gcc man page

gcc — GNU project C and C++ compiler


gcc {{source1.c}} {{source2.c}} -o {{executable}}

gcc {{source.c}} -Wall -Og -o {{executable}}

gcc {{source.c}} -o {{executable}} -I{{header_path}} -L{{library_path}} -l{{library_name}}

gcc -S {{source.c}}

gcc -c {{source.c}}


gcc [-c|-S|-E] [-std=standard]
   [-g] [-pg] [-Olevel]
   [-Wwarn...] [-Wpedantic]
   [-Idir...] [-Ldir...]
   [-Dmacro[=defn]...] [-Umacro]
   [-foption...] [-mmachine-option...]
   [-o outfile] [@file] infile...

Only the most useful options are listed here; see below for the remainder.  g++ accepts mostly the same options as gcc.


When you invoke GCC, it normally does preprocessing, compilation, assembly and linking.  The "overall options" allow you to stop this process at an intermediate stage.  For example, the -c option says not to run the linker.  Then the output consists of object files output by the assembler.

Other options are passed on to one or more stages of processing.  Some options control the preprocessor and others the compiler itself.  Yet other options control the assembler and linker; most of these are not documented here, since you rarely need to use any of them.

Most of the command-line options that you can use with GCC are useful for C programs; when an option is only useful with another language (usually C++), the explanation says so explicitly.  If the description for a particular option does not mention a source language, you can use that option with all supported languages.

The usual way to run GCC is to run the executable called gcc, or machine-gcc when cross-compiling, or machine-gcc-version to run a specific version of GCC. When you compile C++ programs, you should invoke GCC as g++  instead.

The gcc program accepts options and file names as operands.  Many options have multi-letter names; therefore multiple single-letter options may not be grouped: -dv is very different from -d -v.

You can mix options and other arguments.  For the most part, the order you use doesn't matter.  Order does matter when you use several options of the same kind; for example, if you specify -L more than once, the directories are searched in the order specified.  Also, the placement of the -l option is significant.

Many options have long names starting with -f or with -W---for example, -fmove-loop-invariants, -Wformat and so on.  Most of these have both positive and negative forms; the negative form of -ffoo is -fno-foo.  This manual documents only one of these two forms, whichever one is not the default.


Option Summary

Here is a summary of all the options, grouped by type.  Explanations are in the following sections.

Overall Options

-c  -S  -E  -o file  -x language   -v  -###  --help[=class[,...]]  --target-help  --version  -pass-exit-codes  -pipe  -specs=file  -wrapper   @file  -fplugin=file  -fplugin-arg-name=arg   -fdump-ada-spec[-slim]  -fada-spec-parent=unit  -fdump-go-spec=file

C Language Options

-ansi  -std=standard  -fgnu89-inline  -fpermitted-flt-eval-methods=standard  -aux-info filename  -fallow-parameterless-variadic-functions  -fno-asm  -fno-builtin  -fno-builtin-function  -fgimple -fhosted  -ffreestanding  -fopenacc  -fopenmp  -fopenmp-simd  -fms-extensions  -fplan9-extensions  -fsso-struct=endianness  -fallow-single-precision  -fcond-mismatch  -flax-vector-conversions  -fsigned-bitfields  -fsigned-char  -funsigned-bitfields  -funsigned-char

C++ Language Options

-fabi-version=n  -fno-access-control  -faligned-new=n  -fargs-in-order=n  -fcheck-new  -fconstexpr-depth=n  -fconstexpr-loop-limit=n  -ffriend-injection  -fno-elide-constructors  -fno-enforce-eh-specs  -ffor-scope  -fno-for-scope  -fno-gnu-keywords  -fno-implicit-templates  -fno-implicit-inline-templates  -fno-implement-inlines  -fms-extensions  -fnew-inheriting-ctors  -fnew-ttp-matching  -fno-nonansi-builtins  -fnothrow-opt  -fno-operator-names  -fno-optional-diags  -fpermissive  -fno-pretty-templates  -frepo  -fno-rtti  -fsized-deallocation  -ftemplate-backtrace-limit=n  -ftemplate-depth=n  -fno-threadsafe-statics  -fuse-cxa-atexit  -fno-weak  -nostdinc++  -fvisibility-inlines-hidden  -fvisibility-ms-compat  -fext-numeric-literals  -Wabi=n  -Wabi-tag  -Wconversion-null  -Wctor-dtor-privacy  -Wdelete-non-virtual-dtor  -Wliteral-suffix  -Wmultiple-inheritance  -Wnamespaces  -Wnarrowing  -Wnoexcept  -Wnoexcept-type  -Wnon-virtual-dtor  -Wreorder  -Wregister  -Weffc++  -Wstrict-null-sentinel  -Wtemplates  -Wno-non-template-friend  -Wold-style-cast  -Woverloaded-virtual  -Wno-pmf-conversions  -Wsign-promo  -Wvirtual-inheritance

Objective-C and Objective-C++ Language Options

-fconstant-string-class=class-name  -fgnu-runtime  -fnext-runtime  -fno-nil-receivers  -fobjc-abi-version=n  -fobjc-call-cxx-cdtors  -fobjc-direct-dispatch  -fobjc-exceptions  -fobjc-gc  -fobjc-nilcheck  -fobjc-std=objc1  -fno-local-ivars  -fivar-visibility=[public|protected|private|package]  -freplace-objc-classes  -fzero-link  -gen-decls  -Wassign-intercept  -Wno-protocol  -Wselector  -Wstrict-selector-match  -Wundeclared-selector

Diagnostic Message Formatting Options

-fmessage-length=n   -fdiagnostics-show-location=[once|every-line]   -fdiagnostics-color=[auto|never|always]   -fno-diagnostics-show-option  -fno-diagnostics-show-caret  -fdiagnostics-parseable-fixits  -fdiagnostics-generate-patch  -fno-show-column

Warning Options

-fsyntax-only  -fmax-errors=n  -Wpedantic  -pedantic-errors  -w  -Wextra  -Wall  -Waddress  -Waggregate-return   -Walloc-zero  -Walloc-size-larger-than=n -Walloca  -Walloca-larger-than=n  -Wno-aggressive-loop-optimizations  -Warray-bounds  -Warray-bounds=n  -Wno-attributes  -Wbool-compare  -Wbool-operation  -Wno-builtin-declaration-mismatch  -Wno-builtin-macro-redefined  -Wc90-c99-compat  -Wc99-c11-compat  -Wc++-compat  -Wc++11-compat  -Wc++14-compat  -Wcast-align  -Wcast-qual   -Wchar-subscripts -Wchkp  -Wclobbered  -Wcomment   -Wconditionally-supported   -Wconversion  -Wcoverage-mismatch  -Wno-cpp  -Wdangling-else  -Wdate-time  -Wdelete-incomplete  -Wno-deprecated  -Wno-deprecated-declarations  -Wno-designated-init  -Wdisabled-optimization  -Wno-discarded-qualifiers  -Wno-discarded-array-qualifiers  -Wno-div-by-zero  -Wdouble-promotion  -Wduplicated-branches  -Wduplicated-cond  -Wempty-body  -Wenum-compare  -Wno-endif-labels  -Wexpansion-to-defined  -Werror  -Werror=*  -Wfatal-errors  -Wfloat-equal  -Wformat  -Wformat=2  -Wno-format-contains-nul  -Wno-format-extra-args   -Wformat-nonliteral -Wformat-overflow=n  -Wformat-security  -Wformat-signedness  -Wformat-truncation=n  -Wformat-y2k  -Wframe-address  -Wframe-larger-than=len  -Wno-free-nonheap-object  -Wjump-misses-init  -Wignored-qualifiers  -Wignored-attributes  -Wincompatible-pointer-types  -Wimplicit  -Wimplicit-fallthrough  -Wimplicit-fallthrough=n  -Wimplicit-function-declaration  -Wimplicit-int  -Winit-self  -Winline  -Wno-int-conversion  -Wint-in-bool-context  -Wno-int-to-pointer-cast  -Winvalid-memory-model  -Wno-invalid-offsetof  -Winvalid-pch  -Wlarger-than=len  -Wlogical-op  -Wlogical-not-parentheses  -Wlong-long  -Wmain  -Wmaybe-uninitialized  -Wmemset-elt-size  -Wmemset-transposed-args  -Wmisleading-indentation  -Wmissing-braces  -Wmissing-field-initializers  -Wmissing-include-dirs  -Wno-multichar  -Wnonnull  -Wnonnull-compare  -Wnormalized=[none|id|nfc|nfkc]  -Wnull-dereference  -Wodr  -Wno-overflow  -Wopenmp-simd   -Woverride-init-side-effects  -Woverlength-strings  -Wpacked  -Wpacked-bitfield-compat  -Wpadded  -Wparentheses  -Wno-pedantic-ms-format  -Wplacement-new  -Wplacement-new=n  -Wpointer-arith  -Wpointer-compare  -Wno-pointer-to-int-cast  -Wno-pragmas  -Wredundant-decls  -Wrestrict  -Wno-return-local-addr  -Wreturn-type  -Wsequence-point  -Wshadow  -Wno-shadow-ivar  -Wshadow=global,  -Wshadow=local,  -Wshadow=compatible-local  -Wshift-overflow  -Wshift-overflow=n  -Wshift-count-negative  -Wshift-count-overflow  -Wshift-negative-value  -Wsign-compare  -Wsign-conversion  -Wfloat-conversion  -Wno-scalar-storage-order  -Wsizeof-pointer-memaccess  -Wsizeof-array-argument  -Wstack-protector  -Wstack-usage=len  -Wstrict-aliasing  -Wstrict-aliasing=n  -Wstrict-overflow  -Wstrict-overflow=n  -Wstringop-overflow=n  -Wsuggest-attribute=[pure|const|noreturn|format]  -Wsuggest-final-types   -Wsuggest-final-methods  -Wsuggest-override  -Wmissing-format-attribute  -Wsubobject-linkage  -Wswitch  -Wswitch-bool  -Wswitch-default  -Wswitch-enum  -Wswitch-unreachable  -Wsync-nand  -Wsystem-headers  -Wtautological-compare  -Wtrampolines  -Wtrigraphs  -Wtype-limits  -Wundef  -Wuninitialized  -Wunknown-pragmas  -Wunsafe-loop-optimizations  -Wunsuffixed-float-constants  -Wunused  -Wunused-function  -Wunused-label  -Wunused-local-typedefs  -Wunused-macros  -Wunused-parameter  -Wno-unused-result  -Wunused-value  -Wunused-variable  -Wunused-const-variable  -Wunused-const-variable=n  -Wunused-but-set-parameter  -Wunused-but-set-variable  -Wuseless-cast  -Wvariadic-macros  -Wvector-operation-performance  -Wvla  -Wvla-larger-than=n  -Wvolatile-register-var  -Wwrite-strings  -Wzero-as-null-pointer-constant  -Whsa

C and Objective-C-only Warning Options

-Wbad-function-cast  -Wmissing-declarations  -Wmissing-parameter-type  -Wmissing-prototypes  -Wnested-externs  -Wold-style-declaration  -Wold-style-definition  -Wstrict-prototypes  -Wtraditional  -Wtraditional-conversion  -Wdeclaration-after-statement  -Wpointer-sign

Debugging Options

-g  -glevel  -gcoff  -gdwarf  -gdwarf-version  -ggdb  -grecord-gcc-switches  -gno-record-gcc-switches  -gstabs  -gstabs+  -gstrict-dwarf  -gno-strict-dwarf  -gcolumn-info  -gno-column-info  -gvms  -gxcoff  -gxcoff+  -gz[=type]  -fdebug-prefix-map=old=new  -fdebug-types-section  -feliminate-dwarf2-dups  -fno-eliminate-unused-debug-types  -femit-struct-debug-baseonly  -femit-struct-debug-reduced  -femit-struct-debug-detailed[=spec-list]  -feliminate-unused-debug-symbols  -femit-class-debug-always  -fno-merge-debug-strings  -fno-dwarf2-cfi-asm  -fvar-tracking  -fvar-tracking-assignments

Optimization Options

-faggressive-loop-optimizations  -falign-functions[=n]  -falign-jumps[=n]  -falign-labels[=n]  -falign-loops[=n]  -fassociative-math  -fauto-profile  -fauto-profile[=path]  -fauto-inc-dec  -fbranch-probabilities  -fbranch-target-load-optimize  -fbranch-target-load-optimize2  -fbtr-bb-exclusive  -fcaller-saves  -fcombine-stack-adjustments  -fconserve-stack  -fcompare-elim  -fcprop-registers  -fcrossjumping  -fcse-follow-jumps  -fcse-skip-blocks  -fcx-fortran-rules  -fcx-limited-range  -fdata-sections  -fdce  -fdelayed-branch  -fdelete-null-pointer-checks  -fdevirtualize  -fdevirtualize-speculatively  -fdevirtualize-at-ltrans  -fdse  -fearly-inlining  -fipa-sra  -fexpensive-optimizations  -ffat-lto-objects  -ffast-math  -ffinite-math-only  -ffloat-store  -fexcess-precision=style  -fforward-propagate  -ffp-contract=style  -ffunction-sections  -fgcse  -fgcse-after-reload  -fgcse-las  -fgcse-lm  -fgraphite-identity  -fgcse-sm  -fhoist-adjacent-loads  -fif-conversion  -fif-conversion2  -findirect-inlining  -finline-functions  -finline-functions-called-once  -finline-limit=n  -finline-small-functions  -fipa-cp  -fipa-cp-clone  -fipa-bit-cp -fipa-vrp  -fipa-pta  -fipa-profile  -fipa-pure-const  -fipa-reference  -fipa-icf  -fira-algorithm=algorithm  -fira-region=region  -fira-hoist-pressure  -fira-loop-pressure  -fno-ira-share-save-slots  -fno-ira-share-spill-slots  -fisolate-erroneous-paths-dereference  -fisolate-erroneous-paths-attribute  -fivopts  -fkeep-inline-functions  -fkeep-static-functions  -fkeep-static-consts  -flimit-function-alignment  -flive-range-shrinkage  -floop-block  -floop-interchange  -floop-strip-mine  -floop-unroll-and-jam  -floop-nest-optimize  -floop-parallelize-all  -flra-remat  -flto  -flto-compression-level  -flto-partition=alg  -fmerge-all-constants  -fmerge-constants  -fmodulo-sched  -fmodulo-sched-allow-regmoves  -fmove-loop-invariants  -fno-branch-count-reg  -fno-defer-pop  -fno-fp-int-builtin-inexact  -fno-function-cse  -fno-guess-branch-probability  -fno-inline  -fno-math-errno  -fno-peephole  -fno-peephole2  -fno-printf-return-value  -fno-sched-interblock  -fno-sched-spec  -fno-signed-zeros  -fno-toplevel-reorder  -fno-trapping-math  -fno-zero-initialized-in-bss  -fomit-frame-pointer  -foptimize-sibling-calls  -fpartial-inlining  -fpeel-loops  -fpredictive-commoning  -fprefetch-loop-arrays  -fprofile-correction  -fprofile-use  -fprofile-use=path  -fprofile-values  -fprofile-reorder-functions  -freciprocal-math  -free  -frename-registers  -freorder-blocks  -freorder-blocks-algorithm=algorithm  -freorder-blocks-and-partition  -freorder-functions  -frerun-cse-after-loop  -freschedule-modulo-scheduled-loops  -frounding-math  -fsched2-use-superblocks  -fsched-pressure  -fsched-spec-load  -fsched-spec-load-dangerous  -fsched-stalled-insns-dep[=n]  -fsched-stalled-insns[=n]  -fsched-group-heuristic  -fsched-critical-path-heuristic  -fsched-spec-insn-heuristic  -fsched-rank-heuristic  -fsched-last-insn-heuristic  -fsched-dep-count-heuristic  -fschedule-fusion  -fschedule-insns  -fschedule-insns2  -fsection-anchors  -fselective-scheduling  -fselective-scheduling2  -fsel-sched-pipelining  -fsel-sched-pipelining-outer-loops  -fsemantic-interposition  -fshrink-wrap  -fshrink-wrap-separate  -fsignaling-nans  -fsingle-precision-constant  -fsplit-ivs-in-unroller  -fsplit-loops -fsplit-paths  -fsplit-wide-types  -fssa-backprop  -fssa-phiopt  -fstdarg-opt  -fstore-merging  -fstrict-aliasing  -fstrict-overflow  -fthread-jumps  -ftracer  -ftree-bit-ccp  -ftree-builtin-call-dce  -ftree-ccp  -ftree-ch  -ftree-coalesce-vars  -ftree-copy-prop  -ftree-dce  -ftree-dominator-opts  -ftree-dse  -ftree-forwprop  -ftree-fre  -fcode-hoisting  -ftree-loop-if-convert  -ftree-loop-im  -ftree-phiprop  -ftree-loop-distribution  -ftree-loop-distribute-patterns  -ftree-loop-ivcanon  -ftree-loop-linear  -ftree-loop-optimize  -ftree-loop-vectorize  -ftree-parallelize-loops=n  -ftree-pre  -ftree-partial-pre  -ftree-pta  -ftree-reassoc  -ftree-sink  -ftree-slsr  -ftree-sra  -ftree-switch-conversion  -ftree-tail-merge  -ftree-ter  -ftree-vectorize  -ftree-vrp  -funconstrained-commons  -funit-at-a-time  -funroll-all-loops  -funroll-loops  -funsafe-math-optimizations  -funswitch-loops  -fipa-ra  -fvariable-expansion-in-unroller  -fvect-cost-model  -fvpt  -fweb  -fwhole-program  -fwpa  -fuse-linker-plugin  --param name=value -O  -O0  -O1  -O2  -O3  -Os  -Ofast  -Og

Program Instrumentation Options

-p  -pg  -fprofile-arcs  --coverage  -ftest-coverage  -fprofile-dir=path  -fprofile-generate  -fprofile-generate=path  -fsanitize=style  -fsanitize-recover  -fsanitize-recover=style  -fasan-shadow-offset=number  -fsanitize-sections=s1,s2,...  -fsanitize-undefined-trap-on-error  -fbounds-check  -fcheck-pointer-bounds  -fchkp-check-incomplete-type  -fchkp-first-field-has-own-bounds  -fchkp-narrow-bounds  -fchkp-narrow-to-innermost-array  -fchkp-optimize  -fchkp-use-fast-string-functions  -fchkp-use-nochk-string-functions  -fchkp-use-static-bounds  -fchkp-use-static-const-bounds  -fchkp-treat-zero-dynamic-size-as-infinite  -fchkp-check-read  -fchkp-check-read  -fchkp-check-write  -fchkp-store-bounds  -fchkp-instrument-calls  -fchkp-instrument-marked-only  -fchkp-use-wrappers  -fchkp-flexible-struct-trailing-arrays -fstack-protector  -fstack-protector-all  -fstack-protector-strong  -fstack-protector-explicit  -fstack-check  -fstack-limit-register=reg  -fstack-limit-symbol=sym  -fno-stack-limit  -fsplit-stack  -fvtable-verify=[std|preinit|none]  -fvtv-counts  -fvtv-debug  -finstrument-functions  -finstrument-functions-exclude-function-list=sym,sym,...  -finstrument-functions-exclude-file-list=file,file,...

Preprocessor Options

-Aquestion=answer  -A-question[=answer]  -C  -CC  -Dmacro[=defn]  -dD  -dI  -dM  -dN  -dU  -fdebug-cpp  -fdirectives-only  -fdollars-in-identifiers   -fexec-charset=charset  -fextended-identifiers   -finput-charset=charset  -fno-canonical-system-headers  -fpch-deps  -fpch-preprocess  -fpreprocessed  -ftabstop=width  -ftrack-macro-expansion   -fwide-exec-charset=charset  -fworking-directory  -H  -imacros file  -include file  -M  -MD  -MF  -MG  -MM  -MMD  -MP  -MQ  -MT  -no-integrated-cpp  -P  -pthread  -remap  -traditional  -traditional-cpp  -trigraphs  -Umacro  -undef   -Wp,option  -Xpreprocessor option

Assembler Options

-Wa,option  -Xassembler option

Linker Options

object-file-name  -fuse-ld=linker  -llibrary  -nostartfiles  -nodefaultlibs  -nostdlib  -pie  -pthread  -rdynamic  -s  -static  -static-libgcc  -static-libstdc++  -static-libasan  -static-libtsan  -static-liblsan  -static-libubsan  -static-libmpx  -static-libmpxwrappers  -shared  -shared-libgcc  -symbolic  -T script  -Wl,option  -Xlinker option  -u symbol  -z keyword

Directory Options

-Bprefix  -Idir  -I-  -idirafter dir  -imacros file  -imultilib dir  -iplugindir=dir  -iprefix file  -iquote dir  -isysroot dir  -isystem dir  -iwithprefix dir  -iwithprefixbefore dir   -Ldir  -no-canonical-prefixes  --no-sysroot-suffix  -nostdinc  -nostdinc++  --sysroot=dir

Code Generation Options

-fcall-saved-reg  -fcall-used-reg  -ffixed-reg  -fexceptions  -fnon-call-exceptions  -fdelete-dead-exceptions  -funwind-tables  -fasynchronous-unwind-tables  -fno-gnu-unique  -finhibit-size-directive  -fno-common  -fno-ident  -fpcc-struct-return  -fpic  -fPIC  -fpie  -fPIE  -fno-plt  -fno-jump-tables  -frecord-gcc-switches  -freg-struct-return  -fshort-enums  -fshort-wchar  -fverbose-asm  -fpack-struct[=n]   -fleading-underscore  -ftls-model=model  -fstack-reuse=reuse_level  -ftrampolines  -ftrapv  -fwrapv  -fvisibility=[default|internal|hidden|protected]  -fstrict-volatile-bitfields  -fsync-libcalls

Developer Options

-dletters  -dumpspecs  -dumpmachine  -dumpversion  -dumpfullversion  -fchecking  -fchecking=n  -fdbg-cnt-list  -fdbg-cnt=counter-value-list  -fdisable-ipa-pass_name  -fdisable-rtl-pass_name  -fdisable-rtl-pass-name=range-list  -fdisable-tree-pass_name  -fdisable-tree-pass-name=range-list  -fdump-noaddr  -fdump-unnumbered  -fdump-unnumbered-links  -fdump-translation-unit[-n]  -fdump-class-hierarchy[-n]  -fdump-ipa-all  -fdump-ipa-cgraph  -fdump-ipa-inline  -fdump-passes  -fdump-rtl-pass  -fdump-rtl-pass=filename  -fdump-statistics  -fdump-final-insns[=file] -fdump-tree-all  -fdump-tree-switch  -fdump-tree-switch-options  -fdump-tree-switch-options=filename  -fcompare-debug[=opts]  -fcompare-debug-second  -fenable-kind-pass  -fenable-kind-pass=range-list  -fira-verbose=n  -flto-report  -flto-report-wpa  -fmem-report-wpa  -fmem-report  -fpre-ipa-mem-report  -fpost-ipa-mem-report  -fopt-info  -fopt-info-options[=file]  -fprofile-report  -frandom-seed=string  -fsched-verbose=n  -fsel-sched-verbose  -fsel-sched-dump-cfg  -fsel-sched-pipelining-verbose  -fstats  -fstack-usage  -ftime-report  -ftime-report-details  -fvar-tracking-assignments-toggle  -gtoggle  -print-file-name=library  -print-libgcc-file-name  -print-multi-directory  -print-multi-lib  -print-multi-os-directory  -print-prog-name=program  -print-search-dirs  -Q  -print-sysroot  -print-sysroot-headers-suffix  -save-temps  -save-temps=cwd  -save-temps=obj  -time[=file]

Machine-Dependent Options

AArch64 Options -mabi=name  -mbig-endian  -mlittle-endian  -mgeneral-regs-only  -mcmodel=tiny  -mcmodel=small  -mcmodel=large  -mstrict-align  -momit-leaf-frame-pointer  -mno-omit-leaf-frame-pointer  -mtls-dialect=desc  -mtls-dialect=traditional  -mtls-size=size  -mfix-cortex-a53-835769  -mno-fix-cortex-a53-835769  -mfix-cortex-a53-843419  -mno-fix-cortex-a53-843419  -mlow-precision-recip-sqrt  -mno-low-precision-recip-sqrt -mlow-precision-sqrt  -mno-low-precision-sqrt -mlow-precision-div  -mno-low-precision-div  -march=name  -mcpu=name  -mtune=name

Adapteva Epiphany Options -mhalf-reg-file  -mprefer-short-insn-regs  -mbranch-cost=num  -mcmove  -mnops=num  -msoft-cmpsf  -msplit-lohi  -mpost-inc  -mpost-modify  -mstack-offset=num  -mround-nearest  -mlong-calls  -mshort-calls  -msmall16  -mfp-mode=mode  -mvect-double  -max-vect-align=num  -msplit-vecmove-early  -m1reg-reg

ARC Options -mbarrel-shifter  -mcpu=cpu  -mA6  -mARC600  -mA7  -mARC700  -mdpfp  -mdpfp-compact  -mdpfp-fast  -mno-dpfp-lrsr  -mea  -mno-mpy  -mmul32x16  -mmul64  -matomic  -mnorm  -mspfp  -mspfp-compact  -mspfp-fast  -msimd  -msoft-float  -mswap  -mcrc  -mdsp-packa  -mdvbf  -mlock  -mmac-d16  -mmac-24  -mrtsc  -mswape  -mtelephony  -mxy  -misize  -mannotate-align  -marclinux  -marclinux_prof  -mlong-calls  -mmedium-calls  -msdata  -mvolatile-cache  -mtp-regno=regno  -malign-call  -mauto-modify-reg  -mbbit-peephole  -mno-brcc  -mcase-vector-pcrel  -mcompact-casesi  -mno-cond-exec  -mearly-cbranchsi  -mexpand-adddi  -mindexed-loads  -mlra  -mlra-priority-none  -mlra-priority-compact mlra-priority-noncompact  -mno-millicode  -mmixed-code  -mq-class  -mRcq  -mRcw  -msize-level=level  -mtune=cpu  -mmultcost=num  -munalign-prob-threshold=probability  -mmpy-option=multo  -mdiv-rem  -mcode-density  -mll64  -mfpu=fpu

ARM Options -mapcs-frame  -mno-apcs-frame  -mabi=name  -mapcs-stack-check  -mno-apcs-stack-check  -mapcs-reentrant  -mno-apcs-reentrant  -msched-prolog  -mno-sched-prolog  -mlittle-endian  -mbig-endian  -mfloat-abi=name  -mfp16-format=name -mthumb-interwork  -mno-thumb-interwork  -mcpu=name  -march=name  -mfpu=name   -mtune=name  -mprint-tune-info  -mstructure-size-boundary=n  -mabort-on-noreturn  -mlong-calls  -mno-long-calls  -msingle-pic-base  -mno-single-pic-base  -mpic-register=reg  -mnop-fun-dllimport  -mpoke-function-name  -mthumb  -marm  -mtpcs-frame  -mtpcs-leaf-frame  -mcaller-super-interworking  -mcallee-super-interworking  -mtp=name  -mtls-dialect=dialect  -mword-relocations  -mfix-cortex-m3-ldrd  -munaligned-access  -mneon-for-64bits  -mslow-flash-data  -masm-syntax-unified  -mrestrict-it  -mpure-code  -mcmse

AVR Options -mmcu=mcu  -mabsdata  -maccumulate-args  -mbranch-cost=cost  -mcall-prologues  -mint8  -mn_flash=size  -mno-interrupts  -mrelax  -mrmw  -mstrict-X  -mtiny-stack  -mfract-convert-truncate  -nodevicelib  -Waddr-space-convert  -Wmisspelled-isr

Blackfin Options -mcpu=cpu[-sirevision]  -msim  -momit-leaf-frame-pointer  -mno-omit-leaf-frame-pointer  -mspecld-anomaly  -mno-specld-anomaly  -mcsync-anomaly  -mno-csync-anomaly  -mlow-64k  -mno-low64k  -mstack-check-l1  -mid-shared-library  -mno-id-shared-library  -mshared-library-id=n  -mleaf-id-shared-library  -mno-leaf-id-shared-library  -msep-data  -mno-sep-data  -mlong-calls  -mno-long-calls  -mfast-fp  -minline-plt  -mmulticore  -mcorea  -mcoreb  -msdram  -micplb

C6X Options -mbig-endian  -mlittle-endian  -march=cpu  -msim  -msdata=sdata-type

CRIS Options -mcpu=cpu  -march=cpu  -mtune=cpu  -mmax-stack-frame=n  -melinux-stacksize=n  -metrax4  -metrax100  -mpdebug  -mcc-init  -mno-side-effects  -mstack-align  -mdata-align  -mconst-align  -m32-bit  -m16-bit  -m8-bit  -mno-prologue-epilogue  -mno-gotplt  -melf  -maout  -melinux  -mlinux  -sim  -sim2  -mmul-bug-workaround  -mno-mul-bug-workaround

CR16 Options -mmac  -mcr16cplus  -mcr16c  -msim  -mint32  -mbit-ops -mdata-model=model

Darwin Options -all_load  -allowable_client  -arch  -arch_errors_fatal  -arch_only  -bind_at_load  -bundle  -bundle_loader  -client_name  -compatibility_version  -current_version  -dead_strip  -dependency-file  -dylib_file  -dylinker_install_name  -dynamic  -dynamiclib  -exported_symbols_list  -filelist  -flat_namespace  -force_cpusubtype_ALL  -force_flat_namespace  -headerpad_max_install_names  -iframework  -image_base  -init  -install_name  -keep_private_externs  -multi_module  -multiply_defined  -multiply_defined_unused  -noall_load   -no_dead_strip_inits_and_terms  -nofixprebinding  -nomultidefs  -noprebind  -noseglinkedit  -pagezero_size  -prebind  -prebind_all_twolevel_modules  -private_bundle  -read_only_relocs  -sectalign  -sectobjectsymbols  -whyload  -seg1addr  -sectcreate  -sectobjectsymbols  -sectorder  -segaddr  -segs_read_only_addr  -segs_read_write_addr  -seg_addr_table  -seg_addr_table_filename  -seglinkedit  -segprot  -segs_read_only_addr  -segs_read_write_addr  -single_module  -static  -sub_library  -sub_umbrella  -twolevel_namespace  -umbrella  -undefined  -unexported_symbols_list  -weak_reference_mismatches  -whatsloaded  -F  -gused  -gfull  -mmacosx-version-min=version  -mkernel  -mone-byte-bool

DEC Alpha Options -mno-fp-regs  -msoft-float  -mieee  -mieee-with-inexact  -mieee-conformant  -mfp-trap-mode=mode  -mfp-rounding-mode=mode  -mtrap-precision=mode  -mbuild-constants  -mcpu=cpu-type  -mtune=cpu-type  -mbwx  -mmax  -mfix  -mcix  -mfloat-vax  -mfloat-ieee  -mexplicit-relocs  -msmall-data  -mlarge-data  -msmall-text  -mlarge-text  -mmemory-latency=time

FR30 Options -msmall-model  -mno-lsim

FT32 Options -msim  -mlra  -mnodiv

FRV Options -mgpr-32  -mgpr-64  -mfpr-32  -mfpr-64  -mhard-float  -msoft-float  -malloc-cc  -mfixed-cc  -mdword  -mno-dword  -mdouble  -mno-double  -mmedia  -mno-media  -mmuladd  -mno-muladd  -mfdpic  -minline-plt  -mgprel-ro  -multilib-library-pic  -mlinked-fp  -mlong-calls  -malign-labels  -mlibrary-pic  -macc-4  -macc-8  -mpack  -mno-pack  -mno-eflags  -mcond-move  -mno-cond-move  -moptimize-membar  -mno-optimize-membar  -mscc  -mno-scc  -mcond-exec  -mno-cond-exec  -mvliw-branch  -mno-vliw-branch  -mmulti-cond-exec  -mno-multi-cond-exec  -mnested-cond-exec  -mno-nested-cond-exec  -mtomcat-stats  -mTLS  -mtls  -mcpu=cpu

GNU/Linux Options -mglibc  -muclibc  -mmusl  -mbionic  -mandroid  -tno-android-cc  -tno-android-ld

H8/300 Options -mrelax  -mh  -ms  -mn  -mexr  -mno-exr  -mint32  -malign-300

HPPA Options -march=architecture-type  -mcaller-copies  -mdisable-fpregs  -mdisable-indexing  -mfast-indirect-calls  -mgas  -mgnu-ld   -mhp-ld  -mfixed-range=register-range  -mjump-in-delay  -mlinker-opt  -mlong-calls  -mlong-load-store  -mno-disable-fpregs  -mno-disable-indexing  -mno-fast-indirect-calls  -mno-gas  -mno-jump-in-delay  -mno-long-load-store  -mno-portable-runtime  -mno-soft-float  -mno-space-regs  -msoft-float  -mpa-risc-1-0  -mpa-risc-1-1  -mpa-risc-2-0  -mportable-runtime  -mschedule=cpu-type  -mspace-regs  -msio  -mwsio  -munix=unix-std  -nolibdld  -static  -threads

IA-64 Options -mbig-endian  -mlittle-endian  -mgnu-as  -mgnu-ld  -mno-pic  -mvolatile-asm-stop  -mregister-names  -msdata  -mno-sdata  -mconstant-gp  -mauto-pic  -mfused-madd  -minline-float-divide-min-latency  -minline-float-divide-max-throughput  -mno-inline-float-divide  -minline-int-divide-min-latency  -minline-int-divide-max-throughput   -mno-inline-int-divide  -minline-sqrt-min-latency  -minline-sqrt-max-throughput  -mno-inline-sqrt  -mdwarf2-asm  -mearly-stop-bits  -mfixed-range=register-range  -mtls-size=tls-size  -mtune=cpu-type  -milp32  -mlp64  -msched-br-data-spec  -msched-ar-data-spec  -msched-control-spec  -msched-br-in-data-spec  -msched-ar-in-data-spec  -msched-in-control-spec  -msched-spec-ldc  -msched-spec-control-ldc  -msched-prefer-non-data-spec-insns  -msched-prefer-non-control-spec-insns  -msched-stop-bits-after-every-cycle  -msched-count-spec-in-critical-path  -msel-sched-dont-check-control-spec  -msched-fp-mem-deps-zero-cost  -msched-max-memory-insns-hard-limit  -msched-max-memory-insns=max-insns

LM32 Options -mbarrel-shift-enabled  -mdivide-enabled  -mmultiply-enabled  -msign-extend-enabled  -muser-enabled

M32R/D Options -m32r2  -m32rx  -m32r  -mdebug  -malign-loops  -mno-align-loops  -missue-rate=number  -mbranch-cost=number  -mmodel=code-size-model-type  -msdata=sdata-type  -mno-flush-func  -mflush-func=name  -mno-flush-trap  -mflush-trap=number  -G num

M32C Options -mcpu=cpu  -msim  -memregs=number

M680x0 Options -march=arch  -mcpu=cpu  -mtune=tune  -m68000  -m68020  -m68020-40  -m68020-60  -m68030  -m68040  -m68060  -mcpu32  -m5200  -m5206e  -m528x  -m5307  -m5407  -mcfv4e  -mbitfield  -mno-bitfield  -mc68000  -mc68020  -mnobitfield  -mrtd  -mno-rtd  -mdiv  -mno-div  -mshort  -mno-short  -mhard-float  -m68881  -msoft-float  -mpcrel  -malign-int  -mstrict-align  -msep-data  -mno-sep-data  -mshared-library-id=n  -mid-shared-library  -mno-id-shared-library  -mxgot  -mno-xgot  -mlong-jump-table-offsets

MCore Options -mhardlit  -mno-hardlit  -mdiv  -mno-div  -mrelax-immediates  -mno-relax-immediates  -mwide-bitfields  -mno-wide-bitfields  -m4byte-functions  -mno-4byte-functions  -mcallgraph-data  -mno-callgraph-data  -mslow-bytes  -mno-slow-bytes  -mno-lsim  -mlittle-endian  -mbig-endian  -m210  -m340  -mstack-increment

MeP Options -mabsdiff  -mall-opts  -maverage  -mbased=n  -mbitops  -mc=n  -mclip  -mconfig=name  -mcop  -mcop32  -mcop64  -mivc2  -mdc  -mdiv  -meb  -mel  -mio-volatile  -ml  -mleadz  -mm  -mminmax  -mmult  -mno-opts  -mrepeat  -ms  -msatur  -msdram  -msim  -msimnovec  -mtf  -mtiny=n

MicroBlaze Options -msoft-float  -mhard-float  -msmall-divides  -mcpu=cpu  -mmemcpy  -mxl-soft-mul  -mxl-soft-div  -mxl-barrel-shift  -mxl-pattern-compare  -mxl-stack-check  -mxl-gp-opt  -mno-clearbss  -mxl-multiply-high  -mxl-float-convert  -mxl-float-sqrt  -mbig-endian  -mlittle-endian  -mxl-reorder  -mxl-mode-app-model

MIPS Options -EL  -EB  -march=arch  -mtune=arch  -mips1  -mips2  -mips3  -mips4  -mips32  -mips32r2  -mips32r3  -mips32r5  -mips32r6  -mips64  -mips64r2  -mips64r3  -mips64r5  -mips64r6  -mips16  -mno-mips16  -mflip-mips16  -minterlink-compressed  -mno-interlink-compressed  -minterlink-mips16  -mno-interlink-mips16  -mabi=abi  -mabicalls  -mno-abicalls  -mshared  -mno-shared  -mplt  -mno-plt  -mxgot  -mno-xgot  -mgp32  -mgp64  -mfp32  -mfpxx  -mfp64  -mhard-float  -msoft-float  -mno-float  -msingle-float  -mdouble-float  -modd-spreg  -mno-odd-spreg  -mabs=mode  -mnan=encoding  -mdsp  -mno-dsp  -mdspr2  -mno-dspr2  -mmcu  -mmno-mcu  -meva  -mno-eva  -mvirt  -mno-virt  -mxpa  -mno-xpa  -mmicromips  -mno-micromips  -mmsa  -mno-msa  -mfpu=fpu-type  -msmartmips  -mno-smartmips  -mpaired-single  -mno-paired-single  -mdmx  -mno-mdmx  -mips3d  -mno-mips3d  -mmt  -mno-mt  -mllsc  -mno-llsc  -mlong64  -mlong32  -msym32  -mno-sym32  -Gnum  -mlocal-sdata  -mno-local-sdata  -mextern-sdata  -mno-extern-sdata  -mgpopt  -mno-gopt  -membedded-data  -mno-embedded-data  -muninit-const-in-rodata  -mno-uninit-const-in-rodata  -mcode-readable=setting  -msplit-addresses  -mno-split-addresses  -mexplicit-relocs  -mno-explicit-relocs  -mcheck-zero-division  -mno-check-zero-division  -mdivide-traps  -mdivide-breaks  -mload-store-pairs  -mno-load-store-pairs  -mmemcpy  -mno-memcpy  -mlong-calls  -mno-long-calls  -mmad  -mno-mad  -mimadd  -mno-imadd  -mfused-madd  -mno-fused-madd  -nocpp  -mfix-24k  -mno-fix-24k  -mfix-r4000  -mno-fix-r4000  -mfix-r4400  -mno-fix-r4400  -mfix-r10000  -mno-fix-r10000  -mfix-rm7000  -mno-fix-rm7000  -mfix-vr4120  -mno-fix-vr4120  -mfix-vr4130  -mno-fix-vr4130  -mfix-sb1  -mno-fix-sb1  -mflush-func=func  -mno-flush-func  -mbranch-cost=num  -mbranch-likely  -mno-branch-likely  -mcompact-branches=policy  -mfp-exceptions  -mno-fp-exceptions  -mvr4130-align  -mno-vr4130-align  -msynci  -mno-synci  -mlxc1-sxc1 -mno-lxc1-sxc1 -mmadd4 -mno-madd4  -mrelax-pic-calls  -mno-relax-pic-calls  -mmcount-ra-address  -mframe-header-opt  -mno-frame-header-opt

MMIX Options -mlibfuncs  -mno-libfuncs  -mepsilon  -mno-epsilon  -mabi=gnu  -mabi=mmixware  -mzero-extend  -mknuthdiv  -mtoplevel-symbols  -melf  -mbranch-predict  -mno-branch-predict  -mbase-addresses  -mno-base-addresses  -msingle-exit  -mno-single-exit

MN10300 Options -mmult-bug  -mno-mult-bug  -mno-am33  -mam33  -mam33-2  -mam34  -mtune=cpu-type  -mreturn-pointer-on-d0  -mno-crt0  -mrelax  -mliw  -msetlb

Moxie Options -meb  -mel  -mmul.x  -mno-crt0

MSP430 Options -msim  -masm-hex  -mmcu=  -mcpu=  -mlarge  -msmall  -mrelax  -mwarn-mcu  -mcode-region=  -mdata-region=  -msilicon-errata=  -msilicon-errata-warn=  -mhwmult=  -minrt

NDS32 Options -mbig-endian  -mlittle-endian  -mreduced-regs  -mfull-regs  -mcmov  -mno-cmov  -mperf-ext  -mno-perf-ext  -mv3push  -mno-v3push  -m16bit  -mno-16bit  -misr-vector-size=num  -mcache-block-size=num  -march=arch  -mcmodel=code-model  -mctor-dtor  -mrelax

Nios II Options -G num  -mgpopt=option  -mgpopt  -mno-gpopt  -mel  -meb  -mno-bypass-cache  -mbypass-cache  -mno-cache-volatile  -mcache-volatile  -mno-fast-sw-div  -mfast-sw-div  -mhw-mul  -mno-hw-mul  -mhw-mulx  -mno-hw-mulx  -mno-hw-div  -mhw-div  -mcustom-insn=N  -mno-custom-insn  -mcustom-fpu-cfg=name  -mhal  -msmallc  -msys-crt0=name  -msys-lib=name  -march=arch  -mbmx  -mno-bmx  -mcdx  -mno-cdx

Nvidia PTX Options -m32  -m64  -mmainkernel  -moptimize

PDP-11 Options -mfpu  -msoft-float  -mac0  -mno-ac0  -m40  -m45  -m10  -mbcopy  -mbcopy-builtin  -mint32  -mno-int16  -mint16  -mno-int32  -mfloat32  -mno-float64  -mfloat64  -mno-float32  -mabshi  -mno-abshi  -mbranch-expensive  -mbranch-cheap  -munix-asm  -mdec-asm

picoChip Options -mae=ae_type  -mvliw-lookahead=N  -msymbol-as-address  -mno-inefficient-warnings

PowerPC Options See RS/6000 and PowerPC Options.

RISC-V Options -mbranch-cost=N-instruction  -mmemcpy  -mno-memcpy  -mplt  -mno-plt  -mabi=ABI-string  -mfdiv  -mno-fdiv  -mdiv  -mno-div  -march=ISA-string  -mtune=processor-string  -msmall-data-limit=N-bytes  -msave-restore  -mno-save-restore  -mcmodel=code-model  -mexplicit-relocs  -mno-explicit-relocs

RL78 Options -msim  -mmul=none  -mmul=g13  -mmul=g14  -mallregs  -mcpu=g10  -mcpu=g13  -mcpu=g14  -mg10  -mg13  -mg14  -m64bit-doubles  -m32bit-doubles  -msave-mduc-in-interrupts

RS/6000 and PowerPC Options -mcpu=cpu-type  -mtune=cpu-type  -mcmodel=code-model  -mpowerpc64  -maltivec  -mno-altivec  -mpowerpc-gpopt  -mno-powerpc-gpopt  -mpowerpc-gfxopt  -mno-powerpc-gfxopt  -mmfcrf  -mno-mfcrf  -mpopcntb  -mno-popcntb  -mpopcntd  -mno-popcntd  -mfprnd  -mno-fprnd  -mcmpb  -mno-cmpb  -mmfpgpr  -mno-mfpgpr  -mhard-dfp  -mno-hard-dfp  -mfull-toc   -mminimal-toc  -mno-fp-in-toc  -mno-sum-in-toc  -m64  -m32  -mxl-compat  -mno-xl-compat  -mpe  -malign-power  -malign-natural  -msoft-float  -mhard-float  -mmultiple  -mno-multiple  -msingle-float  -mdouble-float  -msimple-fpu  -mstring  -mno-string  -mupdate  -mno-update  -mavoid-indexed-addresses  -mno-avoid-indexed-addresses  -mfused-madd  -mno-fused-madd  -mbit-align  -mno-bit-align  -mstrict-align  -mno-strict-align  -mrelocatable  -mno-relocatable  -mrelocatable-lib  -mno-relocatable-lib  -mtoc  -mno-toc  -mlittle  -mlittle-endian  -mbig  -mbig-endian  -mdynamic-no-pic  -maltivec  -mswdiv  -msingle-pic-base  -mprioritize-restricted-insns=priority  -msched-costly-dep=dependence_type  -minsert-sched-nops=scheme  -mcall-sysv  -mcall-netbsd  -maix-struct-return  -msvr4-struct-return  -mabi=abi-type  -msecure-plt  -mbss-plt  -mblock-move-inline-limit=num  -misel  -mno-isel  -misel=yes  -misel=no  -mspe  -mno-spe  -mspe=yes  -mspe=no  -mpaired  -mgen-cell-microcode  -mwarn-cell-microcode  -mvrsave  -mno-vrsave  -mmulhw  -mno-mulhw  -mdlmzb  -mno-dlmzb  -mfloat-gprs=yes  -mfloat-gprs=no  -mfloat-gprs=single  -mfloat-gprs=double  -mprototype  -mno-prototype  -msim  -mmvme  -mads  -myellowknife  -memb  -msdata  -msdata=opt  -mvxworks  -G num  -mrecip  -mrecip=opt  -mno-recip  -mrecip-precision  -mno-recip-precision  -mveclibabi=type  -mfriz  -mno-friz  -mpointers-to-nested-functions  -mno-pointers-to-nested-functions  -msave-toc-indirect  -mno-save-toc-indirect  -mpower8-fusion  -mno-mpower8-fusion  -mpower8-vector  -mno-power8-vector  -mcrypto  -mno-crypto  -mhtm  -mno-htm  -mdirect-move  -mno-direct-move  -mquad-memory  -mno-quad-memory  -mquad-memory-atomic  -mno-quad-memory-atomic  -mcompat-align-parm  -mno-compat-align-parm  -mupper-regs-df  -mno-upper-regs-df  -mupper-regs-sf  -mno-upper-regs-sf  -mupper-regs-di  -mno-upper-regs-di  -mupper-regs  -mno-upper-regs  -mfloat128  -mno-float128  -mfloat128-hardware  -mno-float128-hardware  -mgnu-attribute  -mno-gnu-attribute  -mstack-protector-guard=guard -mstack-protector-guard-reg=reg  -mstack-protector-guard-offset=offset  -mlra  -mno-lra

RX Options -m64bit-doubles  -m32bit-doubles  -fpu  -nofpu -mcpu= -mbig-endian-data  -mlittle-endian-data  -msmall-data  -msim  -mno-sim -mas100-syntax  -mno-as100-syntax -mrelax -mmax-constant-size= -mint-register= -mpid -mallow-string-insns  -mno-allow-string-insns -mjsr -mno-warn-multiple-fast-interrupts -msave-acc-in-interrupts

S/390 and zSeries Options -mtune=cpu-type  -march=cpu-type  -mhard-float  -msoft-float  -mhard-dfp  -mno-hard-dfp  -mlong-double-64  -mlong-double-128  -mbackchain  -mno-backchain  -mpacked-stack  -mno-packed-stack  -msmall-exec  -mno-small-exec  -mmvcle  -mno-mvcle  -m64  -m31  -mdebug  -mno-debug  -mesa  -mzarch  -mhtm  -mvx  -mzvector  -mtpf-trace  -mno-tpf-trace  -mfused-madd  -mno-fused-madd  -mwarn-framesize  -mwarn-dynamicstack  -mstack-size  -mstack-guard  -mhotpatch=halfwords,halfwords

Score Options -meb  -mel  -mnhwloop  -muls  -mmac  -mscore5  -mscore5u  -mscore7  -mscore7d

SH Options -m1  -m2  -m2e  -m2a-nofpu  -m2a-single-only  -m2a-single  -m2a  -m3  -m3e  -m4-nofpu  -m4-single-only  -m4-single  -m4  -m4a-nofpu  -m4a-single-only  -m4a-single  -m4a  -m4al  -mb  -ml  -mdalign  -mrelax  -mbigtable  -mfmovd  -mrenesas  -mno-renesas  -mnomacsave  -mieee  -mno-ieee  -mbitops  -misize  -minline-ic_invalidate  -mpadstruct  -mprefergot  -musermode  -multcost=number  -mdiv=strategy  -mdivsi3_libfunc=name  -mfixed-range=register-range  -maccumulate-outgoing-args  -matomic-model=atomic-model  -mbranch-cost=num  -mzdcbranch  -mno-zdcbranch  -mcbranch-force-delay-slot  -mfused-madd  -mno-fused-madd  -mfsca  -mno-fsca  -mfsrra  -mno-fsrra  -mpretend-cmove  -mtas

Solaris 2 Options -mclear-hwcap  -mno-clear-hwcap  -mimpure-text  -mno-impure-text  -pthreads

SPARC Options -mcpu=cpu-type  -mtune=cpu-type  -mcmodel=code-model  -mmemory-model=mem-model  -m32  -m64  -mapp-regs  -mno-app-regs  -mfaster-structs  -mno-faster-structs  -mflat  -mno-flat  -mfpu  -mno-fpu  -mhard-float  -msoft-float  -mhard-quad-float  -msoft-quad-float  -mstack-bias  -mno-stack-bias  -mstd-struct-return  -mno-std-struct-return  -munaligned-doubles  -mno-unaligned-doubles  -muser-mode  -mno-user-mode  -mv8plus  -mno-v8plus  -mvis  -mno-vis  -mvis2  -mno-vis2  -mvis3  -mno-vis3  -mvis4 -mno-vis4 -mvis4b -mno-vis4b  -mcbcond  -mno-cbcond  -mfmaf  -mno-fmaf  -mfsmuld  -mno-fsmuld   -mpopc  -mno-popc  -msubxc  -mno-subxc  -mfix-at697f  -mfix-ut699  -mfix-ut700  -mfix-gr712rc  -mlra  -mno-lra

SPU Options -mwarn-reloc  -merror-reloc  -msafe-dma  -munsafe-dma  -mbranch-hints  -msmall-mem  -mlarge-mem  -mstdmain  -mfixed-range=register-range  -mea32  -mea64  -maddress-space-conversion  -mno-address-space-conversion  -mcache-size=cache-size  -matomic-updates  -mno-atomic-updates

System V Options -Qy  -Qn  -YP,paths  -Ym,dir

TILE-Gx Options -mcpu=CPU  -m32  -m64  -mbig-endian  -mlittle-endian  -mcmodel=code-model

TILEPro Options -mcpu=cpu  -m32

V850 Options -mlong-calls  -mno-long-calls  -mep  -mno-ep  -mprolog-function  -mno-prolog-function  -mspace  -mtda=n  -msda=n  -mzda=n  -mapp-regs  -mno-app-regs  -mdisable-callt  -mno-disable-callt  -mv850e2v3  -mv850e2  -mv850e1  -mv850es  -mv850e  -mv850  -mv850e3v5  -mloop  -mrelax  -mlong-jumps  -msoft-float  -mhard-float  -mgcc-abi  -mrh850-abi  -mbig-switch

VAX Options -mg  -mgnu  -munix

Visium Options -mdebug  -msim  -mfpu  -mno-fpu  -mhard-float  -msoft-float  -mcpu=cpu-type  -mtune=cpu-type  -msv-mode  -muser-mode

VMS Options -mvms-return-codes  -mdebug-main=prefix  -mmalloc64  -mpointer-size=size

VxWorks Options -mrtp  -non-static  -Bstatic  -Bdynamic  -Xbind-lazy  -Xbind-now

x86 Options -mtune=cpu-type  -march=cpu-type  -mtune-ctrl=feature-list  -mdump-tune-features  -mno-default  -mfpmath=unit  -masm=dialect  -mno-fancy-math-387  -mno-fp-ret-in-387  -m80387  -mhard-float  -msoft-float  -mno-wide-multiply  -mrtd  -malign-double  -mpreferred-stack-boundary=num  -mincoming-stack-boundary=num  -mcld  -mcx16  -msahf  -mmovbe  -mcrc32  -mrecip  -mrecip=opt  -mvzeroupper  -mprefer-avx128  -mmmx  -msse  -msse2  -msse3  -mssse3  -msse4.1  -msse4.2  -msse4  -mavx  -mavx2  -mavx512f  -mavx512pf  -mavx512er  -mavx512cd  -mavx512vl  -mavx512bw  -mavx512dq  -mavx512ifma  -mavx512vbmi  -msha  -maes  -mpclmul  -mfsgsbase  -mrdrnd  -mf16c  -mfma  -mprefetchwt1  -mclflushopt  -mxsavec  -mxsaves  -msse4a  -m3dnow  -m3dnowa  -mpopcnt  -mabm  -mbmi  -mtbm  -mfma4  -mxop  -mlzcnt  -mbmi2  -mfxsr  -mxsave  -mxsaveopt  -mrtm  -mlwp  -mmpx   -mmwaitx  -mclzero  -mpku  -mthreads  -mms-bitfields  -mno-align-stringops  -minline-all-stringops  -minline-stringops-dynamically  -mstringop-strategy=alg  -mmemcpy-strategy=strategy  -mmemset-strategy=strategy  -mpush-args  -maccumulate-outgoing-args  -m128bit-long-double  -m96bit-long-double  -mlong-double-64  -mlong-double-80  -mlong-double-128  -mregparm=num  -msseregparm  -mveclibabi=type  -mvect8-ret-in-mem  -mpc32  -mpc64  -mpc80  -mstackrealign  -momit-leaf-frame-pointer  -mno-red-zone  -mno-tls-direct-seg-refs  -mcmodel=code-model  -mabi=name  -maddress-mode=mode  -m32  -m64  -mx32  -m16  -miamcu  -mlarge-data-threshold=num  -msse2avx  -mfentry  -mrecord-mcount  -mnop-mcount  -m8bit-idiv  -mavx256-split-unaligned-load  -mavx256-split-unaligned-store  -malign-data=type  -mstack-protector-guard=guard  -mmitigate-rop  -mgeneral-regs-only

x86 Windows Options -mconsole  -mcygwin  -mno-cygwin  -mdll  -mnop-fun-dllimport  -mthread  -municode  -mwin32  -mwindows  -fno-set-stack-executable

Xstormy16 Options -msim

Xtensa Options -mconst16  -mno-const16  -mfused-madd  -mno-fused-madd  -mforce-no-pic  -mserialize-volatile  -mno-serialize-volatile  -mtext-section-literals  -mno-text-section-literals  -mauto-litpools  -mno-auto-litpools  -mtarget-align  -mno-target-align  -mlongcalls  -mno-longcalls

zSeries Options See S/390 and zSeries Options.

Options Controlling the Kind of Output

Compilation can involve up to four stages: preprocessing, compilation proper, assembly and linking, always in that order.  GCC is capable of preprocessing and compiling several files either into several assembler input files, or into one assembler input file; then each assembler input file produces an object file, and linking combines all the object files (those newly compiled, and those specified as input) into an executable file.

For any given input file, the file name suffix determines what kind of compilation is done:


C source code that must be preprocessed.


C source code that should not be preprocessed.


C++ source code that should not be preprocessed.


Objective-C source code.  Note that you must link with the libobjc library to make an Objective-C program work.


Objective-C source code that should not be preprocessed.


Objective-C++ source code.  Note that you must link with the libobjc library to make an Objective-C++ program work.  Note that .M refers to a literal capital M.


Objective-C++ source code that should not be preprocessed.


C, C++, Objective-C or Objective-C++ header file to be turned into a precompiled header (default), or C, C++ header file to be turned into an Ada spec (via the -fdump-ada-spec switch).


C++ source code that must be preprocessed.  Note that in .cxx, the last two letters must both be literally x.  Likewise, .C refers to a literal capital C.


Objective-C++ source code that must be preprocessed.


Objective-C++ source code that should not be preprocessed.


C++ header file to be turned into a precompiled header or Ada spec.


Fixed form Fortran source code that should not be preprocessed.


Fixed form Fortran source code that must be preprocessed (with the traditional preprocessor).


Free form Fortran source code that should not be preprocessed.


Free form Fortran source code that must be preprocessed (with the traditional preprocessor).


Go source code.


BRIG files (binary representation of HSAIL).


Ada source code file that contains a library unit declaration (a declaration of a package, subprogram, or generic, or a generic instantiation), or a library unit renaming declaration (a package, generic, or subprogram renaming declaration).  Such files are also called specs.


Ada source code file containing a library unit body (a subprogram or package body).  Such files are also called bodies.


Assembler code.


Assembler code that must be preprocessed.


An object file to be fed straight into linking. Any file name with no recognized suffix is treated this way.

You can specify the input language explicitly with the -x option:

-x language

Specify explicitly the language for the following input files (rather than letting the compiler choose a default based on the file name suffix).  This option applies to all following input files until the next -x option.  Possible values for language are:

        c  c-header  cpp-output
        c++  c++-header  c++-cpp-output
        objective-c  objective-c-header  objective-c-cpp-output
        objective-c++ objective-c++-header objective-c++-cpp-output
        assembler  assembler-with-cpp
        f77  f77-cpp-input f95  f95-cpp-input
-x none

Turn off any specification of a language, so that subsequent files are handled according to their file name suffixes (as they are if -x has not been used at all).

If you only want some of the stages of compilation, you can use -x (or filename suffixes) to tell gcc where to start, and one of the options -c, -S, or -E to say where gcc is to stop.  Note that some combinations (for example, -x cpp-output -E) instruct gcc to do nothing at all.


Compile or assemble the source files, but do not link.  The linking stage simply is not done.  The ultimate output is in the form of an object file for each source file.

By default, the object file name for a source file is made by replacing the suffix .c, .i, .s, etc., with .o.

Unrecognized input files, not requiring compilation or assembly, are ignored.


Stop after the stage of compilation proper; do not assemble.  The output is in the form of an assembler code file for each non-assembler input file specified.

By default, the assembler file name for a source file is made by replacing the suffix .c, .i, etc., with .s.

Input files that don't require compilation are ignored.


Stop after the preprocessing stage; do not run the compiler proper.  The output is in the form of preprocessed source code, which is sent to the standard output.

Input files that don't require preprocessing are ignored.

-o file

Place output in file file.  This applies to whatever sort of output is being produced, whether it be an executable file, an object file, an assembler file or preprocessed C code.

If -o is not specified, the default is to put an executable file in a.out, the object file for source.suffix in source.o, its assembler file in source.s, a precompiled header file in source.suffix.gch, and all preprocessed C source on standard output.


Print (on standard error output) the commands executed to run the stages of compilation.  Also print the version number of the compiler driver program and of the preprocessor and the compiler proper.


Like -v except the commands are not executed and arguments are quoted unless they contain only alphanumeric characters or "./-_". This is useful for shell scripts to capture the driver-generated command lines.


Print (on the standard output) a description of the command-line options understood by gcc.  If the -v option is also specified then --help is also passed on to the various processes invoked by gcc, so that they can display the command-line options they accept.  If the -Wextra option has also been specified (prior to the --help option), then command-line options that have no documentation associated with them are also displayed.


Print (on the standard output) a description of target-specific command-line options for each tool.  For some targets extra target-specific information may also be printed.


Print (on the standard output) a description of the command-line options understood by the compiler that fit into all specified classes and qualifiers.  These are the supported classes:


Display all of the optimization options supported by the compiler.


Display all of the options controlling warning messages produced by the compiler.


Display target-specific options.  Unlike the --target-help option however, target-specific options of the linker and assembler are not displayed.  This is because those tools do not currently support the extended --help= syntax.


Display the values recognized by the --param option.


Display the options supported for language, where language is the name of one of the languages supported in this version of GCC.


Display the options that are common to all languages.

These are the supported qualifiers:


Display only those options that are undocumented.


Display options taking an argument that appears after an equal sign in the same continuous piece of text, such as: --help=target.


Display options taking an argument that appears as a separate word following the original option, such as: -o output-file.

Thus for example to display all the undocumented target-specific switches supported by the compiler, use:


The sense of a qualifier can be inverted by prefixing it with the ^ character, so for example to display all binary warning options (i.e., ones that are either on or off and that do not take an argument) that have a description, use:


The argument to --help= should not consist solely of inverted qualifiers.

Combining several classes is possible, although this usually restricts the output so much that there is nothing to display.  One case where it does work, however, is when one of the classes is target.  For example, to display all the target-specific optimization options, use:


The --help= option can be repeated on the command line.  Each successive use displays its requested class of options, skipping those that have already been displayed.

If the -Q option appears on the command line before the --help= option, then the descriptive text displayed by --help= is changed.  Instead of describing the displayed options, an indication is given as to whether the option is enabled, disabled or set to a specific value (assuming that the compiler knows this at the point where the --help= option is used).

Here is a truncated example from the ARM port of gcc:

          % gcc -Q -mabi=2 --help=target -c
          The following options are target specific:
          -mabi=                                2
          -mabort-on-noreturn                   [disabled]
          -mapcs                                [disabled]

The output is sensitive to the effects of previous command-line options, so for example it is possible to find out which optimizations are enabled at -O2 by using:

        -Q -O2 --help=optimizers

Alternatively you can discover which binary optimizations are enabled by -O3 by using:

        gcc -c -Q -O3 --help=optimizers > /tmp/O3-opts
        gcc -c -Q -O2 --help=optimizers > /tmp/O2-opts
        diff /tmp/O2-opts /tmp/O3-opts | grep enabled

Display the version number and copyrights of the invoked GCC.


Normally the gcc program exits with the code of 1 if any phase of the compiler returns a non-success return code.  If you specify -pass-exit-codes, the gcc program instead returns with the numerically highest error produced by any phase returning an error indication.  The C, C++, and Fortran front ends return 4 if an internal compiler error is encountered.


Use pipes rather than temporary files for communication between the various stages of compilation.  This fails to work on some systems where the assembler is unable to read from a pipe; but the GNU assembler has no trouble.


Process file after the compiler reads in the standard specs file, in order to override the defaults which the gcc driver program uses when determining what switches to pass to cc1, cc1plus, as, ld, etc.  More than one -specs=file can be specified on the command line, and they are processed in order, from left to right.


Invoke all subcommands under a wrapper program.  The name of the wrapper program and its parameters are passed as a comma separated list.

        gcc -c t.c -wrapper gdb,--args

This invokes all subprograms of gcc under gdb --args, thus the invocation of cc1 is gdb --args cc1 ....


Load the plugin code in file name.so, assumed to be a shared object to be dlopen'd by the compiler.  The base name of the shared object file is used to identify the plugin for the purposes of argument parsing (See -fplugin-arg-name-key=value below). Each plugin should define the callback functions specified in the Plugins API.


Define an argument called key with a value of value for the plugin called name.


For C and C++ source and include files, generate corresponding Ada specs.


In conjunction with -fdump-ada-spec[-slim] above, generate Ada specs as child units of parent unit.


For input files in any language, generate corresponding Go declarations in file.  This generates Go "const", "type", "var", and "func" declarations which may be a useful way to start writing a Go interface to code written in some other language.


Read command-line options from file.  The options read are inserted in place of the original @file option.  If file does not exist, or cannot be read, then the option will be treated literally, and not removed.

Options in file are separated by whitespace.  A whitespace character may be included in an option by surrounding the entire option in either single or double quotes.  Any character (including a backslash) may be included by prefixing the character to be included with a backslash.  The file may itself contain additional @file options; any such options will be processed recursively.

Compiling C++ Programs

C++ source files conventionally use one of the suffixes .C, .cc, .cpp, .CPP, .c++, .cp, or .cxx; C++ header files often use .hh, .hpp, .H, or (for shared template code) .tcc; and preprocessed C++ files use the suffix .ii.  GCC recognizes files with these names and compiles them as C++ programs even if you call the compiler the same way as for compiling C programs (usually with the name gcc).

However, the use of gcc does not add the C++ library. g++ is a program that calls GCC and automatically specifies linking against the C++ library.  It treats .c, .h and .i files as C++ source files instead of C source files unless -x is used.  This program is also useful when precompiling a C header file with a .h extension for use in C++ compilations.  On many systems, g++ is also installed with the name c++.

When you compile C++ programs, you may specify many of the same command-line options that you use for compiling programs in any language; or command-line options meaningful for C and related languages; or options that are meaningful only for C++ programs.

Options Controlling C Dialect

The following options control the dialect of C (or languages derived from C, such as C++, Objective-C and Objective-C++) that the compiler accepts:


In C mode, this is equivalent to -std=c90. In C++ mode, it is equivalent to -std=c++98.

This turns off certain features of GCC that are incompatible with ISO C90 (when compiling C code), or of standard C++ (when compiling C++ code), such as the "asm" and "typeof" keywords, and predefined macros such as "unix" and "vax" that identify the type of system you are using.  It also enables the undesirable and rarely used ISO trigraph feature.  For the C compiler, it disables recognition of C++ style // comments as well as the "inline" keyword.

The alternate keywords "__asm__", "__extension__", "__inline__" and "__typeof__" continue to work despite -ansi.  You would not want to use them in an ISO C program, of course, but it is useful to put them in header files that might be included in compilations done with -ansi.  Alternate predefined macros such as "__unix__" and "__vax__" are also available, with or without -ansi.

The -ansi option does not cause non-ISO programs to be rejected gratuitously.  For that, -Wpedantic is required in addition to -ansi.

The macro "__STRICT_ANSI__" is predefined when the -ansi option is used.  Some header files may notice this macro and refrain from declaring certain functions or defining certain macros that the ISO standard doesn't call for; this is to avoid interfering with any programs that might use these names for other things.

Functions that are normally built in but do not have semantics defined by ISO C (such as "alloca" and "ffs") are not built-in functions when -ansi is used.


Determine the language standard.   This option is currently only supported when compiling C or C++.

The compiler can accept several base standards, such as c90 or c++98, and GNU dialects of those standards, such as gnu90 or gnu++98.  When a base standard is specified, the compiler accepts all programs following that standard plus those using GNU extensions that do not contradict it.  For example, -std=c90 turns off certain features of GCC that are incompatible with ISO C90, such as the "asm" and "typeof" keywords, but not other GNU extensions that do not have a meaning in ISO C90, such as omitting the middle term of a "?:" expression. On the other hand, when a GNU dialect of a standard is specified, all features supported by the compiler are enabled, even when those features change the meaning of the base standard.  As a result, some strict-conforming programs may be rejected.  The particular standard is used by -Wpedantic to identify which features are GNU extensions given that version of the standard. For example -std=gnu90 -Wpedantic warns about C++ style // comments, while -std=gnu99 -Wpedantic does not.

A value for this option must be provided; possible values are


Support all ISO C90 programs (certain GNU extensions that conflict with ISO C90 are disabled). Same as -ansi for C code.


ISO C90 as modified in amendment 1.


ISO C99.  This standard is substantially completely supported, modulo bugs and floating-point issues (mainly but not entirely relating to optional C99 features from Annexes F and G).  See <http://gcc.gnu.org/c99status.html> for more information.  The names c9x and iso9899:199x are deprecated.


ISO C11, the 2011 revision of the ISO C standard.  This standard is substantially completely supported, modulo bugs, floating-point issues (mainly but not entirely relating to optional C11 features from Annexes F and G) and the optional Annexes K (Bounds-checking interfaces) and L (Analyzability).  The name c1x is deprecated.


GNU dialect of ISO C90 (including some C99 features).


GNU dialect of ISO C99.  The name gnu9x is deprecated.


GNU dialect of ISO C11.  This is the default for C code. The name gnu1x is deprecated.


The 1998 ISO C++ standard plus the 2003 technical corrigendum and some additional defect reports. Same as -ansi for C++ code.


GNU dialect of -std=c++98.


The 2011 ISO C++ standard plus amendments. The name c++0x is deprecated.


GNU dialect of -std=c++11. The name gnu++0x is deprecated.


The 2014 ISO C++ standard plus amendments. The name c++1y is deprecated.


GNU dialect of -std=c++14. This is the default for C++ code. The name gnu++1y is deprecated.


The next revision of the ISO C++ standard, tentatively planned for 2017.  Support is highly experimental, and will almost certainly change in incompatible ways in future releases.


GNU dialect of -std=c++1z.  Support is highly experimental, and will almost certainly change in incompatible ways in future releases.


The option -fgnu89-inline tells GCC to use the traditional GNU semantics for "inline" functions when in C99 mode.

Using this option is roughly equivalent to adding the "gnu_inline" function attribute to all inline functions.

The option -fno-gnu89-inline explicitly tells GCC to use the C99 semantics for "inline" when in C99 or gnu99 mode (i.e., it specifies the default behavior). This option is not supported in -std=c90 or -std=gnu90 mode.

The preprocessor macros "__GNUC_GNU_INLINE__" and "__GNUC_STDC_INLINE__" may be used to check which semantics are in effect for "inline" functions.


ISO/IEC TS 18661-3 defines new permissible values for "FLT_EVAL_METHOD" that indicate that operations and constants with a semantic type that is an interchange or extended format should be evaluated to the precision and range of that type.  These new values are a superset of those permitted under C99/C11, which does not specify the meaning of other positive values of "FLT_EVAL_METHOD".  As such, code conforming to C11 may not have been written expecting the possibility of the new values.

-fpermitted-flt-eval-methods specifies whether the compiler should allow only the values of "FLT_EVAL_METHOD" specified in C99/C11, or the extended set of values specified in ISO/IEC TS 18661-3.

style is either "c11" or "ts-18661-3" as appropriate.

The default when in a standards compliant mode (-std=c11 or similar) is -fpermitted-flt-eval-methods=c11.  The default when in a GNU dialect (-std=gnu11 or similar) is -fpermitted-flt-eval-methods=ts-18661-3.

-aux-info filename

Output to the given filename prototyped declarations for all functions declared and/or defined in a translation unit, including those in header files.  This option is silently ignored in any language other than C.

Besides declarations, the file indicates, in comments, the origin of each declaration (source file and line), whether the declaration was implicit, prototyped or unprototyped (I, N for new or O for old, respectively, in the first character after the line number and the colon), and whether it came from a declaration or a definition (C or F, respectively, in the following character).  In the case of function definitions, a K&R-style list of arguments followed by their declarations is also provided, inside comments, after the declaration.


Accept variadic functions without named parameters.

Although it is possible to define such a function, this is not very useful as it is not possible to read the arguments.  This is only supported for C as this construct is allowed by C++.


Do not recognize "asm", "inline" or "typeof" as a keyword, so that code can use these words as identifiers.  You can use the keywords "__asm__", "__inline__" and "__typeof__" instead.  -ansi implies -fno-asm.

In C++, this switch only affects the "typeof" keyword, since "asm" and "inline" are standard keywords.  You may want to use the -fno-gnu-keywords flag instead, which has the same effect.  In C99 mode (-std=c99 or -std=gnu99), this switch only affects the "asm" and "typeof" keywords, since "inline" is a standard keyword in ISO C99.


Don't recognize built-in functions that do not begin with __builtin_ as prefix.

GCC normally generates special code to handle certain built-in functions more efficiently; for instance, calls to "alloca" may become single instructions which adjust the stack directly, and calls to "memcpy" may become inline copy loops.  The resulting code is often both smaller and faster, but since the function calls no longer appear as such, you cannot set a breakpoint on those calls, nor can you change the behavior of the functions by linking with a different library.  In addition, when a function is recognized as a built-in function, GCC may use information about that function to warn about problems with calls to that function, or to generate more efficient code, even if the resulting code still contains calls to that function.  For example, warnings are given with -Wformat for bad calls to "printf" when "printf" is built in and "strlen" is known not to modify global memory.

With the -fno-builtin-function option only the built-in function function is disabled.  function must not begin with __builtin_.  If a function is named that is not built-in in this version of GCC, this option is ignored.  There is no corresponding -fbuiltin-function option; if you wish to enable built-in functions selectively when using -fno-builtin or -ffreestanding, you may define macros such as:

        #define abs(n)          __builtin_abs ((n))
        #define strcpy(d, s)    __builtin_strcpy ((d), (s))

Enable parsing of function definitions marked with "__GIMPLE". This is an experimental feature that allows unit testing of GIMPLE passes.


Assert that compilation targets a hosted environment.  This implies -fbuiltin.  A hosted environment is one in which the entire standard library is available, and in which "main" has a return type of "int".  Examples are nearly everything except a kernel. This is equivalent to -fno-freestanding.


Assert that compilation targets a freestanding environment.  This implies -fno-builtin.  A freestanding environment is one in which the standard library may not exist, and program startup may not necessarily be at "main".  The most obvious example is an OS kernel. This is equivalent to -fno-hosted.


Enable handling of OpenACC directives "#pragma acc" in C/C++ and "!$acc" in Fortran.  When -fopenacc is specified, the compiler generates accelerated code according to the OpenACC Application Programming Interface v2.0 <http://www.openacc.org/>.  This option implies -pthread, and thus is only supported on targets that have support for -pthread.


Specify default compute dimensions for parallel offload regions that do not explicitly specify.  The geom value is a triple of ':'-separated sizes, in order 'gang', 'worker' and, 'vector'.  A size can be omitted, to use a target-specific default value.


Enable handling of OpenMP directives "#pragma omp" in C/C++ and "!$omp" in Fortran.  When -fopenmp is specified, the compiler generates parallel code according to the OpenMP Application Program Interface v4.5 <http://www.openmp.org/>.  This option implies -pthread, and thus is only supported on targets that have support for -pthread. -fopenmp implies -fopenmp-simd.


Enable handling of OpenMP's SIMD directives with "#pragma omp" in C/C++ and "!$omp" in Fortran. Other OpenMP directives are ignored.


Enable the usage of Cilk Plus language extension features for C/C++. When the option -fcilkplus is specified, enable the usage of the Cilk Plus Language extension features for C/C++.  The present implementation follows ABI version 1.2.  This is an experimental feature that is only partially complete, and whose interface may change in future versions of GCC as the official specification changes.  Currently, all features but "_Cilk_for" have been implemented.


When the option -fgnu-tm is specified, the compiler generates code for the Linux variant of Intel's current Transactional Memory ABI specification document (Revision 1.1, May 6 2009).  This is an experimental feature whose interface may change in future versions of GCC, as the official specification changes.  Please note that not all architectures are supported for this feature.

For more information on GCC's support for transactional memory,

Note that the transactional memory feature is not supported with non-call exceptions (-fnon-call-exceptions).


Accept some non-standard constructs used in Microsoft header files.

In C++ code, this allows member names in structures to be similar to previous types declarations.

        typedef int UOW;
        struct ABC {
          UOW UOW;

Some cases of unnamed fields in structures and unions are only accepted with this option.

Note that this option is off for all targets but x86  targets using ms-abi.


Accept some non-standard constructs used in Plan 9 code.

This enables -fms-extensions, permits passing pointers to structures with anonymous fields to functions that expect pointers to elements of the type of the field, and permits referring to anonymous fields declared using a typedef.    This is only supported for C, not C++.


Allow conditional expressions with mismatched types in the second and third arguments.  The value of such an expression is void.  This option is not supported for C++.


Allow implicit conversions between vectors with differing numbers of elements and/or incompatible element types.  This option should not be used for new code.


Let the type "char" be unsigned, like "unsigned char".

Each kind of machine has a default for what "char" should be.  It is either like "unsigned char" by default or like "signed char" by default.

Ideally, a portable program should always use "signed char" or "unsigned char" when it depends on the signedness of an object. But many programs have been written to use plain "char" and expect it to be signed, or expect it to be unsigned, depending on the machines they were written for.  This option, and its inverse, let you make such a program work with the opposite default.

The type "char" is always a distinct type from each of "signed char" or "unsigned char", even though its behavior is always just like one of those two.


Let the type "char" be signed, like "signed char".

Note that this is equivalent to -fno-unsigned-char, which is the negative form of -funsigned-char.  Likewise, the option -fno-signed-char is equivalent to -funsigned-char.


These options control whether a bit-field is signed or unsigned, when the declaration does not use either "signed" or "unsigned".  By default, such a bit-field is signed, because this is consistent: the basic integer types such as "int" are signed types.


Set the default scalar storage order of structures and unions to the specified endianness.  The accepted values are big-endian, little-endian and native for the native endianness of the target (the default).  This option is not supported for C++.

Warning: the -fsso-struct switch causes GCC to generate code that is not binary compatible with code generated without it if the specified endianness is not the native endianness of the target.

Options Controlling C++ Dialect

This section describes the command-line options that are only meaningful for C++ programs.  You can also use most of the GNU compiler options regardless of what language your program is in.  For example, you might compile a file firstClass.C like this:

        g++ -g -fstrict-enums -O -c firstClass.C

In this example, only -fstrict-enums is an option meant only for C++ programs; you can use the other options with any language supported by GCC.

Some options for compiling C programs, such as -std, are also relevant for C++ programs.

Here is a list of options that are only for compiling C++ programs:


Use version n of the C++ ABI.  The default is version 0.

Version 0 refers to the version conforming most closely to the C++ ABI specification.  Therefore, the ABI obtained using version 0 will change in different versions of G++ as ABI bugs are fixed.

Version 1 is the version of the C++ ABI that first appeared in G++ 3.2.

Version 2 is the version of the C++ ABI that first appeared in G++ 3.4, and was the default through G++ 4.9.

Version 3 corrects an error in mangling a constant address as a template argument.

Version 4, which first appeared in G++ 4.5, implements a standard mangling for vector types.

Version 5, which first appeared in G++ 4.6, corrects the mangling of attribute const/volatile on function pointer types, decltype of a plain decl, and use of a function parameter in the declaration of another parameter.

Version 6, which first appeared in G++ 4.7, corrects the promotion behavior of C++11 scoped enums and the mangling of template argument packs, const/static_cast, prefix ++ and --, and a class scope function used as a template argument.

Version 7, which first appeared in G++ 4.8, that treats nullptr_t as a builtin type and corrects the mangling of lambdas in default argument scope.

Version 8, which first appeared in G++ 4.9, corrects the substitution behavior of function types with function-cv-qualifiers.

Version 9, which first appeared in G++ 5.2, corrects the alignment of "nullptr_t".

Version 10, which first appeared in G++ 6.1, adds mangling of attributes that affect type identity, such as ia32 calling convention attributes (e.g. stdcall).

Version 11, which first appeared in G++ 7, corrects the mangling of sizeof... expressions and operator names.  For multiple entities with the same name within a function, that are declared in different scopes, the mangling now changes starting with the twelfth occurrence.  It also implies -fnew-inheriting-ctors.

See also -Wabi.


On targets that support strong aliases, G++ works around mangling changes by creating an alias with the correct mangled name when defining a symbol with an incorrect mangled name. This switch specifies which ABI version to use for the alias.

With -fabi-version=0 (the default), this defaults to 8 (GCC 5 compatibility).  If another ABI version is explicitly selected, this defaults to 0.  For compatibility with GCC versions 3.2 through 4.9, use -fabi-compat-version=2.

If this option is not provided but -Wabi=n is, that version is used for compatibility aliases.  If this option is provided along with -Wabi (without the version), the version from this option is used for the warning.


Turn off all access checking.  This switch is mainly useful for working around bugs in the access control code.


Enable support for C++17 "new" of types that require more alignment than "void* ::operator new(std::size_t)" provides.  A numeric argument such as "-faligned-new=32" can be used to specify how much alignment (in bytes) is provided by that function, but few users will need to override the default of "alignof(std::max_align_t)".


Check that the pointer returned by "operator new" is non-null before attempting to modify the storage allocated.  This check is normally unnecessary because the C++ standard specifies that "operator new" only returns 0 if it is declared "throw()", in which case the compiler always checks the return value even without this option.  In all other cases, when "operator new" has a non-empty exception specification, memory exhaustion is signalled by throwing "std::bad_alloc".  See also new (nothrow).


Enable support for the C++ Extensions for Concepts Technical Specification, ISO 19217 (2015), which allows code like

        template <class T> concept bool Addable = requires (T t) { t + t; };
        template <Addable T> T add (T a, T b) { return a + b; }

Set the maximum nested evaluation depth for C++11 constexpr functions to n.  A limit is needed to detect endless recursion during constant expression evaluation.  The minimum specified by the standard is 512.


Set the maximum number of iterations for a loop in C++14 constexpr functions to n.  A limit is needed to detect infinite loops during constant expression evaluation.  The default is 262144 (1<<18).


Enable deduction of a template type parameter as "std::initializer_list" from a brace-enclosed initializer list, i.e.

        template <class T> auto forward(T t) -> decltype (realfn (t))
          return realfn (t);
        void f()
          forward({1,2}); // call forward<std::initializer_list<int>>

This deduction was implemented as a possible extension to the originally proposed semantics for the C++11 standard, but was not part of the final standard, so it is disabled by default.  This option is deprecated, and may be removed in a future version of G++.


Inject friend functions into the enclosing namespace, so that they are visible outside the scope of the class in which they are declared. Friend functions were documented to work this way in the old Annotated C++ Reference Manual.   However, in ISO C++ a friend function that is not declared in an enclosing scope can only be found using argument dependent lookup.  GCC defaults to the standard behavior.

This option is for compatibility, and may be removed in a future release of G++.


The C++ standard allows an implementation to omit creating a temporary that is only used to initialize another object of the same type. Specifying this option disables that optimization, and forces G++ to call the copy constructor in all cases.  This option also causes G++ to call trivial member functions which otherwise would be expanded inline.

In C++17, the compiler is required to omit these temporaries, but this option still affects trivial member functions.


Don't generate code to check for violation of exception specifications at run time.  This option violates the C++ standard, but may be useful for reducing code size in production builds, much like defining "NDEBUG".  This does not give user code permission to throw exceptions in violation of the exception specifications; the compiler still optimizes based on the specifications, so throwing an unexpected exception results in undefined behavior at run time.


The C++11 and OpenMP standards allow "thread_local" and "threadprivate" variables to have dynamic (runtime) initialization.  To support this, any use of such a variable goes through a wrapper function that performs any necessary initialization. When the use and definition of the variable are in the same translation unit, this overhead can be optimized away, but when the use is in a different translation unit there is significant overhead even if the variable doesn't actually need dynamic initialization.  If the programmer can be sure that no use of the variable in a non-defining TU needs to trigger dynamic initialization (either because the variable is statically initialized, or a use of the variable in the defining TU will be executed before any uses in another TU), they can avoid this overhead with the -fno-extern-tls-init option.

On targets that support symbol aliases, the default is -fextern-tls-init.  On targets that do not support symbol aliases, the default is -fno-extern-tls-init.


If -ffor-scope is specified, the scope of variables declared in a for-init-statement is limited to the "for" loop itself, as specified by the C++ standard. If -fno-for-scope is specified, the scope of variables declared in a for-init-statement extends to the end of the enclosing scope, as was the case in old versions of G++, and other (traditional) implementations of C++.

If neither flag is given, the default is to follow the standard, but to allow and give a warning for old-style code that would otherwise be invalid, or have different behavior.


Do not recognize "typeof" as a keyword, so that code can use this word as an identifier.  You can use the keyword "__typeof__" instead. This option is implied by the strict ISO C++ dialects: -ansi, -std=c++98, -std=c++11, etc.


Never emit code for non-inline templates that are instantiated implicitly (i.e. by use); only emit code for explicit instantiations.


Don't emit code for implicit instantiations of inline templates, either. The default is to handle inlines differently so that compiles with and without optimization need the same set of explicit instantiations.


To save space, do not emit out-of-line copies of inline functions controlled by "#pragma implementation".  This causes linker errors if these functions are not inlined everywhere they are called.


Disable Wpedantic warnings about constructs used in MFC, such as implicit int and getting a pointer to member function via non-standard syntax.


Enable the P0136 adjustment to the semantics of C++11 constructor inheritance.  This is part of C++17 but also considered to be a Defect Report against C++11 and C++14.  This flag is enabled by default unless -fabi-version=10 or lower is specified.


Enable the P0522 resolution to Core issue 150, template template parameters and default arguments: this allows a template with default template arguments as an argument for a template template parameter with fewer template parameters.  This flag is enabled by default for -std=c++1z.


Disable built-in declarations of functions that are not mandated by ANSI/ISO C.  These include "ffs", "alloca", "_exit", "index", "bzero", "conjf", and other related functions.


Treat a "throw()" exception specification as if it were a "noexcept" specification to reduce or eliminate the text size overhead relative to a function with no exception specification.  If the function has local variables of types with non-trivial destructors, the exception specification actually makes the function smaller because the EH cleanups for those variables can be optimized away.  The semantic effect is that an exception thrown out of a function with such an exception specification results in a call to "terminate" rather than "unexpected".


Do not treat the operator name keywords "and", "bitand", "bitor", "compl", "not", "or" and "xor" as synonyms as keywords.


Disable diagnostics that the standard says a compiler does not need to issue.  Currently, the only such diagnostic issued by G++ is the one for a name having multiple meanings within a class.


Downgrade some diagnostics about nonconformant code from errors to warnings.  Thus, using -fpermissive allows some nonconforming code to compile.


When an error message refers to a specialization of a function template, the compiler normally prints the signature of the template followed by the template arguments and any typedefs or typenames in the signature (e.g. "void f(T) [with T = int]" rather than "void f(int)") so that it's clear which template is involved.  When an error message refers to a specialization of a class template, the compiler omits any template arguments that match the default template arguments for that template.  If either of these behaviors make it harder to understand the error message rather than easier, you can use -fno-pretty-templates to disable them.


Enable automatic template instantiation at link time.  This option also implies -fno-implicit-templates.


Disable generation of information about every class with virtual functions for use by the C++ run-time type identification features ("dynamic_cast" and "typeid").  If you don't use those parts of the language, you can save some space by using this flag.  Note that exception handling uses the same information, but G++ generates it as needed. The "dynamic_cast" operator can still be used for casts that do not require run-time type information, i.e. casts to "void *" or to unambiguous base classes.


Enable the built-in global declarations

        void operator delete (void *, std::size_t) noexcept;
        void operator delete[] (void *, std::size_t) noexcept;

as introduced in C++14.  This is useful for user-defined replacement deallocation functions that, for example, use the size of the object to make deallocation faster.  Enabled by default under -std=c++14 and above.  The flag -Wsized-deallocation warns about places that might want to add a definition.


Allow the compiler to optimize using the assumption that a value of enumerated type can only be one of the values of the enumeration (as defined in the C++ standard; basically, a value that can be represented in the minimum number of bits needed to represent all the enumerators).  This assumption may not be valid if the program uses a cast to convert an arbitrary integer value to the enumerated type.


Evaluate member access, array subscripting, and shift expressions in left-to-right order, and evaluate assignment in right-to-left order, as adopted for C++17.  Enabled by default with -std=c++1z. -fstrong-eval-order=some enables just the ordering of member access and shift expressions, and is the default without -std=c++1z.


Set the maximum number of template instantiation notes for a single warning or error to n.  The default value is 10.


Set the maximum instantiation depth for template classes to n. A limit on the template instantiation depth is needed to detect endless recursions during template class instantiation.  ANSI/ISO C++ conforming programs must not rely on a maximum depth greater than 17 (changed to 1024 in C++11).  The default value is 900, as the compiler can run out of stack space before hitting 1024 in some situations.


Do not emit the extra code to use the routines specified in the C++ ABI for thread-safe initialization of local statics.  You can use this option to reduce code size slightly in code that doesn't need to be thread-safe.


Register destructors for objects with static storage duration with the "__cxa_atexit" function rather than the "atexit" function. This option is required for fully standards-compliant handling of static destructors, but only works if your C library supports "__cxa_atexit".


Don't use the "__cxa_get_exception_ptr" runtime routine.  This causes "std::uncaught_exception" to be incorrect, but is necessary if the runtime routine is not available.


This switch declares that the user does not attempt to compare pointers to inline functions or methods where the addresses of the two functions are taken in different shared objects.

The effect of this is that GCC may, effectively, mark inline methods with "__attribute__ ((visibility ("hidden")))" so that they do not appear in the export table of a DSO and do not require a PLT indirection when used within the DSO.  Enabling this option can have a dramatic effect on load and link times of a DSO as it massively reduces the size of the dynamic export table when the library makes heavy use of templates.

The behavior of this switch is not quite the same as marking the methods as hidden directly, because it does not affect static variables local to the function or cause the compiler to deduce that the function is defined in only one shared object.

You may mark a method as having a visibility explicitly to negate the effect of the switch for that method.  For example, if you do want to compare pointers to a particular inline method, you might mark it as having default visibility.  Marking the enclosing class with explicit visibility has no effect.

Explicitly instantiated inline methods are unaffected by this option as their linkage might otherwise cross a shared library boundary.


This flag attempts to use visibility settings to make GCC's C++ linkage model compatible with that of Microsoft Visual Studio.

The flag makes these changes to GCC's linkage model:


It sets the default visibility to "hidden", like -fvisibility=hidden.


Types, but not their members, are not hidden by default.


The One Definition Rule is relaxed for types without explicit visibility specifications that are defined in more than one shared object: those declarations are permitted if they are permitted when this option is not used.

In new code it is better to use -fvisibility=hidden and export those classes that are intended to be externally visible. Unfortunately it is possible for code to rely, perhaps accidentally, on the Visual Studio behavior.

Among the consequences of these changes are that static data members of the same type with the same name but defined in different shared objects are different, so changing one does not change the other; and that pointers to function members defined in different shared objects may not compare equal.  When this flag is given, it is a violation of the ODR to define types with the same name differently.


Do not use weak symbol support, even if it is provided by the linker. By default, G++ uses weak symbols if they are available.  This option exists only for testing, and should not be used by end-users; it results in inferior code and has no benefits.  This option may be removed in a future release of G++.


Do not search for header files in the standard directories specific to C++, but do still search the other standard directories.  (This option is used when building the C++ library.)

In addition, these optimization, warning, and code generation options have meanings only for C++ programs:

-Wabi (C, Objective-C, C++ and Objective-C++ only)

Warn when G++ it generates code that is probably not compatible with the vendor-neutral C++ ABI.  Since G++ now defaults to updating the ABI with each major release, normally -Wabi will warn only if there is a check added later in a release series for an ABI issue discovered since the initial release.  -Wabi will warn about more things if an older ABI version is selected (with -fabi-version=n).

-Wabi can also be used with an explicit version number to warn about compatibility with a particular -fabi-version level, e.g. -Wabi=2 to warn about changes relative to -fabi-version=2.

If an explicit version number is provided and -fabi-compat-version is not specified, the version number from this option is used for compatibility aliases.  If no explicit version number is provided with this option, but -fabi-compat-version is specified, that version number is used for ABI warnings.

Although an effort has been made to warn about all such cases, there are probably some cases that are not warned about, even though G++ is generating incompatible code.  There may also be cases where warnings are emitted even though the code that is generated is compatible.

You should rewrite your code to avoid these warnings if you are concerned about the fact that code generated by G++ may not be binary compatible with code generated by other compilers.

Known incompatibilities in -fabi-version=2 (which was the default from GCC 3.4 to 4.9) include:

  • A template with a non-type template parameter of reference type was mangled incorrectly:

            extern int N;
            template <int &> struct S {};
            void n (S<N>) {2}

    This was fixed in -fabi-version=3.

  • SIMD vector types declared using "__attribute ((vector_size))" were mangled in a non-standard way that does not allow for overloading of functions taking vectors of different sizes.

    The mangling was changed in -fabi-version=4.

  • "__attribute ((const))" and "noreturn" were mangled as type qualifiers, and "decltype" of a plain declaration was folded away.

    These mangling issues were fixed in -fabi-version=5.

  • Scoped enumerators passed as arguments to a variadic function are promoted like unscoped enumerators, causing "va_arg" to complain. On most targets this does not actually affect the parameter passing ABI, as there is no way to pass an argument smaller than "int".

    Also, the ABI changed the mangling of template argument packs, "const_cast", "static_cast", prefix increment/decrement, and a class scope function used as a template argument.

    These issues were corrected in -fabi-version=6.

  • Lambdas in default argument scope were mangled incorrectly, and the ABI changed the mangling of "nullptr_t".

    These issues were corrected in -fabi-version=7.

  • When mangling a function type with function-cv-qualifiers, the un-qualified function type was incorrectly treated as a substitution candidate.

    This was fixed in -fabi-version=8, the default for GCC 5.1.

  • "decltype(nullptr)" incorrectly had an alignment of 1, leading to unaligned accesses.  Note that this did not affect the ABI of a function with a "nullptr_t" parameter, as parameters have a minimum alignment.

    This was fixed in -fabi-version=9, the default for GCC 5.2.

  • Target-specific attributes that affect the identity of a type, such as ia32 calling conventions on a function type (stdcall, regparm, etc.), did not affect the mangled name, leading to name collisions when function pointers were used as template arguments.

    This was fixed in -fabi-version=10, the default for GCC 6.1.

It also warns about psABI-related changes.  The known psABI changes at this point include:

  • For SysV/x86-64, unions with "long double" members are  passed in memory as specified in psABI.  For example:

            union U {
              long double ld;
              int i;

    "union U" is always passed in memory.

-Wabi-tag (C++ and Objective-C++ only)

Warn when a type with an ABI tag is used in a context that does not have that ABI tag.  See C++ Attributes for more information about ABI tags.

-Wctor-dtor-privacy (C++ and Objective-C++ only)

Warn when a class seems unusable because all the constructors or destructors in that class are private, and it has neither friends nor public static member functions.  Also warn if there are no non-private methods, and there's at least one private member function that isn't a constructor or destructor.

-Wdelete-non-virtual-dtor (C++ and Objective-C++ only)

Warn when "delete" is used to destroy an instance of a class that has virtual functions and non-virtual destructor. It is unsafe to delete an instance of a derived class through a pointer to a base class if the base class does not have a virtual destructor.  This warning is enabled by -Wall.

-Wliteral-suffix (C++ and Objective-C++ only)

Warn when a string or character literal is followed by a ud-suffix which does not begin with an underscore.  As a conf