
|
<samp><span class="option">-mtune-ctrl=</span><var>feature-list</var></samp> and <samp><span class="option">-mdump-tune-features</span></samp>.
|
|
|
|
<br><dt><code>-mcld</code><dd><a name="index-mcld-2716"></a>This option instructs GCC to emit a <code>cld</code> instruction in the prologue
|
|
of functions that use string instructions. String instructions depend on
|
|
the DF flag to select between autoincrement or autodecrement mode. While the
|
|
ABI specifies the DF flag to be cleared on function entry, some operating
|
|
systems violate this specification by not clearing the DF flag in their
|
|
exception dispatchers. The exception handler can be invoked with the DF flag
|
|
set, which leads to wrong direction mode when string instructions are used.
|
|
This option can be enabled by default on 32-bit x86 targets by configuring
|
|
GCC with the <samp><span class="option">--enable-cld</span></samp> configure option. Generation of <code>cld</code>
|
|
instructions can be suppressed with the <samp><span class="option">-mno-cld</span></samp> compiler option
|
|
in this case.
|
|
|
|
<br><dt><code>-mvzeroupper</code><dd><a name="index-mvzeroupper-2717"></a>This option instructs GCC to emit a <code>vzeroupper</code> instruction
|
|
before a transfer of control flow out of the function to minimize
|
|
the AVX to SSE transition penalty as well as remove unnecessary <code>zeroupper</code>
|
|
intrinsics.
|
|
|
|
<br><dt><code>-mprefer-avx128</code><dd><a name="index-mprefer_002davx128-2718"></a>This option instructs GCC to use 128-bit AVX instructions instead of
|
|
256-bit AVX instructions in the auto-vectorizer.
|
|
|
|
<br><dt><code>-mcx16</code><dd><a name="index-mcx16-2719"></a>This option enables GCC to generate <code>CMPXCHG16B</code> instructions.
|
|
<code>CMPXCHG16B</code> allows for atomic operations on 128-bit double quadword
|
|
(or oword) data types.
|
|
This is useful for high-resolution counters that can be updated
|
|
by multiple processors (or cores). This instruction is generated as part of
|
|
atomic built-in functions: see <a href="_005f_005fsync-Builtins.html#g_t_005f_005fsync-Builtins">__sync Builtins</a> or
|
|
<a href="_005f_005fatomic-Builtins.html#g_t_005f_005fatomic-Builtins">__atomic Builtins</a> for details.
|
|
|
|
<br><dt><code>-msahf</code><dd><a name="index-msahf-2720"></a>This option enables generation of <code>SAHF</code> instructions in 64-bit code.
|
|
Early Intel Pentium 4 CPUs with Intel 64 support,
|
|
prior to the introduction of Pentium 4 G1 step in December 2005,
|
|
lacked the <code>LAHF</code> and <code>SAHF</code> instructions
|
|
which are supported by AMD64.
|
|
These are load and store instructions, respectively, for certain status flags.
|
|
In 64-bit mode, the <code>SAHF</code> instruction is used to optimize <code>fmod</code>,
|
|
<code>drem</code>, and <code>remainder</code> built-in functions;
|
|
see <a href="Other-Builtins.html#Other-Builtins">Other Builtins</a> for details.
|
|
|
|
<br><dt><code>-mmovbe</code><dd><a name="index-mmovbe-2721"></a>This option enables use of the <code>movbe</code> instruction to implement
|
|
<code>__builtin_bswap32</code> and <code>__builtin_bswap64</code>.
|
|
|
|
<br><dt><code>-mcrc32</code><dd><a name="index-mcrc32-2722"></a>This option enables built-in functions <code>__builtin_ia32_crc32qi</code>,
|
|
<code>__builtin_ia32_crc32hi</code>, <code>__builtin_ia32_crc32si</code> and
|
|
<code>__builtin_ia32_crc32di</code> to generate the <code>crc32</code> machine instruction.
|
|
|
|
<br><dt><code>-mrecip</code><dd><a name="index-mrecip-2723"></a>This option enables use of <code>RCPSS</code> and <code>RSQRTSS</code> instructions
|
|
(and their vectorized variants <code>RCPPS</code> and <code>RSQRTPS</code>)
|
|
with an additional Newton-Raphson step
|
|
to increase precision instead of <code>DIVSS</code> and <code>SQRTSS</code>
|
|
(and their vectorized
|
|
variants) for single-precision floating-point arguments. These instructions
|
|
are generated only when <samp><span class="option">-funsafe-math-optimizations</span></samp> is enabled
|
|
together with <samp><span class="option">-finite-math-only</span></samp> and <samp><span class="option">-fno-trapping-math</span></samp>.
|
|
Note that while the throughput of the sequence is higher than the throughput
|
|
of the non-reciprocal instruction, the precision of the sequence can be
|
|
decreased by up to 2 ulp (i.e. the inverse of 1.0 equals 0.99999994).
|
|
|
|
<p>Note that GCC implements <code>1.0f/sqrtf(</code><var>x</var><code>)</code> in terms of <code>RSQRTSS</code>
|
|
(or <code>RSQRTPS</code>) already with <samp><span class="option">-ffast-math</span></samp> (or the above option
|
|
combination), and doesn't need <samp><span class="option">-mrecip</span></samp>.
|
|
|
|
<p>Also note that GCC emits the above sequence with additional Newton-Raphson step
|
|
for vectorized single-float division and vectorized <code>sqrtf(</code><var>x</var><code>)</code>
|
|
already with <samp><span class="option">-ffast-math</span></samp> (or the above option combination), and
|
|
doesn't need <samp><span class="option">-mrecip</span></samp>.
|
|
|
|
<br><dt><code>-mrecip=</code><var>opt</var><dd><a name="index-mrecip_003dopt-2724"></a>This option controls which reciprocal estimate instructions
|
|
may be used. <var>opt</var> is a comma-separated list of options, which may
|
|
be preceded by a ‘<samp><span class="samp">!</span></samp>’ to invert the option:
|
|
|
|
<dl>
|
|
<dt>‘<samp><span class="samp">all</span></samp>’<dd>Enable all estimate instructions.
|
|
|
|
<br><dt>‘<samp><span class="samp">default</span></samp>’<dd>Enable the default instructions, equivalent to <samp><span class="option">-mrecip</span></samp>.
|
|
|
|
<br><dt>‘<samp><span class="samp">none</span></samp>’<dd>Disable all estimate instructions, equivalent to <samp><span class="option">-mno-recip</span></samp>.
|
|
|
|
<br><dt>‘<samp><span class="samp">div</span></samp>’<dd>Enable the approximation for scalar division.
|
|
|
|
<br><dt>‘<samp><span class="samp">vec-div</span></samp>’<dd>Enable the approximation for vectorized division.
|
|
|
|
<br><dt>‘<samp><span class="samp">sqrt</span></samp>’<dd>Enable the approximation for scalar square root.
|
|
|
|
<br><dt>‘<samp><span class="samp">vec-sqrt</span></samp>’<dd>Enable the approximation for vectorized square root.
|
|
</dl>
|
|
|
|
<p>So, for example, <samp><span class="option">-mrecip=all,!sqrt</span></samp> enables
|
|
all of the reciprocal approximations, except for square root.
|
|
|
|
<br><dt><code>-mveclibabi=</code><var>type</var><dd><a name="index-mveclibabi-2725"></a>Specifies the ABI type to use for vectorizing intrinsics using an
|
|
external library. Supported values for <var>type</var> are ‘<samp><span class="samp">svml</span></samp>’
|
|
for the Intel short
|
|
vector math library and ‘<samp><span class="samp">acml</span></samp>’ for the AMD math core library.
|
|
To use this option, both <samp><span class="option">-ftree-vectorize</span></samp> and
|
|
<samp><span class="option">-funsafe-math-optimizations</span></samp> have to be enabled, and an SVML or ACML
|
|
ABI-compatible library must be specified at link time.
|
|
|
|
<p>GCC currently emits calls to <code>vmldExp2</code>,
|
|
<code>vmldLn2</code>, <code>vmldLog102</code>, <code>vmldLog102</code>, <code>vmldPow2</code>,
|
|
<code>vmldTanh2</code>, <code>vmldTan2</code>, <code>vmldAtan2</code>, <code>vmldAtanh2</code>,
|
|
<code>vmldCbrt2</code>, <code>vmldSinh2</code>, <code>vmldSin2</code>, <code>vmldAsinh2</code>,
|
|
<code>vmldAsin2</code>, <code>vmldCosh2</code>, <code>vmldCos2</code>, <code>vmldAcosh2</code>,
|
|
<code>vmldAcos2</code>, <code>vmlsExp4</code>, <code>vmlsLn4</code>, <code>vmlsLog104</code>,
|
|
<code>vmlsLog104</code>, <code>vmlsPow4</code>, <code>vmlsTanh4</code>, <code>vmlsTan4</code>,
|
|
<code>vmlsAtan4</code>, <code>vmlsAtanh4</code>, <code>vmlsCbrt4</code>, <code>vmlsSinh4</code>,
|
|
<code>vmlsSin4</code>, <code>vmlsAsinh4</code>, <code>vmlsAsin4</code>, <code>vmlsCosh4</code>,
|
|
<code>vmlsCos4</code>, <code>vmlsAcosh4</code> and <code>vmlsAcos4</code> for corresponding
|
|
function type when <samp><span class="option">-mveclibabi=svml</span></samp> is used, and <code>__vrd2_sin</code>,
|
|
<code>__vrd2_cos</code>, <code>__vrd2_exp</code>, <code>__vrd2_log</code>, <code>__vrd2_log2</code>,
|
|
<code>__vrd2_log10</code>, <code>__vrs4_sinf</code>, <code>__vrs4_cosf</code>,
|
|
<code>__vrs4_expf</code>, <code>__vrs4_logf</code>, <code>__vrs4_log2f</code>,
|
|
<code>__vrs4_log10f</code> and <code>__vrs4_powf</code> for the corresponding function type
|
|
when <samp><span class="option">-mveclibabi=acml</span></samp> is used.
|
|
|
|
<br><dt><code>-mabi=</code><var>name</var><dd><a name="index-mabi-2726"></a>Generate code for the specified calling convention. Permissible values
|
|
are ‘<samp><span class="samp">sysv</span></samp>’ for the ABI used on GNU/Linux and other systems, and
|
|
‘<samp><span class="samp">ms</span></samp>’ for the Microsoft ABI. The default is to use the Microsoft
|
|
ABI when targeting Microsoft Windows and the SysV ABI on all other systems.
|
|
You can control this behavior for specific functions by
|
|
using the function attributes <code>ms_abi</code> and <code>sysv_abi</code>.
|
|
See <a href="Function-Attributes.html#Function-Attributes">Function Attributes</a>.
|
|
|
|
<br><dt><code>-mtls-dialect=</code><var>type</var><dd><a name="index-mtls_002ddialect-2727"></a>Generate code to access thread-local storage using the ‘<samp><span class="samp">gnu</span></samp>’ or
|
|
‘<samp><span class="samp">gnu2</span></samp>’ conventions. ‘<samp><span class="samp">gnu</span></samp>’ is the conservative default;
|
|
‘<samp><span class="samp">gnu2</span></samp>’ is more efficient, but it may add compile- and run-time
|
|
requirements that cannot be satisfied on all systems.
|
|
|
|
<br><dt><code>-mpush-args</code><dt><code>-mno-push-args</code><dd><a name="index-mpush_002dargs-2728"></a><a name="index-mno_002dpush_002dargs-2729"></a>Use PUSH operations to store outgoing parameters. This method is shorter
|
|
and usually equally fast as method using SUB/MOV operations and is enabled
|
|
by default. In some cases disabling it may improve performance because of
|
|
improved scheduling and reduced dependencies.
|
|
|
|
<br><dt><code>-maccumulate-outgoing-args</code><dd><a name="index-maccumulate_002doutgoing_002dargs-2730"></a>If enabled, the maximum amount of space required for outgoing arguments is
|
|
computed in the function prologue. This is faster on most modern CPUs
|
|
because of reduced dependencies, improved scheduling and reduced stack usage
|
|
when the preferred stack boundary is not equal to 2. The drawback is a notable
|
|
increase in code size. This switch implies <samp><span class="option">-mno-push-args</span></samp>.
|
|
|
|
<br><dt><code>-mthreads</code><dd><a name="index-mthreads-2731"></a>Support thread-safe exception handling on MinGW. Programs that rely
|
|
on thread-safe exception handling must compile and link all code with the
|
|
<samp><span class="option">-mthreads</span></samp> option. When compiling, <samp><span class="option">-mthreads</span></samp> defines
|
|
<samp><span class="option">-D_MT</span></samp>; when linking, it links in a special thread helper library
|
|
<samp><span class="option">-lmingwthrd</span></samp> which cleans up per-thread exception-handling data.
|
|
|
|
<br><dt><code>-mno-align-stringops</code><dd><a name="index-mno_002dalign_002dstringops-2732"></a>Do not align the destination of inlined string operations. This switch reduces
|
|
code size and improves performance in case the destination is already aligned,
|
|
but GCC doesn't know about it.
|
|
|
|
<br><dt><code>-minline-all-stringops</code><dd><a name="index-minline_002dall_002dstringops-2733"></a>By default GCC inlines string operations only when the destination is
|
|
known to be aligned to least a 4-byte boundary.
|
|
This enables more inlining and increases code
|
|
size, but may improve performance of code that depends on fast
|
|
<code>memcpy</code>, <code>strlen</code>,
|
|
and <code>memset</code> for short lengths.
|
|
|
|
<br><dt><code>-minline-stringops-dynamically</code><dd><a name="index-minline_002dstringops_002ddynamically-2734"></a>For string operations of unknown size, use run-time checks with
|
|
inline code for small blocks and a library call for large blocks.
|
|
|
|
<br><dt><code>-mstringop-strategy=</code><var>alg</var><dd><a name="index-mstringop_002dstrategy_003d_0040var_007balg_007d-2735"></a>Override the internal decision heuristic for the particular algorithm to use
|
|
for inlining string operations. The allowed values for <var>alg</var> are:
|
|
|
|
<dl>
|
|
<dt>‘<samp><span class="samp">rep_byte</span></samp>’<dt>‘<samp><span class="samp">rep_4byte</span></samp>’<dt>‘<samp><span class="samp">rep_8byte</span></samp>’<dd>Expand using i386 <code>rep</code> prefix of the specified size.
|
|
|
|
<br><dt>‘<samp><span class="samp">byte_loop</span></samp>’<dt>‘<samp><span class="samp">loop</span></samp>’<dt>‘<samp><span class="samp">unrolled_loop</span></samp>’<dd>Expand into an inline loop.
|
|
|
|
<br><dt>‘<samp><span class="samp">libcall</span></samp>’<dd>Always use a library call.
|
|
</dl>
|
|
|
|
<br><dt><code>-mmemcpy-strategy=</code><var>strategy</var><dd><a name="index-mmemcpy_002dstrategy_003d_0040var_007bstrategy_007d-2736"></a>Override the internal decision heuristic to decide if <code>__builtin_memcpy</code>
|
|
should be inlined and what inline algorithm to use when the expected size
|
|
of the copy operation is known. <var>strategy</var>
|
|
is a comma-separated list of <var>alg</var>:<var>max_size</var>:<var>dest_align</var> triplets.
|
|
<var>alg</var> is specified in <samp><span class="option">-mstringop-strategy</span></samp>, <var>max_size</var> specifies
|
|
the max byte size with which inline algorithm <var>alg</var> is allowed. For the last
|
|
triplet, the <var>max_size</var> must be <code>-1</code>. The <var>max_size</var> of the triplets
|
|
in the list must be specified in increasing order. The minimal byte size for
|
|
<var>alg</var> is <code>0</code> for the first triplet and <var>max_size</var><code> + 1</code> of the
|
|
preceding range.
|
|
|
|
<br><dt><code>-mmemset-strategy=</code><var>strategy</var><dd><a name="index-mmemset_002dstrategy_003d_0040var_007bstrategy_007d-2737"></a>The option is similar to <samp><span class="option">-mmemcpy-strategy=</span></samp> except that it is to control
|
|
<code>__builtin_memset</code> expansion.
|
|
|
|
<br><dt><code>-momit-leaf-frame-pointer</code><dd><a name="index-momit_002dleaf_002dframe_002dpointer-2738"></a>Don't keep the frame pointer in a register for leaf functions. This
|
|
avoids the instructions to save, set up, and restore frame pointers and
|
|
makes an extra register available in leaf functions. The option
|
|
<samp><span class="option">-fomit-leaf-frame-pointer</span></samp> removes the frame pointer for leaf functions,
|
|
which might make debugging harder.
|
|
|
|
<br><dt><code>-mtls-direct-seg-refs</code><dt><code>-mno-tls-direct-seg-refs</code><dd><a name="index-mtls_002ddirect_002dseg_002drefs-2739"></a>Controls whether TLS variables may be accessed with offsets from the
|
|
TLS segment register (<code>%gs</code> for 32-bit, <code>%fs</code> for 64-bit),
|
|
or whether the thread base pointer must be added. Whether or not this
|
|
is valid depends on the operating system, and whether it maps the
|
|
segment to cover the entire TLS area.
|
|
|
|
<p>For systems that use the GNU C Library, the default is on.
|
|
|
|
<br><dt><code>-msse2avx</code><dt><code>-mno-sse2avx</code><dd><a name="index-msse2avx-2740"></a>Specify that the assembler should encode SSE instructions with VEX
|
|
prefix. The option <samp><span class="option">-mavx</span></samp> turns this on by default.
|
|
|
|
<br><dt><code>-mfentry</code><dt><code>-mno-fentry</code><dd><a name="index-mfentry-2741"></a>If profiling is active (<samp><span class="option">-pg</span></samp>), put the profiling
|
|
counter call before the prologue.
|
|
Note: On x86 architectures the attribute <code>ms_hook_prologue</code>
|
|
isn't possible at the moment for <samp><span class="option">-mfentry</span></samp> and <samp><span class="option">-pg</span></samp>.
|
|
|
|
<br><dt><code>-mrecord-mcount</code><dt><code>-mno-record-mcount</code><dd><a name="index-mrecord_002dmcount-2742"></a>If profiling is active (<samp><span class="option">-pg</span></samp>), generate a __mcount_loc section
|
|
that contains pointers to each profiling call. This is useful for
|
|
automatically patching and out calls.
|
|
|
|
<br><dt><code>-mnop-mcount</code><dt><code>-mno-nop-mcount</code><dd><a name="index-mnop_002dmcount-2743"></a>If profiling is active (<samp><span class="option">-pg</span></samp>), generate the calls to
|
|
the profiling functions as nops. This is useful when they
|
|
should be patched in later dynamically. This is likely only
|
|
useful together with <samp><span class="option">-mrecord-mcount</span></samp>.
|
|
|
|
<br><dt><code>-mskip-rax-setup</code><dt><code>-mno-skip-rax-setup</code><dd><a name="index-mskip_002drax_002dsetup-2744"></a>When generating code for the x86-64 architecture with SSE extensions
|
|
disabled, <samp><span class="option">-mskip-rax-setup</span></samp> can be used to skip setting up RAX
|
|
register when there are no variable arguments passed in vector registers.
|
|
|
|
<p><strong>Warning:</strong> Since RAX register is used to avoid unnecessarily
|
|
saving vector registers on stack when passing variable arguments, the
|
|
impacts of this option are callees may waste some stack space,
|
|
misbehave or jump to a random location. GCC 4.4 or newer don't have
|
|
those issues, regardless the RAX register value.
|
|
|
|
<br><dt><code>-m8bit-idiv</code><dt><code>-mno-8bit-idiv</code><dd><a name="index-m8bit_002didiv-2745"></a>On some processors, like Intel Atom, 8-bit unsigned integer divide is
|
|
much faster than 32-bit/64-bit integer divide. This option generates a
|
|
run-time check. If both dividend and divisor are within range of 0
|
|
to 255, 8-bit unsigned integer divide is used instead of
|
|
32-bit/64-bit integer divide.
|
|
|
|
<br><dt><code>-mavx256-split-unaligned-load</code><dt><code>-mavx256-split-unaligned-store</code><dd><a name="index-mavx256_002dsplit_002dunaligned_002dload-2746"></a><a name="index-mavx256_002dsplit_002dunaligned_002dstore-2747"></a>Split 32-byte AVX unaligned load and store.
|
|
|
|
<br><dt><code>-mstack-protector-guard=</code><var>guard</var><dd><a name="index-mstack_002dprotector_002dguard_003d_0040var_007bguard_007d-2748"></a>Generate stack protection code using canary at <var>guard</var>. Supported
|
|
locations are ‘<samp><span class="samp">global</span></samp>’ for global canary or ‘<samp><span class="samp">tls</span></samp>’ for per-thread
|
|
canary in the TLS block (the default). This option has effect only when
|
|
<samp><span class="option">-fstack-protector</span></samp> or <samp><span class="option">-fstack-protector-all</span></samp> is specified.
|
|
|
|
</dl>
|
|
|
|
<p>These ‘<samp><span class="samp">-m</span></samp>’ switches are supported in addition to the above
|
|
on x86-64 processors in 64-bit environments.
|
|
|
|
<dl>
|
|
<dt><code>-m32</code><dt><code>-m64</code><dt><code>-mx32</code><dt><code>-m16</code><dd><a name="index-m32-2749"></a><a name="index-m64-2750"></a><a name="index-mx32-2751"></a><a name="index-m16-2752"></a>Generate code for a 16-bit, 32-bit or 64-bit environment.
|
|
The <samp><span class="option">-m32</span></samp> option sets <code>int</code>, <code>long</code>, and pointer types
|
|
to 32 bits, and
|
|
generates code that runs on any i386 system.
|
|
|
|
<p>The <samp><span class="option">-m64</span></samp> option sets <code>int</code> to 32 bits and <code>long</code> and pointer
|
|
types to 64 bits, and generates code for the x86-64 architecture.
|
|
For Darwin only the <samp><span class="option">-m64</span></samp> option also turns off the <samp><span class="option">-fno-pic</span></samp>
|
|
and <samp><span class="option">-mdynamic-no-pic</span></samp> options.
|
|
|
|
<p>The <samp><span class="option">-mx32</span></samp> option sets <code>int</code>, <code>long</code>, and pointer types
|
|
to 32 bits, and
|
|
generates code for the x86-64 architecture.
|
|
|
|
<p>The <samp><span class="option">-m16</span></samp> option is the same as <samp><span class="option">-m32</span></samp>, except for that
|
|
it outputs the <code>.code16gcc</code> assembly directive at the beginning of
|
|
the assembly output so that the binary can run in 16-bit mode.
|
|
|
|
<br><dt><code>-mno-red-zone</code><dd><a name="index-mno_002dred_002dzone-2753"></a>Do not use a so-called “red zone” for x86-64 code. The red zone is mandated
|
|
by the x86-64 ABI; it is a 128-byte area beyond the location of the
|
|
stack pointer that is not modified by signal or interrupt handlers
|
|
and therefore can be used for temporary data without adjusting the stack
|
|
pointer. The flag <samp><span class="option">-mno-red-zone</span></samp> disables this red zone.
|
|
|
|
<br><dt><code>-mcmodel=small</code><dd><a name="index-mcmodel_003dsmall-2754"></a>Generate code for the small code model: the program and its symbols must
|
|
be linked in the lower 2 GB of the address space. Pointers are 64 bits.
|
|
Programs can be statically or dynamically linked. This is the default
|
|
code model.
|
|
|
|
<br><dt><code>-mcmodel=kernel</code><dd><a name="index-mcmodel_003dkernel-2755"></a>Generate code for the kernel code model. The kernel runs in the
|
|
negative 2 GB of the address space.
|
|
This model has to be used for Linux kernel code.
|
|
|
|
<br><dt><code>-mcmodel=medium</code><dd><a name="index-mcmodel_003dmedium-2756"></a>Generate code for the medium model: the program is linked in the lower 2
|
|
GB of the address space. Small symbols are also placed there. Symbols
|
|
with sizes larger than <samp><span class="option">-mlarge-data-threshold</span></samp> are put into
|
|
large data or BSS sections and can be located above 2GB. Programs can
|
|
be statically or dynamically linked.
|
|
|
|
<br><dt><code>-mcmodel=large</code><dd><a name="index-mcmodel_003dlarge-2757"></a>Generate code for the large model. This model makes no assumptions
|
|
about addresses and sizes of sections.
|
|
|
|
<br><dt><code>-maddress-mode=long</code><dd><a name="index-maddress_002dmode_003dlong-2758"></a>Generate code for long address mode. This is only supported for 64-bit
|
|
and x32 environments. It is the default address mode for 64-bit
|
|
environments.
|
|
|
|
<br><dt><code>-maddress-mode=short</code><dd><a name="index-maddress_002dmode_003dshort-2759"></a>Generate code for short address mode. This is only supported for 32-bit
|
|
and x32 environments. It is the default address mode for 32-bit and
|
|
x32 environments.
|
|
</dl>
|
|
|
|
</body></html>
|
|
|