Clang command line argument reference¶

Introduction ¶

This page lists the command line arguments currently supported by the GCC-compatible clang and clang++ drivers.

-B<prefix>, --prefix <arg>, --prefix=<arg>¶

Search $prefix/$triple-$file and $prefix$file for executables, libraries, includes, and data files used by the compiler. $prefix may or may not be a directory

-F<arg>¶

Add directory to framework include search path

-ObjC¶

Treat source input files as Objective-C inputs

-ObjC++¶

Treat source input files as Objective-C++ inputs

-Qn, -fno-ident¶

Do not emit metadata containing compiler name and version

-Qunused-arguments¶

Don’t emit warning for unused driver arguments

-Qy, -fident¶

Emit metadata containing compiler name and version

-Wa,<arg>,<arg2>...¶

Pass the comma separated arguments in <arg> to the assembler

-Wlarge-by-value-copy=<arg>¶

-Xarch_<arg1> <arg2>¶

-Xarch_device <arg>¶

Pass <arg> to the CUDA/HIP device compilation

-Xarch_host <arg>¶

Pass <arg> to the CUDA/HIP host compilation

-Xcuda-fatbinary <arg>¶

Pass <arg> to fatbinary invocation

-Xcuda-ptxas <arg>¶

Pass <arg> to the ptxas assembler

-Xopenmp-target <arg>¶

Pass <arg> to the target offloading toolchain.

-Xopenmp-target=<triple> <arg>¶

Pass <arg> to the target offloading toolchain identified by <triple>.

-Z<arg>¶

-a<arg>, --profile-blocks¶

-all_load¶

-allowable_client <arg>¶

--analyze¶

Run the static analyzer

--analyzer-no-default-checks¶

--analyzer-output<arg>¶

-arch <arg>¶

-arch_errors_fatal¶

-arch_only <arg>¶

-arcmt-migrate-emit-errors¶

Emit ARC errors even if the migrator can fix them

-arcmt-migrate-report-output <arg>¶

Output path for the plist report

--autocomplete=<arg>¶

-bind_at_load¶

-bundle¶

-bundle_loader <arg>¶

-client_name<arg>¶

-compatibility_version<arg>¶

--config <arg>¶

Specifies configuration file

--constant-cfstrings¶

--cuda-compile-host-device¶

Compile CUDA code for both host and device (default). Has no effect on non-CUDA compilations.

--cuda-device-only¶

Compile CUDA code for device only

--cuda-host-only¶

Compile CUDA code for host only. Has no effect on non-CUDA compilations.

--cuda-include-ptx=<arg>, --no-cuda-include-ptx=<arg>¶

Include PTX for the following GPU architecture (e.g. sm_35) or ‘all’. May be specified more than once.

--cuda-noopt-device-debug, --no-cuda-noopt-device-debug¶

Enable device-side debug info generation. Disables ptxas optimizations.

-cuid=<arg>¶

An ID for compilation unit, which should be the same for the same compilation unit but different for different compilation units. It is used to externalize device-side static variables for single source offloading languages CUDA and HIP so that they can be accessed by the host code of the same compilation unit.

-current_version<arg>¶

-dead_strip¶

-dependency-dot <arg>¶

Filename to write DOT-formatted header dependencies to

-dependency-file <arg>¶

Filename (or -) to write dependency output to

-dsym-dir<dir>¶

Directory to output dSYM’s (if any) to

-dumpmachine¶

-dumpversion¶

--dyld-prefix=<arg>, --dyld-prefix <arg>¶

-dylib_file <arg>¶

-dylinker¶

-dylinker_install_name<arg>¶

-dynamic¶

-dynamiclib¶

-emit-ast¶

Emit Clang AST files for source inputs

--emit-static-lib¶

Enable linker job to emit a static library.

-enable-trivial-auto-var-init-zero-knowing-it-will-be-removed-from-clang¶

Trivial automatic variable initialization to zero is only here for benchmarks, it’ll eventually be removed, and I’m OK with that because I’m only using it to benchmark

-exported_symbols_list <arg>¶

-faligned-new=<arg>¶

-ffixed-r19¶

Reserve register r19 (Hexagon only)

-fgpu-flush-denormals-to-zero, -fcuda-flush-denormals-to-zero, -fno-gpu-flush-denormals-to-zero¶

Flush denormal floating point values to zero in CUDA/HIP device mode.

-fheinous-gnu-extensions¶

-flat_namespace¶

-fopenmp-targets=<arg1>,<arg2>...¶

Specify comma-separated list of triples OpenMP offloading targets to be supported

-force_cpusubtype_ALL¶

-force_flat_namespace¶

-force_load <arg>¶

-framework <arg>¶

-frtlib-add-rpath, -fno-rtlib-add-rpath¶

Add -rpath with architecture-specific resource directory to the linker flags

-fsanitize-system-ignorelist=<arg>, -fsanitize-system-blacklist=<arg>¶

Path to system ignorelist file for sanitizers

-fshow-skipped-includes¶

#include files may be “skipped” due to include guard optimization: or #pragma once. This flag makes -H show also such includes.

-fsystem-module¶

Build this module as a system module. Only used with -emit-module

-fuse-cuid=<arg>¶

Method to generate ID’s for compilation units for single source offloading languages CUDA and HIP: ‘hash’ (ID’s generated by hashing file path and command line options) | ‘random’ (ID’s generated as random numbers) | ‘none’ (disabled). Default is ‘hash’. This option will be overriden by option ‘-cuid=[ID]’ if it is specified.

--gcc-toolchain=<arg>, -gcc-toolchain <arg>¶

Search for GCC installation in the specified directory on targets which commonly use GCC. The directory usually contains ‘lib{,32,64}/gcc{,-cross}/$triple’ and ‘include’. If specified, sysroot is skipped for GCC detection. Note: executables (e.g. ld) used by the compiler are not overridden by the selected GCC installation

-gcodeview¶

Generate CodeView debug information

-gcodeview-ghash, -gno-codeview-ghash¶

Emit type record hashes in a .debug$H section

--gpu-instrument-lib=<arg>¶

Instrument device library for HIP, which is a LLVM bitcode containing __cyg_profile_func_enter and __cyg_profile_func_exit

--gpu-max-threads-per-block=<arg>¶

Default max threads per block for kernel launch bounds for HIP

-headerpad_max_install_names<arg>¶

-help, --help¶

Display available options

--help-hidden¶

Display help for hidden options

--hip-link¶

Link clang-offload-bundler bundles for HIP

--hip-version=<arg>¶

HIP version in the format of major.minor.patch

-ibuiltininc¶

Enable builtin #include directories even when -nostdinc is used before or after -ibuiltininc. Using -nobuiltininc after the option disables it

-image_base <arg>¶

-index-header-map¶

Make the next included directory (-I or -F) an indexer header map

-init <arg>¶

-install_name <arg>¶

-interface-stub-version=<arg>¶

-keep_private_externs¶

-lazy_framework <arg>¶

-lazy_library <arg>¶

-mbig-endian, -EB¶

-mbranch-protection=<arg>¶

Enforce targets of indirect branches and function returns

-menable-unsafe-fp-math¶

Allow unsafe floating-point math optimizations which may decrease precision

-mharden-sls=<arg>¶

Select straight-line speculation hardening scope

--migrate¶

Run the migrator

-mios-simulator-version-min=<arg>, -miphonesimulator-version-min=<arg>¶

-mlinker-version=<arg>¶

-mlittle-endian, -EL¶

-mllvm <arg>¶

Additional arguments to forward to LLVM’s option processing

-module-dependency-dir <arg>¶

Directory to dump module dependencies to

-mtvos-simulator-version-min=<arg>, -mappletvsimulator-version-min=<arg>¶

-multi_module¶

-multiply_defined <arg>¶

-multiply_defined_unused <arg>¶

-mwatchos-simulator-version-min=<arg>, -mwatchsimulator-version-min=<arg>¶

--no-cuda-version-check¶

Don’t error out if the detected version of the CUDA install is too low for the requested CUDA gpu architecture.

-no-integrated-cpp, --no-integrated-cpp¶

-no_dead_strip_inits_and_terms¶

-nobuiltininc¶

Disable builtin #include directories

-nodefaultlibs¶

-nofixprebinding¶

-nogpuinc, -nocudainc¶

Do not add include paths for CUDA/HIP and do not include the default CUDA/HIP wrapper headers

-nogpulib, -nocudalib¶

Do not link device library for CUDA/HIP device compilation

-nolibc¶

-nomultidefs¶

-nopie, -no-pie¶

-noprebind¶

-noprofilelib¶

-noseglinkedit¶

-nostdinc, --no-standard-includes¶

-nostdinc++¶

Disable standard #include directories for the C++ standard library

-nostdlib++¶

-nostdlibinc¶

-o<file>, --output <arg>, --output=<arg>¶

Write output to <file>

-objcmt-atomic-property¶

Make migration to ‘atomic’ properties

-objcmt-migrate-all¶

Enable migration to modern ObjC

-objcmt-migrate-annotation¶

Enable migration to property and method annotations

-objcmt-migrate-designated-init¶

Enable migration to infer NS_DESIGNATED_INITIALIZER for initializer methods

-objcmt-migrate-instancetype¶

Enable migration to infer instancetype for method result type

-objcmt-migrate-literals¶

Enable migration to modern ObjC literals

-objcmt-migrate-ns-macros¶

Enable migration to NS_ENUM/NS_OPTIONS macros

-objcmt-migrate-property¶

Enable migration to modern ObjC property

-objcmt-migrate-property-dot-syntax¶

Enable migration of setter/getter messages to property-dot syntax

-objcmt-migrate-protocol-conformance¶

Enable migration to add protocol conformance on classes

-objcmt-migrate-readonly-property¶

Enable migration to modern ObjC readonly property

-objcmt-migrate-readwrite-property¶

Enable migration to modern ObjC readwrite property

-objcmt-migrate-subscripting¶

Enable migration to modern ObjC subscripting

-objcmt-ns-nonatomic-iosonly¶

Enable migration to use NS_NONATOMIC_IOSONLY macro for setting property’s ‘atomic’ attribute

-objcmt-returns-innerpointer-property¶

Enable migration to annotate property with NS_RETURNS_INNER_POINTER

-objcmt-whitelist-dir-path=<arg>, -objcmt-white-list-dir-path=<arg>¶

Only modify files with a filename contained in the provided directory path

-object¶

--offload-arch=<arg>, --cuda-gpu-arch=<arg>, --no-offload-arch=<arg>¶

CUDA offloading device architecture (e.g. sm_35), or HIP offloading target ID in the form of a device architecture followed by target ID features delimited by a colon. Each target ID feature is a pre-defined string followed by a plus or minus sign (e.g. gfx908:xnack+:sramecc-). May be specified more than once.

-p, --profile¶

-pagezero_size<arg>¶

-pg¶

Enable mcount instrumentation

-pipe, --pipe¶

Use pipes between commands, when possible

-prebind¶

-prebind_all_twolevel_modules¶

-preload¶

--print-diagnostic-categories¶

-print-effective-triple, --print-effective-triple¶

Print the effective target triple

-print-file-name=<file>, --print-file-name=<file>, --print-file-name <arg>¶

Print the full library path of <file>

-print-ivar-layout¶

Enable Objective-C Ivar layout bitmap print trace

-print-libgcc-file-name, --print-libgcc-file-name¶

Print the library path for the currently used compiler runtime library (“libgcc.a” or “libclang_rt.builtins.*.a”)

-print-multi-directory, --print-multi-directory¶

-print-multi-lib, --print-multi-lib¶

-print-multiarch, --print-multiarch¶

Print the multiarch target triple

-print-prog-name=<name>, --print-prog-name=<name>, --print-prog-name <arg>¶

Print the full program path of <name>

-print-resource-dir, --print-resource-dir¶

Print the resource directory pathname

-print-rocm-search-dirs, --print-rocm-search-dirs¶

Print the paths used for finding ROCm installation

-print-runtime-dir, --print-runtime-dir¶

Print the directory pathname containing clangs runtime libraries

-print-search-dirs, --print-search-dirs¶

Print the paths used for finding libraries and programs

-print-target-triple, --print-target-triple¶

Print the normalized target triple

-print-targets, --print-targets¶

Print the registered targets

-private_bundle¶

-pthread, -no-pthread¶

Support POSIX threads in generated code

-pthreads¶

-read_only_relocs <arg>¶

-relocatable-pch, --relocatable-pch¶

Whether to build a relocatable precompiled header

-remap¶

-rewrite-legacy-objc¶

Rewrite Legacy Objective-C source to C++

-rtlib=<arg>, --rtlib=<arg>, --rtlib <arg>¶

Compiler runtime library to use

-save-stats=<arg>, --save-stats=<arg>, -save-stats (equivalent to -save-stats=cwd), --save-stats (equivalent to -save-stats=cwd)¶

Save llvm statistics.

-save-temps=<arg>, --save-temps=<arg>, -save-temps (equivalent to -save-temps=cwd), --save-temps (equivalent to -save-temps=cwd)¶

Save intermediate compilation results.

-sectalign <arg1> <arg2> <arg3>¶

-sectcreate <arg1> <arg2> <arg3>¶

-sectobjectsymbols <arg1> <arg2>¶

-sectorder <arg1> <arg2> <arg3>¶

-seg1addr<arg>¶

-seg_addr_table <arg>¶

-seg_addr_table_filename <arg>¶

-segaddr <arg1> <arg2>¶

-segcreate <arg1> <arg2> <arg3>¶

-seglinkedit¶

-segprot <arg1> <arg2> <arg3>¶

-segs_read_<arg>¶

-segs_read_only_addr <arg>¶

-segs_read_write_addr <arg>¶

-serialize-diagnostics <arg>, --serialize-diagnostics <arg>¶

Serialize compiler diagnostics to a file

-shared-libgcc¶

-shared-libsan, -shared-libasan¶

Dynamically link the sanitizer runtime

-single_module¶

-static-libgcc¶

-static-libsan¶

Statically link the sanitizer runtime

-static-libstdc++¶

-static-openmp¶

Use the static host OpenMP runtime while linking.

-std-default=<arg>¶

-stdlib=<arg>, --stdlib=<arg>, --stdlib <arg>¶

C++ standard library to use

-sub_library<arg>¶

-sub_umbrella<arg>¶

--sysroot=<arg>, --sysroot <arg>¶

--target-help¶

--target=<arg>, -target <arg>¶

Generate code for the given target

-time¶

Time individual commands

-traditional, --traditional¶

-traditional-cpp, --traditional-cpp¶

Enable some traditional CPP emulation

-twolevel_namespace¶

-twolevel_namespace_hints¶

-umbrella <arg>¶

-unexported_symbols_list <arg>¶

-unwindlib=<arg>, --unwindlib=<arg>¶

Unwind library to use

-v, --verbose¶

Show commands to run and use verbose output

--verify-debug-info¶

Verify the binary representation of debug output

--version¶

Print version information

-w, --no-warnings¶

Suppress all warnings

-weak-l<arg>¶

-weak_framework <arg>¶

-weak_library <arg>¶

-weak_reference_mismatches <arg>¶

-whatsloaded¶

-why_load, -whyload¶

-working-directory<arg>, -working-directory=<arg>¶

Resolve file paths relative to the specified directory

-x<language>, --language <arg>, --language=<arg>¶

Treat subsequent input files as having type <language>

-y<arg>¶

Actions ¶

The action to perform on the input.

-E, --preprocess¶

Only run the preprocessor

-S, --assemble¶

Only run preprocess and compilation steps

-c, --compile¶

Only run preprocess, compile, and assemble steps

-emit-interface-stubs¶

Generate Interface Stub Files.

-emit-llvm¶

Use the LLVM representation for assembler and object files

-emit-merged-ifs¶

Generate Interface Stub Files, emit merged text not binary.

-fsyntax-only¶

-module-file-info¶

Provide information about a particular module file

--precompile¶

Only precompile the input

-rewrite-objc¶

Rewrite Objective-C source to C++

-verify-pch¶

Load and verify that a pre-compiled header file is not stale

Compilation flags ¶

Flags controlling the behavior of Clang during compilation. These flags have no effect during actions that do not perform compilation.

-Xassembler <arg>¶

Pass <arg> to the assembler

-Xclang <arg>¶

Pass <arg> to the clang compiler

-ansi, --ansi¶

-fc++-abi=<arg>¶

C++ ABI to use. This will override the target C++ ABI.

-fclang-abi-compat=<version>¶

Attempt to match the ABI of Clang <version>

-fcomment-block-commands=<arg>,<arg2>...¶

Treat each comma separated argument in <arg> as a documentation comment block command

-fcomplete-member-pointers, -fno-complete-member-pointers¶

Require member pointer base types to be complete if they would be significant under the Microsoft ABI

-fcrash-diagnostics-dir=<dir>¶

Put crash-report files in <dir>

-fdeclspec, -fno-declspec¶

Allow __declspec as a keyword

-fdepfile-entry=<arg>¶

-fdiagnostics-fixit-info, -fno-diagnostics-fixit-info¶

-fdiagnostics-format=<arg>¶

-fdiagnostics-parseable-fixits¶

Print fix-its in machine parseable form

-fdiagnostics-print-source-range-info¶

Print source range spans in numeric form

-fdiagnostics-show-category=<arg>¶

-fdiscard-value-names, -fno-discard-value-names¶

Discard value names in LLVM IR

-fexperimental-relative-c++-abi-vtables, -fno-experimental-relative-c++-abi-vtables¶

Use the experimental C++ class ABI for classes with virtual tables

-fexperimental-strict-floating-point¶

Enables experimental strict floating point in LLVM.

-ffine-grained-bitfield-accesses, -fno-fine-grained-bitfield-accesses¶

Use separate accesses for consecutive bitfield runs with legal widths and alignments.

-fglobal-isel, -fexperimental-isel, -fno-global-isel¶

Enables the global instruction selector

-finline-functions, -fno-inline-functions¶

Inline suitable functions

-finline-hint-functions¶

Inline functions which are (explicitly or implicitly) marked inline

-flegacy-pass-manager, -fno-experimental-new-pass-manager, -fno-legacy-pass-manager¶

Use the legacy pass manager in LLVM

-fno-crash-diagnostics¶

Disable auto-generation of preprocessed source files and a script for reproduction during a clang crash

-fno-sanitize-ignorelist, -fno-sanitize-blacklist¶

Don’t use ignorelist file for sanitizers

-fparse-all-comments¶

-frecord-command-line, -fno-record-command-line, -frecord-gcc-switches¶

-fsanitize-address-destructor=<arg>¶

Set destructor type used in ASan instrumentation

-fsanitize-address-field-padding=<arg>¶

Level of field padding for AddressSanitizer

-fsanitize-address-globals-dead-stripping¶

Enable linker dead stripping of globals in AddressSanitizer

-fsanitize-address-outline-instrumentation, -fno-sanitize-address-outline-instrumentation¶

Always generate function calls for address sanitizer instrumentation

-fsanitize-address-poison-custom-array-cookie, -fno-sanitize-address-poison-custom-array-cookie¶

Enable poisoning array cookies when using custom operator new[] in AddressSanitizer

-fsanitize-address-use-after-return=<mode>¶

Select the mode of detecting stack use-after-return in AddressSanitizer: never | runtime (default) | always

-fsanitize-address-use-after-scope, -fno-sanitize-address-use-after-scope¶

Enable use-after-scope detection in AddressSanitizer

-fsanitize-address-use-odr-indicator, -fno-sanitize-address-use-odr-indicator¶

Enable ODR indicator globals to avoid false ODR violation reports in partially sanitized programs at the cost of an increase in binary size

-fsanitize-cfi-canonical-jump-tables, -fno-sanitize-cfi-canonical-jump-tables¶

Make the jump table addresses canonical in the symbol table

-fsanitize-cfi-cross-dso, -fno-sanitize-cfi-cross-dso¶

Enable control flow integrity (CFI) checks for cross-DSO calls.

-fsanitize-cfi-icall-generalize-pointers¶

Generalize pointers in CFI indirect call type signature checks

-fsanitize-coverage-allowlist=<arg>, -fsanitize-coverage-whitelist=<arg>¶

Restrict sanitizer coverage instrumentation exclusively to modules and functions that match the provided special case list, except the blocked ones

-fsanitize-coverage-ignorelist=<arg>, -fsanitize-coverage-blacklist=<arg>¶

Disable sanitizer coverage instrumentation for modules and functions that match the provided special case list, even the allowed ones

-fsanitize-coverage=<arg1>,<arg2>..., -fno-sanitize-coverage=<arg1>,<arg2>...¶

Specify the type of coverage instrumentation for Sanitizers

-fsanitize-hwaddress-abi=<arg>¶

Select the HWAddressSanitizer ABI to target (interceptor or platform, default interceptor). This option is currently unused.

-fsanitize-hwaddress-experimental-aliasing, -fno-sanitize-hwaddress-experimental-aliasing¶

Enable aliasing mode in HWAddressSanitizer

-fsanitize-ignorelist=<arg>, -fsanitize-blacklist=<arg>¶

Path to ignorelist file for sanitizers

-fsanitize-link-c++-runtime, -fno-sanitize-link-c++-runtime¶

-fsanitize-link-runtime, -fno-sanitize-link-runtime¶

-fsanitize-memory-track-origins, -fno-sanitize-memory-track-origins¶

Enable origins tracking in MemorySanitizer

-fsanitize-memory-track-origins=<arg>¶

Enable origins tracking in MemorySanitizer

-fsanitize-memory-use-after-dtor, -fno-sanitize-memory-use-after-dtor¶

Enable use-after-destroy detection in MemorySanitizer

-fsanitize-minimal-runtime, -fno-sanitize-minimal-runtime¶

-fsanitize-recover=<arg1>,<arg2>..., -fno-sanitize-recover=<arg1>,<arg2>..., -fsanitize-recover (equivalent to -fsanitize-recover=all)¶

Enable recovery for specified sanitizers

-fsanitize-stats, -fno-sanitize-stats¶

Enable sanitizer statistics gathering.

-fsanitize-thread-atomics, -fno-sanitize-thread-atomics¶

Enable atomic operations instrumentation in ThreadSanitizer (default)

-fsanitize-thread-func-entry-exit, -fno-sanitize-thread-func-entry-exit¶

Enable function entry/exit instrumentation in ThreadSanitizer (default)

-fsanitize-thread-memory-access, -fno-sanitize-thread-memory-access¶

Enable memory access instrumentation in ThreadSanitizer (default)

-fsanitize-trap=<arg1>,<arg2>..., -fno-sanitize-trap=<arg1>,<arg2>..., -fsanitize-trap (equivalent to -fsanitize-trap=all), -fsanitize-undefined-trap-on-error (equivalent to -fsanitize-trap=undefined)¶

Enable trapping for specified sanitizers

-fsanitize-undefined-strip-path-components=<number>¶

Strip (or keep only, if negative) a given number of path components when emitting check metadata.

-fsanitize=<check>,<arg2>..., -fno-sanitize=<arg1>,<arg2>...¶

Turn on runtime checks for various forms of undefined or suspicious behavior. See user manual for available checks

-moutline, -mno-outline¶

Enable function outlining (AArch64 only)

-moutline-atomics, -mno-outline-atomics¶

Generate local calls to out-of-line atomic operations

--param <arg>, --param=<arg>¶

-print-supported-cpus, --print-supported-cpus, -mcpu=?, -mtune=?¶

Print supported cpu models for the given target (if target is not specified, it will print the supported cpus for the default target)

-std=<arg>, --std=<arg>, --std <arg>¶

Language standard to compile for

Preprocessor flags ¶

Flags controlling the behavior of the Clang preprocessor.

-C, --comments¶

Include comments in preprocessed output

-CC, --comments-in-macros¶

Include comments from within macros in preprocessed output

-D<macro>=<value>, --define-macro <arg>, --define-macro=<arg>¶

Define <macro> to <value> (or 1 if <value> omitted)

-H, --trace-includes¶

Show header includes and nesting depth

-P, --no-line-commands¶

Disable linemarker output in -E mode

-U<macro>, --undefine-macro <arg>, --undefine-macro=<arg>¶

Undefine macro <macro>

-Wp,<arg>,<arg2>...¶

Pass the comma separated arguments in <arg> to the preprocessor

-Xpreprocessor <arg>¶

Pass <arg> to the preprocessor

Include path management ¶

Flags controlling how #includes are resolved to files.

-I<dir>, --include-directory <arg>, --include-directory=<arg>¶

Add directory to include search path. For C++ inputs, if there are multiple -I options, these directories are searched in the order they are given before the standard system directories are searched. If the same directory is in the SYSTEM include search paths, for example if also specified with -isystem, the -I option will be ignored

-I-, --include-barrier¶

Restrict all prior -I flags to double-quoted inclusion and remove current directory from include path

--amdgpu-arch-tool=<arg>¶

Tool used for detecting AMD GPU arch in the system.

--cuda-path-ignore-env¶

Ignore environment variables to detect CUDA installation

--cuda-path=<arg>¶

CUDA installation path

-cxx-isystem<directory>¶

Add directory to the C++ SYSTEM include search path

-fbuild-session-file=<file>¶

Use the last modification time of <file> as the build session timestamp

-fbuild-session-timestamp=<time since Epoch in seconds>¶

Time when the current build session started

-fmodule-file=[<name>=]<file>¶

Specify the mapping of module name to precompiled module file, or load a module file if name is omitted.

-fmodules-cache-path=<directory>¶

Specify the module cache path

-fmodules-disable-diagnostic-validation¶

Disable validation of the diagnostic options when loading the module

-fmodules-prune-after=<seconds>¶

Specify the interval (in seconds) after which a module file will be considered unused

-fmodules-prune-interval=<seconds>¶

Specify the interval (in seconds) between attempts to prune the module cache

-fmodules-user-build-path <directory>¶

Specify the module user build path

-fmodules-validate-once-per-build-session¶

Don’t verify input files for the modules if the module has been successfully validated or loaded during this build session

-fmodules-validate-system-headers, -fno-modules-validate-system-headers¶

Validate the system headers that a module depends on when loading the module

-fprebuilt-module-path=<directory>¶

Specify the prebuilt module path

--hip-path=<arg>¶

HIP runtime installation path, used for finding HIP version and adding HIP include path.

-idirafter<arg>, --include-directory-after <arg>, --include-directory-after=<arg>¶

Add directory to AFTER include search path

-iframework<arg>¶

Add directory to SYSTEM framework search path

-iframeworkwithsysroot<directory>¶

Add directory to SYSTEM framework search path, absolute paths are relative to -isysroot

-imacros<file>, --imacros<file>, --imacros=<arg>¶

Include macros from file before parsing

-include<file>, --include<file>, --include=<arg>¶

Include file before parsing

-include-pch <file>¶

Include precompiled header file

-iprefix<dir>, --include-prefix <arg>, --include-prefix=<arg>¶

Set the -iwithprefix/-iwithprefixbefore prefix

-iquote<directory>¶

Add directory to QUOTE include search path

-isysroot<dir>¶

Set the system root directory (usually /)

-isystem<directory>¶

Add directory to SYSTEM include search path

-isystem-after<directory>¶

Add directory to end of the SYSTEM include search path

-ivfsoverlay<arg>¶

Overlay the virtual filesystem described by file over the real file system

-iwithprefix<dir>, --include-with-prefix <arg>, --include-with-prefix-after <arg>, --include-with-prefix-after=<arg>, --include-with-prefix=<arg>¶

Set directory to SYSTEM include search path with prefix

-iwithprefixbefore<dir>, --include-with-prefix-before <arg>, --include-with-prefix-before=<arg>¶

Set directory to include search path with prefix

-iwithsysroot<directory>¶

Add directory to SYSTEM include search path, absolute paths are relative to -isysroot

--libomptarget-amdgcn-bc-path=<arg>¶

Path to libomptarget-amdgcn bitcode library

--libomptarget-nvptx-bc-path=<arg>¶

Path to libomptarget-nvptx bitcode library

--ptxas-path=<arg>¶

Path to ptxas (used for compiling CUDA code)

--rocm-path=<arg>¶

ROCm installation path, used for finding and automatically linking required bitcode libraries.

-stdlib++-isystem<directory>¶

Use directory as the C++ standard library include path

--system-header-prefix=<prefix>, --no-system-header-prefix=<prefix>, --system-header-prefix <arg>¶

Treat all #include paths starting with <prefix> as including a system header.

Dependency file generation ¶

Flags controlling generation of a dependency file for make-like build systems.

-M, --dependencies¶

Like -MD, but also implies -E and writes to stdout by default

-MD, --write-dependencies¶

Write a depfile containing user and system headers

-MF<file>¶

Write depfile output from -MMD, -MD, -MM, or -M to <file>

-MG, --print-missing-file-dependencies¶

Add missing headers to depfile

-MJ<arg>¶

Write a compilation database entry per input

-MM, --user-dependencies¶

Like -MMD, but also implies -E and writes to stdout by default

-MMD, --write-user-dependencies¶

Write a depfile containing user headers

-MP¶

Create phony target for each dependency (other than main file)

-MQ<arg>¶

Specify name of main file output to quote in depfile

-MT<arg>¶

Specify name of main file output in depfile

-MV¶

Use NMake/Jom format for the depfile

Dumping preprocessor state ¶

Flags allowing the state of the preprocessor to be dumped in various ways.

-d¶

-d<arg>¶

-dD¶

Print macro definitions in -E mode in addition to normal output

-dI¶

Print include directives in -E mode in addition to normal output

-dM¶

Print macro definitions in -E mode instead of normal output

Diagnostic flags ¶

Flags controlling which warnings, errors, and remarks Clang will generate. See the full list of warning and remark flags.

-R<remark>¶

Enable the specified remark

-Rpass-analysis=<arg>¶

Report transformation analysis from optimization passes whose name matches the given POSIX regular expression

-Rpass-missed=<arg>¶

Report missed transformations by optimization passes whose name matches the given POSIX regular expression

-Rpass=<arg>¶

Report transformations performed by optimization passes whose name matches the given POSIX regular expression

-W<warning>, --extra-warnings, --warn-<arg>, --warn-=<arg>¶

Enable the specified warning

-Wdeprecated, -Wno-deprecated¶

Enable warnings for deprecated constructs and define __DEPRECATED

-Wframe-larger-than=<arg>, -Wframe-larger-than¶

-Wnonportable-cfstrings<arg>, -Wno-nonportable-cfstrings<arg>¶

Target-independent compilation options ¶

-fPIC, -fno-PIC¶

-fPIE, -fno-PIE¶

-faccess-control, -fno-access-control¶

-faddrsig, -fno-addrsig¶

Emit an address-significance table

-falign-functions, -fno-align-functions¶

-falign-functions=<arg>¶

-faligned-allocation, -faligned-new, -fno-aligned-allocation¶

Enable C++17 aligned allocation functions

-fallow-editor-placeholders, -fno-allow-editor-placeholders¶

Treat editor placeholders as valid source code

-fallow-unsupported¶

-faltivec, -fno-altivec¶

-faltivec-src-compat=<arg>¶

Source-level compatibility for Altivec vectors (for PowerPC targets). This includes results of vector comparison (scalar for ‘xl’, vector for ‘gcc’) as well as behavior when initializing with a scalar (splatting for ‘xl’, element zero only for ‘gcc’). For ‘mixed’, the compatibility is as ‘gcc’ for ‘vector bool/vector pixel’ and as ‘xl’ for other types. Current default is ‘mixed’.

-fansi-escape-codes¶

Use ANSI escape codes for diagnostics

-fapple-kext, -findirect-virtual-calls, -fterminated-vtables¶

Use Apple’s kernel extensions ABI

-fapple-link-rtlib¶

Force linking the clang builtins runtime library

-fapple-pragma-pack, -fno-apple-pragma-pack¶

Enable Apple gcc-compatible #pragma pack handling

-fapplication-extension, -fno-application-extension¶

Restrict code to those available for App Extensions

-fasm, -fno-asm¶

-fasm-blocks, -fno-asm-blocks¶

-fassociative-math, -fno-associative-math¶

-fassume-sane-operator-new, -fno-assume-sane-operator-new¶

-fast¶

-fastcp¶

-fastf¶

-fasync-exceptions, -fno-async-exceptions¶

Enable EH Asynchronous exceptions

-fasynchronous-unwind-tables, -fno-asynchronous-unwind-tables¶

-fautolink, -fno-autolink¶

-fbasic-block-sections=<arg>¶

Generate labels for each basic block or place each basic block or a subset of basic blocks in its own section.

-fbinutils-version=<major.minor>¶

Produced object files can use all ELF features supported by this binutils version and newer. If -fno-integrated-as is specified, the generated assembly will consider GNU as support. ‘none’ means that all ELF features can be used, regardless of binutils support. Defaults to 2.26.

-fblocks, -fno-blocks¶

Enable the ‘blocks’ language feature

-fbootclasspath=<arg>, --bootclasspath <arg>, --bootclasspath=<arg>¶

-fborland-extensions, -fno-borland-extensions¶

Accept non-standard constructs supported by the Borland compiler

-fbracket-depth=<arg>¶

-fbuiltin, -fno-builtin¶

-fbuiltin-module-map¶

Load the clang builtins module map file.

-fc++-static-destructors, -fno-c++-static-destructors¶

-fcaret-diagnostics, -fno-caret-diagnostics¶

-fcf-protection=<arg>, -fcf-protection (equivalent to -fcf-protection=full)¶

Instrument control-flow architecture protection. Options: return, branch, full, none.

-fcf-runtime-abi=<arg>¶

-fchar8_t, -fno-char8_t¶

Enable C++ builtin type char8_t

-fclasspath=<arg>, --CLASSPATH <arg>, --CLASSPATH=<arg>, --classpath <arg>, --classpath=<arg>¶

-fcolor-diagnostics, -fno-color-diagnostics¶

Enable colors in diagnostics

-fcommon, -fno-common¶

Place uninitialized global variables in a common block

-fcompile-resource=<arg>, --resource <arg>, --resource=<arg>¶

-fconstant-cfstrings, -fno-constant-cfstrings¶

-fconstant-string-class=<arg>¶

-fconstexpr-backtrace-limit=<arg>¶

-fconstexpr-depth=<arg>¶

-fconstexpr-steps=<arg>¶

-fconvergent-functions¶

Assume functions may be convergent

-fcoroutines-ts, -fno-coroutines-ts¶

Enable support for the C++ Coroutines TS

-fcoverage-compilation-dir=<arg>¶

The compilation directory to embed in the coverage mapping.

-fcoverage-mapping, -fno-coverage-mapping¶

Generate coverage mapping to enable code coverage analysis

-fcoverage-prefix-map=<arg>¶

remap file source paths in coverage mapping

-fcreate-profile¶

-fcs-profile-generate¶

Generate instrumented code to collect context sensitive execution counts into default.profraw (overridden by LLVM_PROFILE_FILE env var)

-fcs-profile-generate=<directory>¶

Generate instrumented code to collect context sensitive execution counts into <directory>/default.profraw (overridden by LLVM_PROFILE_FILE env var)

-fcuda-approx-transcendentals, -fno-cuda-approx-transcendentals¶

Use approximate transcendental functions

-fcuda-short-ptr, -fno-cuda-short-ptr¶

Use 32-bit pointers for accessing const/local/shared address spaces

-fcxx-exceptions, -fno-cxx-exceptions¶

Enable C++ exceptions

-fcxx-modules, -fno-cxx-modules¶

-fdata-sections, -fno-data-sections¶

Place each data in its own section

-fdebug-compilation-dir=<arg>, -fdebug-compilation-dir <arg>¶

The compilation directory to embed in the debug info

-fdebug-default-version=<arg>¶

Default DWARF version to use, if a -g option caused DWARF debug info to be produced

-fdebug-info-for-profiling, -fno-debug-info-for-profiling¶

Emit extra debug info to make sample profile more accurate

-fdebug-macro, -fno-debug-macro¶

Emit macro debug information

-fdebug-pass-arguments¶

-fdebug-pass-structure¶

-fdebug-prefix-map=<arg>¶

remap file source paths in debug info

-fdebug-ranges-base-address, -fno-debug-ranges-base-address¶

Use DWARF base address selection entries in .debug_ranges

-fdebug-types-section, -fno-debug-types-section¶

Place debug types in their own section (ELF Only)

-fdelayed-template-parsing, -fno-delayed-template-parsing¶

Parse templated function definitions at the end of the translation unit

-fdelete-null-pointer-checks, -fno-delete-null-pointer-checks¶

Treat usage of null pointers as undefined behavior (default)

-fdenormal-fp-math=<arg>¶

-fdiagnostics-absolute-paths¶

Print absolute paths in diagnostics

-fdiagnostics-color, -fno-diagnostics-color¶

-fdiagnostics-color=<arg>¶

-fdiagnostics-hotness-threshold=<value>¶

Prevent optimization remarks from being output if they do not have at least this profile count. Use ‘auto’ to apply the threshold from profile summary

-fdiagnostics-show-hotness, -fno-diagnostics-show-hotness¶

Enable profile hotness information in diagnostic line

-fdiagnostics-show-note-include-stack, -fno-diagnostics-show-note-include-stack¶

Display include stacks for diagnostic notes

-fdiagnostics-show-option, -fno-diagnostics-show-option¶

Print option name with mappable diagnostics

-fdiagnostics-show-template-tree¶

Print a template comparison tree for differing templates

-fdigraphs, -fno-digraphs¶

Enable alternative token representations ‘<:’, ‘:>’, ‘<%’, ‘%>’, ‘%:’, ‘%:%:’ (default)

-fdirect-access-external-data, -fno-direct-access-external-data¶

Don’t use GOT indirection to reference external data symbols

-fdollars-in-identifiers, -fno-dollars-in-identifiers¶

Allow ‘$’ in identifiers

-fdouble-square-bracket-attributes, -fno-double-square-bracket-attributes¶

Enable ‘[[]]’ attributes in all C and C++ language modes

-fdwarf-directory-asm, -fno-dwarf-directory-asm¶

-fdwarf-exceptions¶

Use DWARF style exceptions

-felide-constructors, -fno-elide-constructors¶

-feliminate-unused-debug-symbols, -fno-eliminate-unused-debug-symbols¶

-feliminate-unused-debug-types, -fno-eliminate-unused-debug-types¶

Do not emit debug info for defined but unused types

-fembed-bitcode=<option>, -fembed-bitcode (equivalent to -fembed-bitcode=all), -fembed-bitcode-marker (equivalent to -fembed-bitcode=marker)¶

Embed LLVM bitcode (option: off, all, bitcode, marker)

-femit-all-decls¶

Emit all declarations, even if unused

-femulated-tls, -fno-emulated-tls¶

Use emutls functions to access thread_local variables

-fenable-matrix¶

Enable matrix data type and related builtin functions

-fencoding=<arg>, --encoding <arg>, --encoding=<arg>¶

-ferror-limit=<arg>¶

-fescaping-block-tail-calls, -fno-escaping-block-tail-calls¶

-fexceptions, -fno-exceptions¶

Enable support for exception handling

-fexec-charset=<arg>¶

-fexperimental-new-constant-interpreter¶

Enable the experimental new constant interpreter

-fextdirs=<arg>, --extdirs <arg>, --extdirs=<arg>¶

-fextend-arguments=<arg>¶

Controls how scalar integer arguments are extended in calls to unprototyped and varargs functions

-ffast-math, -fno-fast-math¶

Allow aggressive, lossy floating-point optimizations

-ffile-compilation-dir=<arg>¶

The compilation directory to embed in the debug info and coverage mapping.

-ffile-prefix-map=<arg>¶

remap file source paths in debug info, predefined preprocessor macros and __builtin_FILE()

-ffinite-loops, -fno-finite-loops¶

Assume all loops are finite.

-ffinite-math-only, -fno-finite-math-only¶

-ffixed-point, -fno-fixed-point¶

Enable fixed point types

-ffor-scope, -fno-for-scope¶

-fforce-dwarf-frame, -fno-force-dwarf-frame¶

Always emit a debug frame section

-fforce-emit-vtables, -fno-force-emit-vtables¶

Emits more virtual tables to improve devirtualization

-fforce-enable-int128, -fno-force-enable-int128¶

Enable support for int128_t type

-ffp-contract=<arg>¶

Form fused FP ops (e.g. FMAs): fast (fuses across statements disregarding pragmas) | on (only fuses in the same statement unless dictated by pragmas) | off (never fuses) | fast-honor-pragmas (fuses across statements unless diectated by pragmas). Default is ‘fast’ for CUDA, ‘fast-honor-pragmas’ for HIP, and ‘on’ otherwise.

-ffp-exception-behavior=<arg>¶

Specifies the exception behavior of floating-point operations.

-ffp-model=<arg>¶

Controls the semantics of floating-point calculations.

-ffreestanding¶

Assert that the compilation takes place in a freestanding environment

-ffunction-sections, -fno-function-sections¶

Place each function in its own section

-fgnu-inline-asm, -fno-gnu-inline-asm¶

-fgnu-keywords, -fno-gnu-keywords¶

Allow GNU-extension keywords regardless of language standard

-fgnu-runtime¶

Generate output compatible with the standard GNU Objective-C runtime

-fgnu89-inline, -fno-gnu89-inline¶

Use the gnu89 inline semantics

-fgnuc-version=<arg>¶

Sets various macros to claim compatibility with the given GCC version (default is 4.2.1)

-fgpu-allow-device-init, -fno-gpu-allow-device-init¶

Allow device side init function in HIP (experimental)

-fgpu-defer-diag, -fno-gpu-defer-diag¶

Defer host/device related diagnostic messages for CUDA/HIP

-fgpu-rdc, -fcuda-rdc, -fno-gpu-rdc¶

Generate relocatable device code, also known as separate compilation mode

-fgpu-sanitize, -fno-gpu-sanitize¶

Enable sanitizer for AMDGPU target

-fhip-fp32-correctly-rounded-divide-sqrt, -fno-hip-fp32-correctly-rounded-divide-sqrt¶

Specify that single precision floating-point divide and sqrt used in the program source are correctly rounded (HIP device compilation only)

-fhip-new-launch-api, -fno-hip-new-launch-api¶

Use new kernel launching API for HIP

-fhonor-infinities, -fhonor-infinites, -fno-honor-infinities¶

-fhonor-nans, -fno-honor-nans¶

-fhosted¶

-fignore-exceptions¶

Enable support for ignoring exception handling constructs

-fimplicit-module-maps, -fmodule-maps, -fno-implicit-module-maps¶

Implicitly search the file system for module map files.

-fimplicit-modules, -fno-implicit-modules¶

-finput-charset=<arg>¶

Specify the default character set for source files

-finstrument-function-entry-bare¶

Instrument function entry only, after inlining, without arguments to the instrumentation call

-finstrument-functions¶

Generate calls to instrument function entry and exit

-finstrument-functions-after-inlining¶

Like -finstrument-functions, but insert the calls after inlining

-fintegrated-as, -fno-integrated-as, -integrated-as¶

Enable the integrated assembler

-fintegrated-cc1, -fno-integrated-cc1¶

Run cc1 in-process

-fjump-tables, -fno-jump-tables¶

Use jump tables for lowering switches

-fkeep-static-consts, -fno-keep-static-consts¶

Keep static const variables if unused

-flax-vector-conversions=<arg>, -flax-vector-conversions (equivalent to -flax-vector-conversions=integer), -fno-lax-vector-conversions (equivalent to -flax-vector-conversions=none)¶

Enable implicit vector bit-casts

-flimited-precision=<arg>¶

-flto, -fno-lto¶

Enable LTO in ‘full’ mode

-flto-jobs=<arg>¶

Controls the backend parallelism of -flto=thin (default of 0 means the number of threads will be derived from the number of CPUs detected)

-flto=<arg>¶

Set LTO mode to either ‘full’ or ‘thin’

-flto=auto¶

-flto=jobserver¶

-fmacro-backtrace-limit=<arg>¶

-fmacro-prefix-map=<arg>¶

remap file source paths in predefined preprocessor macros and __builtin_FILE()

-fmath-errno, -fno-math-errno¶

Require math functions to indicate errors by setting errno

-fmax-tokens=<arg>¶

Max total number of preprocessed tokens for -Wmax-tokens.

-fmax-type-align=<arg>¶

Specify the maximum alignment to enforce on pointers lacking an explicit alignment

-fmemory-profile, -fno-memory-profile¶

Enable heap memory profiling

-fmemory-profile=<directory>¶

Enable heap memory profiling and dump results into <directory>

-fmerge-all-constants, -fno-merge-all-constants¶

Allow merging of constants

-fmessage-length=<arg>¶

Format message diagnostics so that they fit within N columns

-fmodule-file-deps, -fno-module-file-deps¶

-fmodule-map-file=<file>¶

Load this module map file

-fmodule-name=<name>, -fmodule-implementation-of <arg>¶

Specify the name of the module to build

-fmodules, -fno-modules¶

Enable the ‘modules’ language feature

-fmodules-decluse, -fno-modules-decluse¶

Require declaration of modules used within a module

-fmodules-ignore-macro=<arg>¶

Ignore the definition of the given macro when building and loading modules

-fmodules-search-all, -fno-modules-search-all¶

Search even non-imported modules to resolve references

-fmodules-strict-decluse¶

Like -fmodules-decluse but requires all headers to be in modules

-fmodules-ts¶

Enable support for the C++ Modules TS

-fmodules-validate-input-files-content¶

Validate PCM input files based on content if mtime differs

-fms-compatibility, -fno-ms-compatibility¶

Enable full Microsoft Visual C++ compatibility

-fms-compatibility-version=<arg>¶

Dot-separated value representing the Microsoft compiler version number to report in _MSC_VER (0 = don’t define it (default))

-fms-extensions, -fno-ms-extensions¶

Accept some non-standard constructs supported by the Microsoft compiler

-fms-memptr-rep=<arg>¶

-fms-volatile¶

-fmsc-version=<arg>¶

Microsoft compiler version number to report in _MSC_VER (0 = don’t define it (default))

-fmudflap¶

-fmudflapth¶

-fnested-functions¶

-fnew-alignment=<align>, -fnew-alignment <arg>¶

Specifies the largest alignment guaranteed by ‘::operator new(size_t)’

-fnext-runtime¶

-fno-builtin-<arg>¶

Disable implicit builtin knowledge of a specific function

-fno-elide-type¶

Do not elide types when printing diagnostics

-fno-max-type-align¶

-fno-strict-modules-decluse¶

-fno-temp-file¶

Directly create compilation output files. This may lead to incorrect incremental builds if the compiler crashes

-fno-working-directory¶

-fno_modules-validate-input-files-content¶

-fno_pch-validate-input-files-content¶

-fnoxray-link-deps¶

-fobjc-abi-version=<arg>¶

-fobjc-arc, -fno-objc-arc¶

Synthesize retain and release calls for Objective-C pointers

-fobjc-arc-exceptions, -fno-objc-arc-exceptions¶

Use EH-safe code when synthesizing retains and releases in -fobjc-arc

-fobjc-convert-messages-to-runtime-calls, -fno-objc-convert-messages-to-runtime-calls¶

-fobjc-disable-direct-methods-for-testing¶

Ignore attribute objc_direct so that direct methods can be tested

-fobjc-encode-cxx-class-template-spec, -fno-objc-encode-cxx-class-template-spec¶

Fully encode c++ class template specialization

-fobjc-exceptions, -fno-objc-exceptions¶

Enable Objective-C exceptions

-fobjc-infer-related-result-type, -fno-objc-infer-related-result-type¶

-fobjc-legacy-dispatch, -fno-objc-legacy-dispatch¶

-fobjc-link-runtime¶

-fobjc-nonfragile-abi, -fno-objc-nonfragile-abi¶

-fobjc-nonfragile-abi-version=<arg>¶

-fobjc-runtime=<arg>¶

Specify the target Objective-C runtime kind and version

-fobjc-sender-dependent-dispatch¶

-fobjc-weak, -fno-objc-weak¶

Enable ARC-style weak references in Objective-C

-foffload-lto, -fno-offload-lto¶

Enable LTO in ‘full’ mode for offload compilation

-foffload-lto=<arg>¶

Set LTO mode to either ‘full’ or ‘thin’ for offload compilation

-fomit-frame-pointer, -fno-omit-frame-pointer¶

-fopenmp, -fno-openmp¶

Parse OpenMP pragmas and generate parallel code.

-fopenmp-simd, -fno-openmp-simd¶

Emit OpenMP code only for SIMD-based constructs.

-fopenmp-target-new-runtime, -fno-openmp-target-new-runtime¶

Use the new bitcode library for OpenMP offloading

-fopenmp-version=<arg>¶

-fopenmp=<arg>¶

-foperator-arrow-depth=<arg>¶

-foperator-names, -fno-operator-names¶

-foptimization-record-file=<file>¶

Specify the output name of the file containing the optimization remarks. Implies -fsave-optimization-record. On Darwin platforms, this cannot be used with multiple -arch <arch> options.

-foptimization-record-passes=<regex>¶

Only include passes which match a specified regular expression in the generated optimization record (by default, include all passes)

-foptimize-sibling-calls, -fno-optimize-sibling-calls¶

-forder-file-instrumentation¶

Generate instrumented code to collect order file into default.profraw file (overridden by ‘=’ form of option or LLVM_PROFILE_FILE env var)

-foutput-class-dir=<arg>, --output-class-directory <arg>, --output-class-directory=<arg>¶

-fpack-struct, -fno-pack-struct¶

-fpack-struct=<arg>¶

Specify the default maximum struct packing alignment

-fpascal-strings, -fno-pascal-strings, -mpascal-strings¶

Recognize and construct Pascal-style string literals

-fpass-plugin=<dsopath>¶

Load pass plugin from a dynamic shared object file (only with new pass manager).

-fpatchable-function-entry=<N,M>¶

Generate M NOPs before function entry and N-M NOPs after function entry

-fpcc-struct-return¶

Override the default ABI to return all structs on the stack

-fpch-codegen, -fno-pch-codegen¶

Generate code for uses of this PCH that assumes an explicit object file will be built for the PCH

-fpch-debuginfo, -fno-pch-debuginfo¶

Generate debug info for types in an object file built from this PCH and do not generate them elsewhere

-fpch-instantiate-templates, -fno-pch-instantiate-templates¶

Instantiate templates already while building a PCH

-fpch-preprocess¶

-fpch-validate-input-files-content¶

Validate PCH input files based on content if mtime differs

-fpic, -fno-pic¶

-fpie, -fno-pie¶

-fplt, -fno-plt¶

-fplugin=<dsopath>¶

Load the named plugin (dynamic shared object)

-fprebuilt-implicit-modules, -fno-prebuilt-implicit-modules¶

Look up implicit modules in the prebuilt module path

-fpreserve-as-comments, -fno-preserve-as-comments¶

-fproc-stat-report<arg>¶

Print subprocess statistics

-fproc-stat-report=<arg>¶

Save subprocess statistics to the given file

-fprofile-arcs, -fno-profile-arcs¶

-fprofile-dir=<arg>¶

-fprofile-exclude-files=<arg>¶

Instrument only functions from files where names don’t match all the regexes separated by a semi-colon

-fprofile-filter-files=<arg>¶

Instrument only functions from files where names match any regex separated by a semi-colon

-fprofile-generate, -fno-profile-generate¶

Generate instrumented code to collect execution counts into default.profraw (overridden by LLVM_PROFILE_FILE env var)

-fprofile-generate=<directory>¶

Generate instrumented code to collect execution counts into <directory>/default.profraw (overridden by LLVM_PROFILE_FILE env var)

-fprofile-instr-generate, -fno-profile-instr-generate¶

Generate instrumented code to collect execution counts into default.profraw file (overridden by ‘=’ form of option or LLVM_PROFILE_FILE env var)

-fprofile-instr-generate=<file>¶

Generate instrumented code to collect execution counts into <file> (overridden by LLVM_PROFILE_FILE env var)

-fprofile-instr-use, -fno-profile-instr-use, -fprofile-use¶

-fprofile-instr-use=<arg>¶

Use instrumentation data for profile-guided optimization

-fprofile-list=<arg>¶

Filename defining the list of functions/files to instrument

-fprofile-remapping-file=<file>¶

Use the remappings described in <file> to match the profile data against names in the program

-fprofile-sample-accurate, -fauto-profile-accurate, -fno-profile-sample-accurate¶

Specifies that the sample profile is accurate. If the sample: profile is accurate, callsites without profile samples are marked as cold. Otherwise, treat callsites without profile samples as if we have no profile

-fprofile-sample-use, -fauto-profile, -fno-profile-sample-use¶

-fprofile-sample-use=<arg>, -fauto-profile=<arg>¶

Enable sample-based profile guided optimizations

-fprofile-update=<method>¶

Set update method of profile counters (atomic,prefer-atomic,single)

-fprofile-use=<pathname>¶

Use instrumentation data for profile-guided optimization. If pathname is a directory, it reads from <pathname>/default.profdata. Otherwise, it reads from file <pathname>.

-fprotect-parens, -fno-protect-parens¶

Determines whether the optimizer honors parentheses when floating-point expressions are evaluated

-fpseudo-probe-for-profiling, -fno-pseudo-probe-for-profiling¶

Emit pseudo probes for sample profiling

-freciprocal-math, -fno-reciprocal-math¶

Allow division operations to be reassociated

-freg-struct-return¶

Override the default ABI to return small structs in registers

-fregister-global-dtors-with-atexit, -fno-register-global-dtors-with-atexit¶

Use atexit or __cxa_atexit to register global destructors

-frelaxed-template-template-args, -fno-relaxed-template-template-args¶

Enable C++17 relaxed template template argument matching

-freroll-loops, -fno-reroll-loops¶

Turn on loop reroller

-fretain-comments-from-system-headers¶

-frewrite-imports, -fno-rewrite-imports¶

-frewrite-includes, -fno-rewrite-includes¶

-frewrite-map-file=<arg>¶

-fropi, -fno-ropi¶

Generate read-only position independent code (ARM only)

-frounding-math, -fno-rounding-math¶

-frtti, -fno-rtti¶

-frtti-data, -fno-rtti-data¶

-frwpi, -fno-rwpi¶

Generate read-write position independent code (ARM only)

-fsave-optimization-record, -fno-save-optimization-record¶

Generate a YAML optimization record file

-fsave-optimization-record=<format>¶

Generate an optimization record file in a specific format

-fseh-exceptions¶

Use SEH style exceptions

-fsemantic-interposition, -fno-semantic-interposition¶

-fshort-enums, -fno-short-enums¶

Allocate to an enum type only as many bytes as it needs for the declared range of possible values

-fshort-wchar, -fno-short-wchar¶

Force wchar_t to be a short unsigned int

-fshow-column, -fno-show-column¶

-fshow-overloads=<arg>¶

Which overload candidates to show when overload resolution fails: best|all; defaults to all

-fshow-source-location, -fno-show-source-location¶

-fsignaling-math, -fno-signaling-math¶

-fsigned-bitfields¶

-fsigned-char, -fno-signed-char, --signed-char¶

char is signed

-fsigned-zeros, -fno-signed-zeros¶

-fsized-deallocation, -fno-sized-deallocation¶

Enable C++14 sized global deallocation functions

-fsjlj-exceptions¶

Use SjLj style exceptions

-fslp-vectorize, -fno-slp-vectorize, -ftree-slp-vectorize¶

Enable the superword-level parallelism vectorization passes

-fspell-checking, -fno-spell-checking¶

-fspell-checking-limit=<arg>¶

-fsplit-dwarf-inlining, -fno-split-dwarf-inlining¶

Provide minimal debug info in the object/executable to facilitate online symbolication/stack traces in the absence of .dwo/.dwp files when using Split DWARF

-fsplit-lto-unit, -fno-split-lto-unit¶

Enables splitting of the LTO unit

-fsplit-machine-functions, -fno-split-machine-functions¶

Enable late function splitting using profile information (x86 ELF)

-fsplit-stack, -fno-split-stack¶

Use segmented stack

-fstack-clash-protection, -fno-stack-clash-protection¶

Enable stack clash protection

-fstack-protector, -fno-stack-protector¶

Enable stack protectors for some functions vulnerable to stack smashing. This uses a loose heuristic which considers functions vulnerable if they contain a char (or 8bit integer) array or constant sized calls to alloca , which are of greater size than ssp-buffer-size (default: 8 bytes). All variable sized calls to alloca are considered vulnerable. A function with a stack protector has a guard value added to the stack frame that is checked on function exit. The guard value must be positioned in the stack frame such that a buffer overflow from a vulnerable variable will overwrite the guard value before overwriting the function’s return address. The reference stack guard value is stored in a global variable.

-fstack-protector-all¶

Enable stack protectors for all functions

-fstack-protector-strong¶

Enable stack protectors for some functions vulnerable to stack smashing. Compared to -fstack-protector, this uses a stronger heuristic that includes functions containing arrays of any size (and any type), as well as any calls to alloca or the taking of an address from a local variable

-fstack-size-section, -fno-stack-size-section¶

Emit section containing metadata on function stack sizes

-fstack-usage¶

Emit .su file containing information on function stack sizes

-fstandalone-debug, -fno-limit-debug-info, -fno-standalone-debug¶

Emit full debug info for all types used by the program

-fstrict-aliasing, -fno-strict-aliasing¶

-fstrict-enums, -fno-strict-enums¶

Enable optimizations based on the strict definition of an enum’s value range

-fstrict-float-cast-overflow, -fno-strict-float-cast-overflow¶

Assume that overflowing float-to-int casts are undefined (default)

-fstrict-overflow, -fno-strict-overflow¶

-fstrict-return, -fno-strict-return¶

-fstrict-vtable-pointers, -fno-strict-vtable-pointers¶

Enable optimizations based on the strict rules for overwriting polymorphic C++ objects

-fstruct-path-tbaa, -fno-struct-path-tbaa¶

-fsymbol-partition=<arg>¶

-ftabstop=<arg>¶

-ftemplate-backtrace-limit=<arg>¶

-ftemplate-depth-<arg>¶

-ftemplate-depth=<arg>¶

-ftest-coverage, -fno-test-coverage¶

-fthin-link-bitcode=<arg>¶

Write minimized bitcode to <file> for the ThinLTO thin link only

-fthinlto-index=<arg>¶

Perform ThinLTO importing using provided function summary index

-fthreadsafe-statics, -fno-threadsafe-statics¶

-ftime-report¶

-ftime-report=<arg>¶

(For new pass manager) “per-pass”: one report for each pass; “per-pass-run”: one report for each pass invocation

-ftime-trace¶

Turn on time profiler. Generates JSON file based on output filename. Results can be analyzed with chrome://tracing or Speedscope App for flamegraph visualization.

-ftime-trace-granularity=<arg>¶

Minimum time granularity (in microseconds) traced by time profiler

-ftls-model=<arg>¶

-ftrap-function=<arg>¶

Issue call to specified function rather than a trap instruction

-ftrapping-math, -fno-trapping-math¶

-ftrapv¶

Trap on integer overflow

-ftrapv-handler <arg>¶

-ftrapv-handler=<function name>¶

Specify the function to be called on overflow

-ftrigraphs, -fno-trigraphs, -trigraphs, --trigraphs¶

Process trigraph sequences

-ftrivial-auto-var-init-stop-after=<arg>¶

Stop initializing trivial automatic stack variables after the specified number of instances

-ftrivial-auto-var-init=<arg>¶

Initialize trivial automatic stack variables: uninitialized (default) | pattern

-funique-basic-block-section-names, -fno-unique-basic-block-section-names¶

Use unique names for basic block sections (ELF Only)

-funique-internal-linkage-names, -fno-unique-internal-linkage-names¶

Uniqueify Internal Linkage Symbol Names by appending the MD5 hash of the module path

-funique-section-names, -fno-unique-section-names¶

-funit-at-a-time, -fno-unit-at-a-time¶

-funroll-loops, -fno-unroll-loops¶

Turn on loop unroller

-funsafe-math-optimizations, -fno-unsafe-math-optimizations¶

-funsigned-bitfields¶

-funsigned-char, -fno-unsigned-char, --unsigned-char¶

-funwind-tables, -fno-unwind-tables¶

-fuse-cxa-atexit, -fno-use-cxa-atexit¶

-fuse-init-array, -fno-use-init-array¶

-fuse-ld=<arg>¶

-fuse-line-directives, -fno-use-line-directives¶

Use #line in preprocessed output

-fvalidate-ast-input-files-content¶

Compute and store the hash of input files used to build an AST. Files with mismatching mtime’s are considered valid if both contents is identical

-fveclib=<arg>¶

Use the given vector functions library

-fvectorize, -fno-vectorize, -ftree-vectorize¶

Enable the loop vectorization passes

-fverbose-asm, -dA, -fno-verbose-asm¶

Generate verbose assembly output

-fvirtual-function-elimination, -fno-virtual-function-elimination¶

Enables dead virtual function elimination optimization. Requires -flto=full

-fvisibility-dllexport=<arg>¶

The visibility for dllexport defintions [-fvisibility-from-dllstorageclass]

-fvisibility-externs-dllimport=<arg>¶

The visibility for dllimport external declarations [-fvisibility-from-dllstorageclass]

-fvisibility-externs-nodllstorageclass=<arg>¶

The visibility for external declarations without an explicit DLL dllstorageclass [-fvisibility-from-dllstorageclass]

-fvisibility-from-dllstorageclass, -fno-visibility-from-dllstorageclass¶

Set the visiblity of symbols in the generated code from their DLL storage class

-fvisibility-global-new-delete-hidden¶

Give global C++ operator new and delete declarations hidden visibility

-fvisibility-inlines-hidden, -fno-visibility-inlines-hidden¶

Give inline C++ member functions hidden visibility by default

-fvisibility-inlines-hidden-static-local-var, -fno-visibility-inlines-hidden-static-local-var¶

When -fvisibility-inlines-hidden is enabled, static variables in inline C++ member functions will also be given hidden visibility by default

-fvisibility-ms-compat¶

Give global types ‘default’ visibility and global functions and variables ‘hidden’ visibility by default

-fvisibility-nodllstorageclass=<arg>¶

The visibility for defintiions without an explicit DLL export class [-fvisibility-from-dllstorageclass]

-fvisibility=<arg>¶

Set the default symbol visibility for all global declarations

-fwasm-exceptions¶

Use WebAssembly style exceptions

-fwhole-program-vtables, -fno-whole-program-vtables¶

Enables whole-program vtable optimization. Requires -flto

-fwrapv, -fno-wrapv¶

Treat signed integer overflow as two’s complement

-fwritable-strings¶

Store string literals as writable data

-fxl-pragma-pack, -fno-xl-pragma-pack¶

Enable IBM XL #pragma pack handling

-fxray-always-emit-customevents, -fno-xray-always-emit-customevents¶

Always emit __xray_customevent(…) calls even if the containing function is not always instrumented

-fxray-always-emit-typedevents, -fno-xray-always-emit-typedevents¶

Always emit __xray_typedevent(…) calls even if the containing function is not always instrumented

-fxray-always-instrument=<arg>¶

DEPRECATED: Filename defining the whitelist for imbuing the ‘always instrument’ XRay attribute.

-fxray-attr-list=<arg>¶

Filename defining the list of functions/types for imbuing XRay attributes.

-fxray-function-groups=<arg>¶

Only instrument 1 of N groups

-fxray-function-index, -fno-xray-function-index¶

-fxray-ignore-loops, -fno-xray-ignore-loops¶

Don’t instrument functions with loops unless they also meet the minimum function size

-fxray-instruction-threshold<arg>¶

-fxray-instruction-threshold=<arg>¶

Sets the minimum function size to instrument with XRay

-fxray-instrument, -fno-xray-instrument¶

Generate XRay instrumentation sleds on function entry and exit

-fxray-instrumentation-bundle=<arg>¶

Select which XRay instrumentation points to emit. Options: all, none, function-entry, function-exit, function, custom. Default is ‘all’. ‘function’ includes both ‘function-entry’ and ‘function-exit’.

-fxray-link-deps¶

Tells clang to add the link dependencies for XRay.

-fxray-modes=<arg>¶

List of modes to link in by default into XRay instrumented binaries.

-fxray-never-instrument=<arg>¶

DEPRECATED: Filename defining the whitelist for imbuing the ‘never instrument’ XRay attribute.

-fxray-selected-function-group=<arg>¶

When using -fxray-function-groups, select which group of functions to instrument. Valid range is 0 to fxray-function-groups - 1

-fzero-initialized-in-bss, -fno-zero-initialized-in-bss¶

-fzvector, -fno-zvector, -mzvector¶

Enable System z vector language extension

--gpu-bundle-output, --no-gpu-bundle-output¶

Bundle output files of HIP device compilation

-pedantic, --pedantic, -no-pedantic, --no-pedantic¶

Warn on language extensions

-pedantic-errors, --pedantic-errors¶

OpenCL flags ¶

-cl-denorms-are-zero¶

OpenCL only. Allow denormals to be flushed to zero.

-cl-fast-relaxed-math¶

OpenCL only. Sets -cl-finite-math-only and -cl-unsafe-math-optimizations, and defines __FAST_RELAXED_MATH__.

-cl-finite-math-only¶

OpenCL only. Allow floating-point optimizations that assume arguments and results are not NaNs or +-Inf.

-cl-fp32-correctly-rounded-divide-sqrt¶

OpenCL only. Specify that single precision floating-point divide and sqrt used in the program source are correctly rounded.

-cl-kernel-arg-info¶

OpenCL only. Generate kernel argument metadata.

-cl-mad-enable¶

OpenCL only. Allow use of less precise MAD computations in the generated binary.

-cl-no-signed-zeros¶

OpenCL only. Allow use of less precise no signed zeros computations in the generated binary.

-cl-no-stdinc¶

OpenCL only. Disables all standard includes containing non-native compiler types and functions.

-cl-opt-disable¶

OpenCL only. This option disables all optimizations. By default optimizations are enabled.

-cl-single-precision-constant¶

OpenCL only. Treat double precision floating-point constant as single precision constant.

-cl-std=<arg>¶

OpenCL language standard to compile for.

-cl-strict-aliasing¶

OpenCL only. This option is added for compatibility with OpenCL 1.0.

-cl-uniform-work-group-size¶

OpenCL only. Defines that the global work-size be a multiple of the work-group size specified to clEnqueueNDRangeKernel

-cl-unsafe-math-optimizations¶

OpenCL only. Allow unsafe floating-point optimizations. Also implies -cl-no-signed-zeros and -cl-mad-enable.

SYCL flags ¶

-fsycl, -fno-sycl¶

Enables SYCL kernels compilation for device

-sycl-std=<arg>¶

SYCL language standard to compile for.

Target-dependent compilation options ¶

-G<size>, -G=<arg>, -msmall-data-limit=<arg>, -msmall-data-threshold=<arg>¶

Put objects of at most <size> bytes into small data section (MIPS / Hexagon)

-ffixed-x1¶

Reserve the x1 register (AArch64/RISC-V only)

-ffixed-x10¶

Reserve the x10 register (AArch64/RISC-V only)

-ffixed-x11¶

Reserve the x11 register (AArch64/RISC-V only)

-ffixed-x12¶

Reserve the x12 register (AArch64/RISC-V only)

-ffixed-x13¶

Reserve the x13 register (AArch64/RISC-V only)

-ffixed-x14¶

Reserve the x14 register (AArch64/RISC-V only)

-ffixed-x15¶

Reserve the x15 register (AArch64/RISC-V only)

-ffixed-x16¶

Reserve the x16 register (AArch64/RISC-V only)

-ffixed-x17¶

Reserve the x17 register (AArch64/RISC-V only)

-ffixed-x18¶

Reserve the x18 register (AArch64/RISC-V only)

-ffixed-x19¶

Reserve the x19 register (AArch64/RISC-V only)

-ffixed-x2¶

Reserve the x2 register (AArch64/RISC-V only)

-ffixed-x20¶

Reserve the x20 register (AArch64/RISC-V only)

-ffixed-x21¶

Reserve the x21 register (AArch64/RISC-V only)

-ffixed-x22¶

Reserve the x22 register (AArch64/RISC-V only)

-ffixed-x23¶

Reserve the x23 register (AArch64/RISC-V only)

-ffixed-x24¶

Reserve the x24 register (AArch64/RISC-V only)

-ffixed-x25¶

Reserve the x25 register (AArch64/RISC-V only)

-ffixed-x26¶

Reserve the x26 register (AArch64/RISC-V only)

-ffixed-x27¶

Reserve the x27 register (AArch64/RISC-V only)

-ffixed-x28¶

Reserve the x28 register (AArch64/RISC-V only)

-ffixed-x29¶

Reserve the x29 register (AArch64/RISC-V only)

-ffixed-x3¶

Reserve the x3 register (AArch64/RISC-V only)

-ffixed-x30¶

Reserve the x30 register (AArch64/RISC-V only)

-ffixed-x31¶

Reserve the x31 register (AArch64/RISC-V only)

-ffixed-x4¶

Reserve the x4 register (AArch64/RISC-V only)

-ffixed-x5¶

Reserve the x5 register (AArch64/RISC-V only)

-ffixed-x6¶

Reserve the x6 register (AArch64/RISC-V only)

-ffixed-x7¶

Reserve the x7 register (AArch64/RISC-V only)

-ffixed-x8¶

Reserve the x8 register (AArch64/RISC-V only)

-ffixed-x9¶

Reserve the x9 register (AArch64/RISC-V only)

-m16¶

-m32¶

-m64¶

-mabi=<arg>¶

-mabi=vec-default¶

Enable the default Altivec ABI on AIX (AIX only). Uses only volatile vector registers.

-mabi=vec-extabi¶

Enable the extended Altivec ABI on AIX (AIX only). Uses volatile and nonvolatile vector registers

-maix-struct-return¶

Return all structs in memory (PPC32 only)

-malign-branch-boundary=<arg>¶

Specify the boundary’s size to align branches

-malign-branch=<arg1>,<arg2>...¶

Specify types of branches to align

-malign-double¶

Align doubles to two words in structs (x86 only)

-mamdgpu-ieee, -mno-amdgpu-ieee¶

Sets the IEEE bit in the expected default floating point mode register. Floating point opcodes that support exception flag gathering quiet and propagate signaling NaN inputs per IEEE 754-2008. This option changes the ABI. (AMDGPU only)

-march=<arg>¶

-masm=<arg>¶

-mbackchain, -mno-backchain¶

Link stack frames through backchain on System Z

-mbranches-within-32B-boundaries¶

Align selected branches (fused, jcc, jmp) within 32-byte boundary

-mcmodel=<arg>, -mcmodel=medany (equivalent to -mcmodel=medium), -mcmodel=medlow (equivalent to -mcmodel=small)¶

-mcode-object-v3, -mno-code-object-v3¶

Legacy option to specify code object ABI V3 (AMDGPU only)

-mcode-object-version=<version>¶

Specify code object ABI version. Defaults to 3. (AMDGPU only)

-mconsole<arg>¶

-mcpu=<arg>, -mv5 (equivalent to -mcpu=hexagonv5), -mv55 (equivalent to -mcpu=hexagonv55), -mv60 (equivalent to -mcpu=hexagonv60), -mv62 (equivalent to -mcpu=hexagonv62), -mv65 (equivalent to -mcpu=hexagonv65), -mv66 (equivalent to -mcpu=hexagonv66), -mv67 (equivalent to -mcpu=hexagonv67), -mv67t (equivalent to -mcpu=hexagonv67t), -mv68 (equivalent to -mcpu=hexagonv68)¶

-mcrc, -mno-crc¶

Allow use of CRC instructions (ARM/Mips only)

-mdefault-build-attributes<arg>, -mno-default-build-attributes<arg>¶

-mdll<arg>¶

-mdouble=<arg>¶

Force double to be 32 bits or 64 bits

-mdynamic-no-pic<arg>¶

-meabi <arg>¶

Set EABI type, e.g. 4, 5 or gnu (default depends on triple)

-menable-experimental-extensions¶

Enable use of experimental RISC-V extensions.

-mfentry¶

Insert calls to fentry at function entry (x86/SystemZ only)

-mfloat-abi=<arg>¶

-mfpmath=<arg>¶

-mfpu=<arg>¶

-mgeneral-regs-only¶

Generate code which only uses the general purpose registers (AArch64/x86 only)

-mglobal-merge, -mno-global-merge¶

Enable merging of globals

-mhard-float¶

-mhwdiv=<arg>, --mhwdiv <arg>, --mhwdiv=<arg>¶

-mhwmult=<arg>¶

-miamcu, -mno-iamcu¶

Use Intel MCU ABI

-mignore-xcoff-visibility¶

Not emit the visibility attribute for asm in AIX OS or give all symbols ‘unspecified’ visibility in XCOFF object file

-mimplicit-float, -mno-implicit-float¶

-mimplicit-it=<arg>¶

-mincremental-linker-compatible, -mno-incremental-linker-compatible¶

(integrated-as) Emit an object file which can be used with an incremental linker

-miphoneos-version-min=<arg>, -mios-version-min=<arg>¶

-mkernel¶

-mlong-calls, -mno-long-calls¶

Generate branches with extended addressability, usually via indirect jumps.

-mlvi-cfi, -mno-lvi-cfi¶

Enable only control-flow mitigations for Load Value Injection (LVI)

-mlvi-hardening, -mno-lvi-hardening¶

Enable all mitigations for Load Value Injection (LVI)

-mmacosx-version-min=<arg>, -mmacos-version-min=<arg>¶

Set Mac OS X deployment target

-mmcu=<arg>¶

-mms-bitfields, -mno-ms-bitfields¶

Set the default structure layout to be compatible with the Microsoft compiler standard

-mnop-mcount¶

Generate mcount/__fentry__ calls as nops. To activate they need to be patched in.

-momit-leaf-frame-pointer, -mno-omit-leaf-frame-pointer¶

Omit frame pointer setup for leaf functions

-moslib=<arg>¶

-mpacked-stack, -mno-packed-stack¶

Use packed stack layout (SystemZ only).

-mpad-max-prefix-size=<arg>¶

Specify maximum number of prefixes to use for padding

-mprefer-vector-width=<arg>¶

Specifies preferred vector width for auto-vectorization. Defaults to ‘none’ which allows target specific decisions.

-mqdsp6-compat¶

Enable hexagon-qdsp6 backward compatibility

-mrecip¶

-mrecip=<arg1>,<arg2>...¶

-mrecord-mcount¶

Generate a __mcount_loc section entry for each __fentry__ call.

-mred-zone, -mno-red-zone¶

-mregparm=<arg>¶

-mrelax, -mno-relax¶

Enable linker relaxation

-mrelax-all, -mno-relax-all¶

(integrated-as) Relax all machine instructions

-mretpoline, -mno-retpoline¶

-mrtd, -mno-rtd¶

Make StdCall calling convention the default

-mseses, -mno-seses¶

Enable speculative execution side effect suppression (SESES). Includes LVI control flow integrity mitigations

-msign-return-address=<arg>¶

Select return address signing scope

-msim¶

-msoft-float, -mno-soft-float¶

Use software floating point

-mspeculative-load-hardening, -mno-speculative-load-hardening¶

-mstack-alignment=<arg>¶

Set the stack alignment

-mstack-arg-probe, -mno-stack-arg-probe¶

Enable stack probes

-mstack-probe-size=<arg>¶

Set the stack probe size

-mstack-protector-guard-offset=<arg>¶

Use the given offset for addressing the stack-protector guard

-mstack-protector-guard-reg=<arg>¶

Use the given reg for addressing the stack-protector guard

-mstack-protector-guard=<arg>¶

Use the given guard (global, tls) for addressing the stack-protector guard

-mstackrealign, -mno-stackrealign¶

Force realign the stack at entry to every function

-msvr4-struct-return¶

Return small structs in registers (PPC32 only)

-mthread-model <arg>¶

The thread model to use, e.g. posix, single (posix by default)

-mthreads<arg>¶

-mthumb, -mno-thumb¶

-mtls-direct-seg-refs, -mno-tls-direct-seg-refs¶

Enable direct TLS access through segment registers (default)

-mtls-size=<arg>¶

Specify bit size of immediate TLS offsets (AArch64 ELF only): 12 (for 4KB) | 24 (for 16MB, default) | 32 (for 4GB) | 48 (for 256TB, needs -mcmodel=large)

-mtune=<arg>¶

Only supported on X86 and RISC-V. Otherwise accepted for compatibility with GCC.

-mtvos-version-min=<arg>, -mappletvos-version-min=<arg>¶

-municode<arg>¶

-munsafe-fp-atomics, -mno-unsafe-fp-atomics¶

Enable unsafe floating point atomic instructions (AMDGPU only)

-mvx, -mno-vx¶

-mwarn-nonportable-cfstrings, -mno-warn-nonportable-cfstrings¶

-mwatchos-version-min=<arg>¶

-mwavefrontsize64, -mno-wavefrontsize64¶

Specify wavefront size 64 mode (AMDGPU only)

-mwindows<arg>¶

-mx32¶

AARCH64 ¶

-fcall-saved-x10¶

Make the x10 register call-saved (AArch64 only)

-fcall-saved-x11¶

Make the x11 register call-saved (AArch64 only)

-fcall-saved-x12¶

Make the x12 register call-saved (AArch64 only)

-fcall-saved-x13¶

Make the x13 register call-saved (AArch64 only)

-fcall-saved-x14¶

Make the x14 register call-saved (AArch64 only)

-fcall-saved-x15¶

Make the x15 register call-saved (AArch64 only)

-fcall-saved-x18¶

Make the x18 register call-saved (AArch64 only)

-fcall-saved-x8¶

Make the x8 register call-saved (AArch64 only)

-fcall-saved-x9¶

Make the x9 register call-saved (AArch64 only)

-mfix-cortex-a53-835769, -mno-fix-cortex-a53-835769¶

Workaround Cortex-A53 erratum 835769 (AArch64 only)

-mmark-bti-property¶

Add .note.gnu.property with BTI to assembly files (AArch64 only)

-msve-vector-bits=<arg>¶

Specify the size in bits of an SVE vector register. Defaults to the vector length agnostic value of “scalable”. (AArch64 only)

AMDGPU ¶

-mcumode, -mno-cumode¶

Specify CU wavefront execution mode (AMDGPU only)

-mtgsplit, -mno-tgsplit¶

Enable threadgroup split execution mode (AMDGPU only)

ARM ¶

-faapcs-bitfield-load¶

Follows the AAPCS standard that all volatile bit-field write generates at least one load. (ARM only).

-faapcs-bitfield-width, -fno-aapcs-bitfield-width¶

Follow the AAPCS standard requirement stating that volatile bit-field width is dictated by the field container type. (ARM only).

-ffixed-r9¶

Reserve the r9 register (ARM only)

-mcmse¶

Allow use of CMSE (Armv8-M Security Extensions)

-mexecute-only, -mno-execute-only, -mpure-code¶

Disallow generation of data access to code sections (ARM only)

-mno-movt¶

Disallow use of movt/movw pairs (ARM only)

-mno-neg-immediates¶

Disallow converting instructions with negative immediates to their negation or inversion.

-mnocrc¶

Disallow use of CRC instructions (ARM only)

-mrestrict-it, -mno-restrict-it¶

Disallow generation of deprecated IT blocks for ARMv8. It is on by default for ARMv8 Thumb mode.

-mtp=<arg>¶

Thread pointer access method (AArch32/AArch64 only)

-munaligned-access, -mno-unaligned-access¶

Allow memory accesses to be unaligned (AArch32/AArch64 only)

Hexagon ¶

-mieee-rnd-near¶

-mmemops, -mno-memops¶

Enable generation of memop instructions

-mnvj, -mno-nvj¶

Enable generation of new-value jumps

-mnvs, -mno-nvs¶

Enable generation of new-value stores

-mpackets, -mno-packets¶

Enable generation of instruction packets

Hexagon ¶

-mhvx, -mno-hvx¶

Enable Hexagon Vector eXtensions

-mhvx-length=<arg>¶

Set Hexagon Vector Length

-mhvx=<arg>¶

Enable Hexagon Vector eXtensions

M68k ¶

-ffixed-a0¶

Reserve the a0 register (M68k only)

-ffixed-a1¶

Reserve the a1 register (M68k only)

-ffixed-a2¶

Reserve the a2 register (M68k only)

-ffixed-a3¶

Reserve the a3 register (M68k only)

-ffixed-a4¶

Reserve the a4 register (M68k only)

-ffixed-a5¶

Reserve the a5 register (M68k only)

-ffixed-a6¶

Reserve the a6 register (M68k only)

-ffixed-d0¶

Reserve the d0 register (M68k only)

-ffixed-d1¶

Reserve the d1 register (M68k only)

-ffixed-d2¶

Reserve the d2 register (M68k only)

-ffixed-d3¶

Reserve the d3 register (M68k only)

-ffixed-d4¶

Reserve the d4 register (M68k only)

-ffixed-d5¶

Reserve the d5 register (M68k only)

-ffixed-d6¶

Reserve the d6 register (M68k only)

-ffixed-d7¶

Reserve the d7 register (M68k only)

-m68000¶

-m68010¶

-m68020¶

-m68030¶

-m68040¶

-m68060¶

MIPS ¶

-mabicalls, -mno-abicalls¶

Enable SVR4-style position-independent code (Mips only)

-mabs=<arg>¶

-mcheck-zero-division, -mno-check-zero-division¶

-mcompact-branches=<arg>¶

-mdouble-float¶

-mdsp, -mno-dsp¶

-mdspr2, -mno-dspr2¶

-membedded-data, -mno-embedded-data¶

Place constants in the .rodata section instead of the .sdata section even if they meet the -G <size> threshold (MIPS)

-mextern-sdata, -mno-extern-sdata¶

Assume that externally defined data is in the small data if it meets the -G <size> threshold (MIPS)

-mfp32¶

Use 32-bit floating point registers (MIPS only)

-mfp64¶

Use 64-bit floating point registers (MIPS only)

-mginv, -mno-ginv¶

-mgpopt, -mno-gpopt¶

Use GP relative accesses for symbols known to be in a small data section (MIPS)

-mindirect-jump=<arg>¶

Change indirect jump instructions to inhibit speculation

-mips16¶

-mldc1-sdc1, -mno-ldc1-sdc1¶

-mlocal-sdata, -mno-local-sdata¶

Extend the -G behaviour to object local data (MIPS)

-mmadd4, -mno-madd4¶

Enable the generation of 4-operand madd.s, madd.d and related instructions.

-mmicromips, -mno-micromips¶

-mmsa, -mno-msa¶

Enable MSA ASE (MIPS only)

-mmt, -mno-mt¶

Enable MT ASE (MIPS only)

-mnan=<arg>¶

-mno-mips16¶

-msingle-float¶

-mvirt, -mno-virt¶

-mxgot, -mno-xgot¶

PowerPC ¶

-maltivec, -mno-altivec¶

-mcmpb, -mno-cmpb¶

-mcrbits, -mno-crbits¶

-mcrypto, -mno-crypto¶

-mdirect-move, -mno-direct-move¶

-mefpu2¶

-mfloat128, -mno-float128¶

-mfprnd, -mno-fprnd¶

-mhtm, -mno-htm¶

-minvariant-function-descriptors, -mno-invariant-function-descriptors¶

-misel, -mno-isel¶

-mlongcall, -mno-longcall¶

-mmfocrf, -mmfcrf, -mno-mfocrf¶

-mmma, -mno-mma¶

-mpaired-vector-memops, -mno-paired-vector-memops¶

-mpcrel, -mno-pcrel¶

-mpopcntd, -mno-popcntd¶

-mpower10-vector, -mno-power10-vector¶

-mpower8-vector, -mno-power8-vector¶

-mpower9-vector, -mno-power9-vector¶

-mprefixed, -mno-prefixed¶

-mprivileged¶

-mrop-protect¶

-msecure-plt¶

-mspe, -mno-spe¶

-mvsx, -mno-vsx¶

WebAssembly ¶

-matomics, -mno-atomics¶

-mbulk-memory, -mno-bulk-memory¶

-mexception-handling, -mno-exception-handling¶

-mmultivalue, -mno-multivalue¶

-mmutable-globals, -mno-mutable-globals¶

-mnontrapping-fptoint, -mno-nontrapping-fptoint¶

-mreference-types, -mno-reference-types¶

-msign-ext, -mno-sign-ext¶

-msimd128, -mno-simd128¶

-mtail-call, -mno-tail-call¶

WebAssembly Driver ¶

-mexec-model=<arg>¶

Execution model (WebAssembly only)

X86 ¶

-m3dnow, -mno-3dnow¶

-m3dnowa, -mno-3dnowa¶

-madx, -mno-adx¶

-maes, -mno-aes¶

-mamx-bf16, -mno-amx-bf16¶

-mamx-int8, -mno-amx-int8¶

-mamx-tile, -mno-amx-tile¶

-mavx, -mno-avx¶

-mavx2, -mno-avx2¶

-mavx512bf16, -mno-avx512bf16¶

-mavx512bitalg, -mno-avx512bitalg¶

-mavx512bw, -mno-avx512bw¶

-mavx512cd, -mno-avx512cd¶

-mavx512dq, -mno-avx512dq¶

-mavx512er, -mno-avx512er¶

-mavx512f, -mno-avx512f¶

-mavx512ifma, -mno-avx512ifma¶

-mavx512pf, -mno-avx512pf¶

-mavx512vbmi, -mno-avx512vbmi¶

-mavx512vbmi2, -mno-avx512vbmi2¶

-mavx512vl, -mno-avx512vl¶

-mavx512vnni, -mno-avx512vnni¶

-mavx512vp2intersect, -mno-avx512vp2intersect¶

-mavx512vpopcntdq, -mno-avx512vpopcntdq¶

-mavxvnni, -mno-avxvnni¶

-mbmi, -mno-bmi¶

-mbmi2, -mno-bmi2¶

-mcldemote, -mno-cldemote¶

-mclflushopt, -mno-clflushopt¶

-mclwb, -mno-clwb¶

-mclzero, -mno-clzero¶

-mcx16, -mno-cx16¶

-menqcmd, -mno-enqcmd¶

-mf16c, -mno-f16c¶

-mfma, -mno-fma¶

-mfma4, -mno-fma4¶

-mfsgsbase, -mno-fsgsbase¶

-mfxsr, -mno-fxsr¶

-mgfni, -mno-gfni¶

-mhreset, -mno-hreset¶

-minvpcid, -mno-invpcid¶

-mkl, -mno-kl¶

-mlwp, -mno-lwp¶

-mlzcnt, -mno-lzcnt¶

-mmmx, -mno-mmx¶

-mmovbe, -mno-movbe¶

-mmovdir64b, -mno-movdir64b¶

-mmovdiri, -mno-movdiri¶

-mmwaitx, -mno-mwaitx¶

-mpclmul, -mno-pclmul¶

-mpconfig, -mno-pconfig¶

-mpku, -mno-pku¶

-mpopcnt, -mno-popcnt¶

-mprefetchwt1, -mno-prefetchwt1¶

-mprfchw, -mno-prfchw¶

-mptwrite, -mno-ptwrite¶

-mrdpid, -mno-rdpid¶

-mrdrnd, -mno-rdrnd¶

-mrdseed, -mno-rdseed¶

-mretpoline-external-thunk, -mno-retpoline-external-thunk¶

-mrtm, -mno-rtm¶

-msahf, -mno-sahf¶

-mserialize, -mno-serialize¶

-msgx, -mno-sgx¶

-msha, -mno-sha¶

-mshstk, -mno-shstk¶

-msse, -mno-sse¶

-msse2, -mno-sse2¶

-msse3, -mno-sse3¶

-msse4.1, -mno-sse4.1¶

-msse4.2, -mno-sse4.2, -msse4¶

-msse4a, -mno-sse4a¶

-mssse3, -mno-ssse3¶

-mtbm, -mno-tbm¶

-mtsxldtrk, -mno-tsxldtrk¶

-muintr, -mno-uintr¶

-mvaes, -mno-vaes¶

-mvpclmulqdq, -mno-vpclmulqdq¶

-mvzeroupper, -mno-vzeroupper¶

-mwaitpkg, -mno-waitpkg¶

-mwbnoinvd, -mno-wbnoinvd¶

-mwidekl, -mno-widekl¶

-mx87, -m80387, -mno-x87¶

-mxop, -mno-xop¶

-mxsave, -mno-xsave¶

-mxsavec, -mno-xsavec¶

-mxsaveopt, -mno-xsaveopt¶

-mxsaves, -mno-xsaves¶

RISCV ¶

-msave-restore, -mno-save-restore¶

Enable using library calls for save and restore

Long double flags ¶

Selects the long double implementation

-mlong-double-128¶

Force long double to be 128 bits

-mlong-double-64¶

Force long double to be 64 bits

-mlong-double-80¶

Force long double to be 80 bits, padded to 128 bits for storage

Optimization level ¶

Flags controlling how much optimization should be performed.

-O<arg>, -O (equivalent to -O1), --optimize, --optimize=<arg>¶

-Ofast<arg>¶

Debug information generation ¶

Flags controlling how much and what kind of debug information should be generated.

Kind and level of debug information ¶

-g, --debug, --debug=<arg>¶

Generate source-level debug information

-gdwarf¶

Generate source-level debug information with the default dwarf version

-gdwarf-2¶

Generate source-level debug information with dwarf version 2

-gdwarf-3¶

Generate source-level debug information with dwarf version 3

-gdwarf-4¶

Generate source-level debug information with dwarf version 4

-gdwarf-5¶

Generate source-level debug information with dwarf version 5

-gdwarf32¶

Enables DWARF32 format for ELF binaries, if debug information emission is enabled.

-gdwarf64¶

Enables DWARF64 format for ELF binaries, if debug information emission is enabled.

-gfull¶

-ginline-line-tables, -gno-inline-line-tables¶

-gused¶

Debug level ¶

-g0¶

-g2¶

-g3¶

-ggdb0¶

-ggdb1¶

-ggdb2¶

-ggdb3¶

-gline-directives-only¶

Emit debug line info directives only

-gline-tables-only, -g1, -gmlt¶

Emit debug line number tables only

-gmodules¶

Generate debug info with external references to clang modules or precompiled headers

Debugger to tune debug information for ¶

-gdbx¶

-ggdb¶

-glldb¶

-gsce¶

Debug information flags ¶

-gcolumn-info, -gno-column-info¶

-gdwarf-aranges¶

-gembed-source, -gno-embed-source¶

Embed source text in DWARF debug sections

-ggnu-pubnames, -gno-gnu-pubnames¶

-gpubnames, -gno-pubnames¶

-grecord-command-line, -gno-record-command-line, -grecord-gcc-switches¶

-gsplit-dwarf, -gno-split-dwarf¶

-gsplit-dwarf=<arg>¶

Set DWARF fission mode to either ‘split’ or ‘single’

-gstrict-dwarf, -gno-strict-dwarf¶

-gz=<arg>, -gz (equivalent to -gz=zlib)¶

DWARF debug sections compression type

Static analyzer flags ¶

Flags controlling the behavior of the Clang Static Analyzer.

-Xanalyzer <arg>¶

Pass <arg> to the static analyzer

Fortran compilation flags ¶

Flags that will be passed onto the gfortran compiler when Clang is given a Fortran input.

-A<arg>, --assert <arg>, --assert=<arg>¶

-A-<arg>¶

-faggressive-function-elimination, -fno-aggressive-function-elimination¶

-falign-commons, -fno-align-commons¶

-fall-intrinsics, -fno-all-intrinsics¶

-fautomatic, -fno-automatic¶

-fbacktrace, -fno-backtrace¶

-fblas-matmul-limit=<arg>¶

-fbounds-check, -fno-bounds-check¶

-fcheck-array-temporaries, -fno-check-array-temporaries¶

-fcheck=<arg>¶

-fcoarray=<arg>¶

-fconvert=<arg>¶

-fcray-pointer, -fno-cray-pointer¶

-fd-lines-as-code, -fno-d-lines-as-code¶

-fd-lines-as-comments, -fno-d-lines-as-comments¶

-fdollar-ok, -fno-dollar-ok¶

-fdump-fortran-optimized, -fno-dump-fortran-optimized¶

-fdump-fortran-original, -fno-dump-fortran-original¶

-fdump-parse-tree, -fno-dump-parse-tree¶

-fexternal-blas, -fno-external-blas¶

-ff2c, -fno-f2c¶

-ffpe-trap=<arg>¶

-ffree-line-length-<arg>¶

-ffrontend-optimize, -fno-frontend-optimize¶

-finit-character=<arg>¶

-finit-integer=<arg>¶

-finit-local-zero, -fno-init-local-zero¶

-finit-logical=<arg>¶

-finit-real=<arg>¶

-finteger-4-integer-8, -fno-integer-4-integer-8¶

-fmax-array-constructor=<arg>¶

-fmax-errors=<arg>¶

-fmax-identifier-length, -fno-max-identifier-length¶

-fmax-stack-var-size=<arg>¶

-fmax-subrecord-length=<arg>¶

-fmodule-private, -fno-module-private¶

-fpack-derived, -fno-pack-derived¶

-frange-check, -fno-range-check¶

-freal-4-real-10, -fno-real-4-real-10¶

-freal-4-real-16, -fno-real-4-real-16¶

-freal-4-real-8, -fno-real-4-real-8¶

-freal-8-real-10, -fno-real-8-real-10¶

-freal-8-real-16, -fno-real-8-real-16¶

-freal-8-real-4, -fno-real-8-real-4¶

-frealloc-lhs, -fno-realloc-lhs¶

-frecord-marker=<arg>¶

-frecursive, -fno-recursive¶

-frepack-arrays, -fno-repack-arrays¶

-fsecond-underscore, -fno-second-underscore¶

-fsign-zero, -fno-sign-zero¶

-fstack-arrays, -fno-stack-arrays¶

-funderscoring, -fno-underscoring¶

-fwhole-file, -fno-whole-file¶

-imultilib <arg>¶

-static-libgfortran¶

Linker flags ¶

Flags that are passed on to the linker

-L<dir>, --library-directory <arg>, --library-directory=<arg>¶

Add directory to library search path

-Mach¶

-T<script>¶

Specify <script> as linker script

-Tbss<addr>¶

Set starting address of BSS to <addr>

-Tdata<addr>¶

Set starting address of DATA to <addr>

-Ttext<addr>¶

Set starting address of TEXT to <addr>

-Wl,<arg>,<arg2>...¶

Pass the comma separated arguments in <arg> to the linker

-X¶

-Xlinker <arg>, --for-linker <arg>, --for-linker=<arg>¶

Pass <arg> to the linker

-Z¶

-coverage, --coverage¶

-e<arg>, --entry¶

-filelist <arg>¶

--hip-device-lib=<arg>¶

HIP device library

-l<arg>¶

--ld-path=<arg>¶

-nostartfiles¶

-nostdlib, --no-standard-libraries¶

-pie¶

-r¶

-rdynamic¶

--rocm-device-lib-path=<arg>, --hip-device-lib-path=<arg>¶

ROCm device library path. Alternative to rocm-path.

-rpath <arg>¶

-s¶

-shared, --shared¶

-specs=<arg>, --specs=<arg>¶

-static, --static¶

-static-pie¶

-t¶

-u<arg>, --force-link <arg>, --force-link=<arg>¶

-undef¶

undef all system defines

-undefined<arg>, --no-undefined¶

-z <arg>¶

Pass -z <arg> to the linker