CLANG(1) User Commands CLANG(1)

# [NAME](#NAME)

clang - Wolfi Clang 23.1.4-r1

# [DESCRIPTION](#DESCRIPTION)

OVERVIEW: clang LLVM compiler

USAGE: clang-23 \[options] file...

## [DEBUG/DEVELOPMENT OPTIONS:](#DEBUG/DEVELOPMENT_OPTIONS:)

[**-ccc-install-dir** &lt;value&gt;](#ccc)

Simulate installation in the given directory

**-ccc-print-bindings** Show bindings of tools to actions

[**-gen-cdb-fragment-path** &lt;value&gt;](#gen)

Emit a compilation database fragment to the specified directory

[**-gen-reproducer**](#gen~2)

Auto-generates preprocessed source files and a reproduction script

[**-via-file-asm**](#via)

Write assembly to file for input to assemble jobs

## [DRIVER OPTIONS:](#DRIVER_OPTIONS:)

[**-ccc-gcc-name** &lt;gcc-path&gt;](#ccc~2)

Name for native GCC compiler

**--driver-mode=**&lt;value&gt; Set the driver mode to either 'gcc', 'g++', 'cpp', 'cl' or 'flang'

**--rsp-quoting=**&lt;value&gt; Set the rsp quoting to either 'posix', or 'windows'

## [OPTIONS:](#OPTIONS:)

-###

Print (but do not run) the commands to run for this compilation

[**--analyzer-output** &lt;value&gt;](#analyzer)

Static analyzer report output format (html|plist|plist-multi-file|plist-html|sarif|sarif-html|text).

[**--analyze**](#analyze)

Run the static analyzer

[**-B** &lt;prefix&gt;](#B)

Search $prefix$file for executables, libraries, and data files. If $prefix is a directory, search $prefix/$file

[**-b** &lt;arg&gt;](#b)

Pass **-b** &lt;arg&gt; to the linker on AIX

[**-canonical-prefixes**](#canonical)

Use absolute paths for invoking subcommands (default)

[**-ccc-print-phases**](#ccc~3)

Dump list of actions to perform

[**-CC**](#CC)

Include comments from within macros in preprocessed output

[**-cl-denorms-are-zero**](#cl)

OpenCL only. Allow denormals to be flushed to zero.

[**-cl-ext=**&lt;value&gt;](#cl~2)

OpenCL only. Enable or disable OpenCL extensions/optional features. The argument is a comma-separated sequence of one or more extension names, each prefixed by '+' or '-'.

[**-cl-fast-relaxed-math**](#cl~3)

OpenCL only. Sets **-cl-finite-math-only** and **-cl-unsafe-math-optimizations**, and defines \_\_FAST\_RELAXED\_MATH\_\_.

[**-cl-finite-math-only**](#cl~4)

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

[**-cl-fp32-correctly-rounded-divide-sqrt**](#cl~5)

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

[**-cl-kernel-arg-info**](#cl~6)

OpenCL only. Generate kernel argument metadata.

[**-cl-mad-enable**](#cl~7)

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

[**-cl-no-signed-zeros**](#cl~8)

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

[**-cl-no-stdinc**](#cl~9)

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

[**-cl-opt-disable**](#cl~10)

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

[**-cl-single-precision-constant**](#cl~11)

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

[**-cl-std=**&lt;value&gt;](#cl~12)

OpenCL language standard to compile for.

[**-cl-strict-aliasing**](#cl~13)

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

[**-cl-uniform-work-group-size**](#cl~14)

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

[**-cl-unsafe-math-optimizations**](#cl~15)

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

**-clangir-disable-passes** Disable CIR transformations pipeline

[**-clangir-disable-verifier**](#clangir)

ClangIR: Disable MLIR module verifier

[**-clangir-enable-idiom-recognizer**](#clangir~2)

ClangIR: Enable Idiom Recognizer pass

[**-clangir-lib-opt=**&lt;value&gt;](#clangir~3)

Enable C/C++ library based optimizations (with options)

[**-clangir-lib-opt**](#clangir~4)

Enable C/C++ library based optimizations

[**--config-system-dir=**&lt;value&gt;](#config)

System directory for configuration files

[**--config-user-dir=**&lt;value&gt;](#config~2)

User directory for configuration files

[**--config=**&lt;file&gt;](#config=)

Specify configuration file

[**--cstdlib=**&lt;value&gt;](#cstdlib=)

C standard library to use

[**--cuda-compile-host-device**](#cuda)

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

[**--cuda-device-only**](#cuda~2)

Compile CUDA code for device only

[**--cuda-feature=**&lt;value&gt;](#cuda~3)

Manually specify the CUDA feature to use

[**--cuda-host-only**](#cuda~4)

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

[**--cuda-include-ptx=**&lt;value&gt;](#cuda~5)

Include PTX for the following GPU architecture (e.g. sm\_35) or 'all'. May be specified more than once.

[**--cuda-noopt-device-debug**](#cuda~6)

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

[**--cuda-path-ignore-env**](#cuda~7)

Ignore environment variables to detect CUDA installation

[**--cuda-path=**&lt;value&gt;](#cuda~8)

CUDA installation path

[**-cuid=**&lt;value&gt;](#cuid=)

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.

[**-cxx-isystem** &lt;directory&gt;](#cxx)

Add directory to the C++ SYSTEM include search path

[**-C**](#C)

Include comments in preprocessed output

[**-c**](#c)

Only run preprocess, compile, and assemble steps

[**-darwin-target-variant-triple** &lt;value&gt;](#darwin)

Specify the darwin target variant triple

[**-darwin-target-variant** &lt;value&gt;](#darwin~2)

Generate code for an additional runtime variant of the deployment target

[**-dD**](#dD)

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

**-dependency-dot** &lt;value&gt; Filename to write DOT-formatted header dependencies to

[**-dependency-file** &lt;value&gt;](#dependency)

Filename (or -) to write dependency output to

[**-dI**](#dI)

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

[**-dM**](#dM)

Print macro definitions in **-E** mode instead of normal output

[**-dsym-dir** &lt;dir&gt;](#dsym)

Directory to output dSYM's (if any) to

[**-dumpdir** &lt;dumppfx&gt;](#dumpdir)

Use &lt;dumpfpx&gt; as a prefix to form auxiliary and dump file names

[**-dumpmachine**](#dumpmachine)

Display the compiler's target processor

[**-dumpversion**](#dumpversion)

Display the version of the compiler

[**-D** &lt;macro&gt;=&lt;value&gt;](#D)

Define &lt;macro&gt; to &lt;value&gt; (or 1 if &lt;value&gt; omitted)

[**--embed-dir=**&lt;dir&gt;](#embed)

Add directory to embed search path

[**-emit-ast**](#emit)

Emit Clang AST files for source inputs

[**-emit-cir**](#emit~2)

Build ASTs and then lower to ClangIR

[**--emit-extension-symbol-graphs**](#emit~3)

Generate additional symbol graphs for extended modules.

[**-emit-interface-stubs**](#emit~4)

Generate Interface Stub Files.

[**-emit-llvm**](#emit~5)

Use the LLVM representation for assembler and object files

[**-emit-merged-ifs**](#emit~6)

Generate Interface Stub Files, emit merged text not binary.

[**--emit-static-lib**](#emit~7)

Enable linker job to emit a static library.

[**-emit-symbol-graph**](#emit~8)

Generate Extract API information as a side effect of compilation.

[**--end-no-unused-arguments**](#end)

Start emitting warnings for unused driver arguments

[**--extract-api-ignores=**&lt;value&gt;](#extract)

Comma separated list of files containing a new line separated list of API symbols to ignore when extracting API information.

[**-extract-api**](#extract~2)

Extract API information

[**-E**](#E)

Only run the preprocessor

[**-faapcs-bitfield-load**](#faapcs)

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

[**-faapcs-bitfield-width**](#faapcs~2)

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

[**-faarch64-jump-table-hardening**](#faarch64)

Use hardened lowering for jump-table dispatch

[**-faddrsig**](#faddrsig)

Emit an address-significance table

[**-falign-loops=**&lt;N&gt;](#falign)

N must be a power of two. Align loops to the boundary

[**-faligned-allocation**](#faligned)

Enable C++17 aligned allocation functions

[**-falloc-token-max=**&lt;N&gt;](#falloc)

Limit to maximum N allocation tokens (0 = target SIZE\_MAX)

[**-fallow-editor-placeholders**](#fallow)

Treat editor placeholders as valid source code

[**-fallow-runtime-check-skip-hot-cutoff=**&lt;value&gt;](#fallow~2)

Exclude \_\_builtin\_allow\_runtime\_check for the top hottest code responsible for the given fraction of PGO counters (0.0 \[default] = skip none; 1.0 = skip all). Argument format: &lt;value&gt;

[**-faltivec-src-compat=**&lt;value&gt;](#faltivec)

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**](#fansi)

Use ANSI escape codes for diagnostics

[**-fapinotes-modules**](#fapinotes)

Enable module-based external API notes support

[**-fapinotes-swift-version=**&lt;version&gt;](#fapinotes~2)

Specify the Swift version to use when filtering API notes

[**-fapinotes**](#fapinotes~3)

Enable external API notes support

[**-fapple-kext**](#fapple)

Use Apple's kernel extensions ABI

[**-fapple-link-rtlib**](#fapple~2)

Force linking the clang builtins runtime library

[**-fapple-pragma-pack**](#fapple~3)

Enable Apple gcc-compatible #pragma pack handling

**-fapplication-extension** Restrict code to those available for App Extensions

[**-fapprox-func**](#fapprox)

Allow certain math function calls to be replaced with an approximately equivalent calculation

[**-fassume-nothrow-exception-dtor**](#fassume)

Assume that exception objects' destructors are non-throwing

[**-fasync-exceptions**](#fasync)

Enable EH Asynchronous exceptions

[**-fatomic-fine-grained-memory**](#fatomic)

May have atomic operations on fine-grained memory

[**-fatomic-ignore-denormal-mode**](#fatomic~2)

Allow atomic operations to ignore denormal mode

[**-fatomic-remote-memory**](#fatomic~3)

May have atomic operations on remote memory

[**-fauto-import**](#fauto)

MinGW specific. Enable code generation support for automatic dllimport, and enable support for it in the linker. Enabled by default.

[**-fbasic-block-address-map**](#fbasic)

Emit the basic block address map section.

[**-fbasic-block-sections=**&lt;value&gt;](#fbasic~2)

Place each function's basic blocks in unique sections (ELF Only)

[**-fbinutils-version=**&lt;major.minor&gt;](#fbinutils)

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**](#fblocks)

Enable the 'blocks' language feature

[**-fborland-extensions**](#fborland)

Accept non-standard constructs supported by the Borland compiler

[**-fbuild-session-file=**&lt;file&gt;](#fbuild)

Use the last modification time of &lt;file&gt; as the build session timestamp

[**-fbuild-session-timestamp=**&lt;time since Epoch in seconds&gt;](#fbuild~2)

Time when the current build session started

[**-fbuiltin-module-map**](#fbuiltin)

Load the clang builtins module map file.

[**-fc**++-abi=&lt;value&gt;](#fc)

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

[**-fc**++-static-destructors=&lt;value&gt;](#fc~2)

Controls which variables C++ static destructors are registered for

[**-fcall-saved-x10**](#fcall)

Make the x10 register call-saved (AArch64 only)

[**-fcall-saved-x11**](#fcall~2)

Make the x11 register call-saved (AArch64 only)

[**-fcall-saved-x12**](#fcall~3)

Make the x12 register call-saved (AArch64 only)

[**-fcall-saved-x13**](#fcall~4)

Make the x13 register call-saved (AArch64 only)

[**-fcall-saved-x14**](#fcall~5)

Make the x14 register call-saved (AArch64 only)

[**-fcall-saved-x15**](#fcall~6)

Make the x15 register call-saved (AArch64 only)

[**-fcall-saved-x18**](#fcall~7)

Make the x18 register call-saved (AArch64 only)

[**-fcall-saved-x8**](#fcall~8)

Make the x8 register call-saved (AArch64 only)

[**-fcall-saved-x9**](#fcall~9)

Make the x9 register call-saved (AArch64 only)

[**-fcaret-diagnostics-max-lines=**&lt;value&gt;](#fcaret)

Set the maximum number of source lines to show in a caret diagnostic (0 = no limit).

**-fcf-protection=**&lt;value&gt; Instrument control-flow architecture protection

[**-fcf-protection**](#fcf)

Enable cf-protection in 'full' mode

[**-fchar8\_t**](#fchar8_t)

Enable C++ builtin type char8\_t

[**-fcheck-new**](#fcheck)

Do not assume C++ operator new may not return NULL

[**-fclang-abi-compat=**&lt;version&gt;](#fclang)

Attempt to match the ABI of Clang &lt;version&gt;

[**-fclangir**](#fclangir)

Use the ClangIR pipeline to compile

[**-fcodegen-data-generate=**&lt;path&gt;](#fcodegen)

Emit codegen data into the object file. LLD for MachO (currently) merges them into the specified &lt;path&gt;.

**-fcodegen-data-generate** Emit codegen data into the object file. LLD for MachO (currently) merges them into default.cgdata.

[**-fcodegen-data-use=**&lt;path&gt;](#fcodegen~2)

Use codegen data read from the specified &lt;path&gt;.

[**-fcodegen-data-use**](#fcodegen~3)

Use codegen data read from default.cgdata to optimize the binary

[**-fcolor-diagnostics**](#fcolor)

Enable colors in diagnostics

[**-fcomment-block-commands=**&lt;arg&gt;](#fcomment)

Treat each comma separated argument in &lt;arg&gt; as a documentation comment block command

[**-fcommon**](#fcommon)

Place uninitialized global variables in a common block

[**-fcomplete-member-pointers**](#fcomplete)

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

[**-fcomplex-arithmetic=**&lt;value&gt;](#fcomplex)

Controls the calculation methods of complex number multiplication and division.

[**-fconstant-nsarray-literals**](#fconstant)

Enable compile-time constant NSArray literals

[**-fconstant-nsdictionary-literals**](#fconstant~2)

Enable compile-time constant NSDictionary literals

[**-fconstant-nsnumber-literals**](#fconstant~3)

Enable compile-time constant NSNumber literals

[**-fconstexpr-backtrace-limit=**&lt;value&gt;](#fconstexpr)

Set the maximum number of entries to print in a constexpr evaluation backtrace (0 = no limit)

[**-fconstexpr-depth=**&lt;value&gt;](#fconstexpr~2)

Set the maximum depth of recursive constexpr function calls

[**-fconstexpr-steps=**&lt;value&gt;](#fconstexpr~3)

Set the maximum number of steps in constexpr function evaluation (0 = no limit)

[**-fconvergent-functions**](#fconvergent)

Assume all functions may be convergent.

[**-fcoro-aligned-allocation**](#fcoro)

Prefer aligned allocation for C++ Coroutines

[**-fcoroutines**](#fcoroutines)

Enable support for the C++ Coroutines

[**-fcoverage-compilation-dir=**&lt;value&gt;](#fcoverage)

The compilation directory to embed in the coverage mapping.

[**-fcoverage-mapping**](#fcoverage~2)

Generate coverage mapping to enable code coverage analysis

[**-fcoverage-mcdc**](#fcoverage~3)

Enable MC/DC criteria when generating code coverage

[**-fcoverage-prefix-map=**&lt;old&gt;=&lt;new&gt;](#fcoverage~4)

remap file source paths &lt;old&gt; to &lt;new&gt; in coverage mapping. If there are multiple options, prefix replacement is applied in reverse order starting from the last one

[**-fcrash-diagnostics-dir=**&lt;dir&gt;](#fcrash)

Put crash-report files in &lt;dir&gt;

[**-fcrash-diagnostics-tar=**&lt;path&gt;](#fcrash~2)

Put crash-report tarball at &lt;path&gt;

[**-fcrash-diagnostics=**&lt;value&gt;](#fcrash~3)

Set level of crash diagnostic reporting, (option: off, compiler, all)

[**-fcrash-diagnostics**](#fcrash~4)

Enable crash diagnostic reporting (default)

[**-fcs-profile-generate=**&lt;directory&gt;](#fcs)

Generate instrumented code to collect context sensitive execution counts into &lt;directory&gt;/default.profraw (overridden by LLVM\_PROFILE\_FILE env var)

[**-fcs-profile-generate**](#fcs~2)

Generate instrumented code to collect context sensitive execution counts into default.profraw (overridden by LLVM\_PROFILE\_FILE env var)

[**-fcuda-short-ptr**](#fcuda)

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

[**-fcx-fortran-rules**](#fcx)

Range reduction is enabled for complex arithmetic operations.

[**-fcx-limited-range**](#fcx~2)

Basic algebraic expansions of complex arithmetic operations involving are enabled.

[**-fcxx-exceptions**](#fcxx)

Enable C++ exceptions

[**-fcxx-modules**](#fcxx~2)

Enable modules for C++

[**-fdata-sections**](#fdata)

Place each data in its own section

[**-fdebug-compilation-dir=**&lt;value&gt;](#fdebug)

The compilation directory to embed in the debug info

[**-fdebug-default-version=**&lt;value&gt;](#fdebug~2)

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

[**-fdebug-info-for-profiling**](#fdebug~3)

Emit extra debug info to make sample profile more accurate

[**-fdebug-macro**](#fdebug~4)

Emit macro debug information

[**-fdebug-prefix-map=**&lt;old&gt;=&lt;new&gt;](#fdebug~5)

For paths in debug info, remap directory &lt;old&gt; to &lt;new&gt;. If multiple options match a path, the last option wins

[**-fdebug-ranges-base-address**](#fdebug~6)

Use DWARF base address selection entries in .debug\_ranges

[**-fdebug-record-sysroot**](#fdebug~7)

Record the sysroot path in debug info as DW\_AT\_LLVM\_sysroot and in DW\_AT\_LLVM\_include\_path (used by LLDB to locate the SDK for Clang module imports).

[**-fdebug-types-section**](#fdebug~8)

Place debug types in their own section (ELF Only)

[**-fdeclspec**](#fdeclspec)

Allow \_\_declspec as a keyword

[**-fdefer-ts**](#fdefer)

Enable support for the C '\_Defer' Technical Specification

[**-fdefine-target-os-macros**](#fdefine)

Enable predefined target OS macros

[**-fdelayed-template-parsing**](#fdelayed)

Parse templated function definitions at the end of the translation unit

[**-fdelete-null-pointer-checks**](#fdelete)

Treat usage of null pointers as undefined behavior (default)

[**-fdevirtualize-speculatively**](#fdevirtualize)

Enables speculative devirtualization optimization.

[**-fdiagnostics-absolute-paths**](#fdiagnostics)

Print absolute paths in diagnostics

[**-fdiagnostics-color=**&lt;value&gt;](#fdiagnostics~2)

When to use colors in diagnostics

[**-fdiagnostics-hotness-threshold=**&lt;value&gt;](#fdiagnostics~3)

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-misexpect-tolerance=**&lt;value&gt;](#fdiagnostics~4)

Prevent misexpect diagnostics from being output if the profile counts are within N% of the expected.

[**-fdiagnostics-parseable-fixits**](#fdiagnostics~5)

Print fix-its in machine parseable form

[**-fdiagnostics-print-source-range-info**](#fdiagnostics~6)

Print source range spans in numeric form

[**-fdiagnostics-show-hotness**](#fdiagnostics~7)

Enable profile hotness information in diagnostic line

[**-fdiagnostics-show-inlining-chain**](#fdiagnostics~8)

Show inlining chain notes for \[\[gnu::warning/error]] diagnostics

[**-fdiagnostics-show-note-include-stack**](#fdiagnostics~9)

Display include stacks for diagnostic notes

[**-fdiagnostics-show-option**](#fdiagnostics~10)

Print option name with mappable diagnostics

[**-fdiagnostics-show-template-tree**](#fdiagnostics~11)

Print a template comparison tree for differing templates

[**-fdigraphs**](#fdigraphs)

Enable alternative token representations '&lt;:', ':&gt;', '&lt;%', '%&gt;', '%:', '%:%:' (default)

[**-fdirect-access-external-data**](#fdirect)

Don't use GOT indirection to reference external data symbols

[**-fdisable-block-signature-string**](#fdisable)

Disable block signature string)

[**-fdiscard-value-names**](#fdiscard)

Discard value names in LLVM IR

[**-fdollars-in-identifiers**](#fdollars)

Allow '$' in identifiers

[**-fdriver-only**](#fdriver)

Only run the driver.

[**-fdwarf-exceptions**](#fdwarf)

Use DWARF style exceptions

[**-feliminate-unused-debug-types**](#feliminate)

Do not emit debug info for defined but unused types

[**-fembed-bitcode-marker**](#fembed)

Embed placeholder LLVM IR data as a marker

[**-fembed-bitcode=**&lt;option&gt;](#fembed~2)

Embed LLVM bitcode

[**-fembed-bitcode**](#fembed~3)

Embed LLVM IR bitcode as data

[**-fembed-offload-object=**&lt;value&gt;](#fembed~4)

Embed Offloading device-side binary into host object file as a section.

[**-femit-all-decls**](#femit)

Emit all declarations, even if unused

[**-femit-compact-unwind-non-canonical**](#femit~2)

Try emitting Compact-Unwind for non-canonical entries. Maybe overridden by other constraints

[**-femit-dwarf-unwind=**&lt;value&gt;](#femit~3)

When to emit DWARF unwind (EH frame) info

[**-femulated-tls**](#femulated)

Use emutls functions to access thread\_local variables

[**-fenable-matrix**](#fenable)

Enable matrix data type and related builtin functions

**-fenable-new-pm-codegen** Use the NewPM for the Codegen Pipeline

[**-fexceptions**](#fexceptions)

Enable support for exception handling

[**-fexcess-precision=**&lt;value&gt;](#fexcess)

Allows control over excess precision on targets where native support for the precision types is not available. By default, excess precision is used to calculate intermediate results following the rules specified in ISO C99.

[**-fexperimental-allow-pointer-field-protection-attr**](#fexperimental)

Allow the use of the \[\[clang::pointer\_field\_protection]] attribute

[**-fexperimental-call-graph-section**](#fexperimental~2)

Emit a call graph section

[**-fexperimental-late-parse-attributes**](#fexperimental~3)

Enable experimental late parsing of attributes

[**-fexperimental-library**](#fexperimental~4)

Control whether unstable and experimental library features are enabled. This option enables various library features that are either experimental (also known as TSes), or have been but are not stable yet in the selected Standard Library implementation. It is not recommended to use this option in production code, since neither ABI nor API stability are guaranteed. This is intended to provide a preview of features that will ship in the future for experimentation purposes

[**-fexperimental-loop-fusion**](#fexperimental~5)

EnableEnable the loop fusion pass

[**-fexperimental-openacc-macro-override** &lt;value&gt;](#fexperimental~6)

Overrides the \_OPENACC macro value for experimental testing during OpenACC support development

[**-fexperimental-pointer-field-protection-abi**](#fexperimental~7)

Enable pointer field protection on all non-standard layout struct types

[**-fexperimental-pointer-field-protection-tagged**](#fexperimental~8)

Use pointer identity (tag) to discriminate pointers of non-trivially-copyable types

[**-fexperimental-relative-c**++-abi-vtables](#fexperimental~9)

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

[**-fexperimental-sanitize-metadata-ignorelist=**&lt;value&gt;](#fexperimental~10)

Disable sanitizer metadata for modules and functions that match the provided special case list

[**-fexperimental-sanitize-metadata=**&lt;value&gt;](#fexperimental~11)

Specify the type of metadata to emit for binary analysis sanitizers

[**-fexperimental-strict-floating-point**](#fexperimental~12)

Enables the use of non-default rounding modes and non-default exception handling on targets that are not currently ready.

[**-fextend-arguments=**&lt;value&gt;](#fextend)

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

[**-fextend-variable-liveness=**&lt;value&gt;](#fextend~2)

Extend the liveness of user variables through optimizations to prevent stale or optimized-out variable values when debugging.

[**-fextend-variable-liveness**](#fextend~3)

Alias for **-fextend-variable-liveness**=*all*.

[**-ffast-math**](#ffast)

Allow aggressive, lossy floating-point optimizations

[**-ffat-lto-objects**](#ffat)

Enable fat LTO object support

[**-ffile-compilation-dir=**&lt;value&gt;](#ffile)

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

[**-ffile-prefix-map=**&lt;value&gt;](#ffile~2)

remap file source paths in debug info, coverage mapping, predefined preprocessor macros and \_\_builtin\_FILE(). Implies **-ffile-reproducible**.

[**-ffile-reproducible**](#ffile~3)

Use the target's platform-specific path separator character when expanding the \_\_FILE\__ macro

[**-ffine-grained-bitfield-accesses**](#ffine)

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

[**-ffinite-loops**](#ffinite)

Assume all non-trivial loops are finite.

[**-ffinite-math-only**](#ffinite~2)

Allow floating-point optimizations that assume arguments and results are not NaNs or +-inf. This defines the \\\_\\\_FINITE\\\_MATH\\\_ONLY\\\_\\_ preprocessor macro.

[**-ffixed-a0**](#ffixed)

Reserve the a0 register (M68k only)

[**-ffixed-a1**](#ffixed~2)

Reserve the a1 register (M68k only)

[**-ffixed-a2**](#ffixed~3)

Reserve the a2 register (M68k only)

[**-ffixed-a3**](#ffixed~4)

Reserve the a3 register (M68k only)

[**-ffixed-a4**](#ffixed~5)

Reserve the a4 register (M68k only)

[**-ffixed-a5**](#ffixed~6)

Reserve the a5 register (M68k only)

[**-ffixed-a6**](#ffixed~7)

Reserve the a6 register (M68k only)

[**-ffixed-d0**](#ffixed~8)

Reserve the d0 register (M68k only)

[**-ffixed-d1**](#ffixed~9)

Reserve the d1 register (M68k only)

[**-ffixed-d2**](#ffixed~10)

Reserve the d2 register (M68k only)

[**-ffixed-d3**](#ffixed~11)

Reserve the d3 register (M68k only)

[**-ffixed-d4**](#ffixed~12)

Reserve the d4 register (M68k only)

[**-ffixed-d5**](#ffixed~13)

Reserve the d5 register (M68k only)

[**-ffixed-d6**](#ffixed~14)

Reserve the d6 register (M68k only)

[**-ffixed-d7**](#ffixed~15)

Reserve the d7 register (M68k only)

[**-ffixed-edi**](#ffixed~16)

Reserve the edi register (x86 only)

[**-ffixed-g1**](#ffixed~17)

Reserve the G1 register (SPARC only)

[**-ffixed-g2**](#ffixed~18)

Reserve the G2 register (SPARC only)

[**-ffixed-g3**](#ffixed~19)

Reserve the G3 register (SPARC only)

[**-ffixed-g4**](#ffixed~20)

Reserve the G4 register (SPARC only)

[**-ffixed-g5**](#ffixed~21)

Reserve the G5 register (SPARC only)

[**-ffixed-g6**](#ffixed~22)

Reserve the G6 register (SPARC only)

[**-ffixed-g7**](#ffixed~23)

Reserve the G7 register (SPARC only)

[**-ffixed-i0**](#ffixed~24)

Reserve the I0 register (SPARC only)

[**-ffixed-i1**](#ffixed~25)

Reserve the I1 register (SPARC only)

[**-ffixed-i2**](#ffixed~26)

Reserve the I2 register (SPARC only)

[**-ffixed-i3**](#ffixed~27)

Reserve the I3 register (SPARC only)

[**-ffixed-i4**](#ffixed~28)

Reserve the I4 register (SPARC only)

[**-ffixed-i5**](#ffixed~29)

Reserve the I5 register (SPARC only)

[**-ffixed-l0**](#ffixed~30)

Reserve the L0 register (SPARC only)

[**-ffixed-l1**](#ffixed~31)

Reserve the L1 register (SPARC only)

[**-ffixed-l2**](#ffixed~32)

Reserve the L2 register (SPARC only)

[**-ffixed-l3**](#ffixed~33)

Reserve the L3 register (SPARC only)

[**-ffixed-l4**](#ffixed~34)

Reserve the L4 register (SPARC only)

[**-ffixed-l5**](#ffixed~35)

Reserve the L5 register (SPARC only)

[**-ffixed-l6**](#ffixed~36)

Reserve the L6 register (SPARC only)

[**-ffixed-l7**](#ffixed~37)

Reserve the L7 register (SPARC only)

[**-ffixed-o0**](#ffixed~38)

Reserve the O0 register (SPARC only)

[**-ffixed-o1**](#ffixed~39)

Reserve the O1 register (SPARC only)

[**-ffixed-o2**](#ffixed~40)

Reserve the O2 register (SPARC only)

[**-ffixed-o3**](#ffixed~41)

Reserve the O3 register (SPARC only)

[**-ffixed-o4**](#ffixed~42)

Reserve the O4 register (SPARC only)

[**-ffixed-o5**](#ffixed~43)

Reserve the O5 register (SPARC only)

[**-ffixed-point**](#ffixed~44)

Enable fixed point types

[**-ffixed-r10**](#ffixed~45)

Reserve the r10 register (Hexagon/x86\_64 only)

[**-ffixed-r11**](#ffixed~46)

Reserve the r11 register (Hexagon/x86\_64 only)

[**-ffixed-r12**](#ffixed~47)

Reserve the r12 register (Hexagon/x86\_64 only)

[**-ffixed-r13**](#ffixed~48)

Reserve the r13 register (Hexagon/x86\_64 only)

[**-ffixed-r14**](#ffixed~49)

Reserve the r14 register (Hexagon/x86\_64 only)

[**-ffixed-r15**](#ffixed~50)

Reserve the r15 register (Hexagon/x86\_64 only)

[**-ffixed-r16**](#ffixed~51)

Reserve the r16 register (Hexagon/x86\_64 only; x86\_64 requires APX EGPR)

[**-ffixed-r17**](#ffixed~52)

Reserve the r17 register (Hexagon/x86\_64 only; x86\_64 requires APX EGPR)

[**-ffixed-r18**](#ffixed~53)

Reserve the r18 register (Hexagon/x86\_64 only; x86\_64 requires APX EGPR)

[**-ffixed-r19**](#ffixed~54)

Reserve register r19 (Hexagon/x86\_64 only; x86\_64 requires APX EGPR)

[**-ffixed-r20**](#ffixed~55)

Reserve the r20 register (Hexagon/x86\_64 only; x86\_64 requires APX EGPR)

[**-ffixed-r21**](#ffixed~56)

Reserve the r21 register (Hexagon/x86\_64 only; x86\_64 requires APX EGPR)

[**-ffixed-r22**](#ffixed~57)

Reserve the r22 register (Hexagon/x86\_64 only; x86\_64 requires APX EGPR)

[**-ffixed-r23**](#ffixed~58)

Reserve the r23 register (Hexagon/x86\_64 only; x86\_64 requires APX EGPR)

[**-ffixed-r24**](#ffixed~59)

Reserve the r24 register (Hexagon/x86\_64 only; x86\_64 requires APX EGPR)

[**-ffixed-r25**](#ffixed~60)

Reserve the r25 register (Hexagon/x86\_64 only; x86\_64 requires APX EGPR)

[**-ffixed-r26**](#ffixed~61)

Reserve the r26 register (Hexagon/x86\_64 only; x86\_64 requires APX EGPR)

[**-ffixed-r27**](#ffixed~62)

Reserve the r27 register (Hexagon/x86\_64 only; x86\_64 requires APX EGPR)

[**-ffixed-r28**](#ffixed~63)

Reserve the r28 register (Hexagon/x86\_64 only; x86\_64 requires APX EGPR)

[**-ffixed-r29**](#ffixed~64)

Reserve the r29 register (Hexagon/x86\_64 only; x86\_64 requires APX EGPR)

[**-ffixed-r30**](#ffixed~65)

Reserve the r30 register (Hexagon/x86\_64 only; x86\_64 requires APX EGPR)

[**-ffixed-r31**](#ffixed~66)

Reserve the r31 register (Hexagon/x86\_64 only; x86\_64 requires APX EGPR)

[**-ffixed-r6**](#ffixed~67)

Reserve register r6 (Hexagon only)

[**-ffixed-r7**](#ffixed~68)

Reserve register r7 (Hexagon only)

[**-ffixed-r8**](#ffixed~69)

Reserve the r8 register (Hexagon/x86\_64 only)

[**-ffixed-r9**](#ffixed~70)

Reserve the r9 register (ARM/Hexagon/x86\_64 only)

[**-ffixed-x10**](#ffixed~71)

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

[**-ffixed-x11**](#ffixed~72)

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

[**-ffixed-x12**](#ffixed~73)

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

[**-ffixed-x13**](#ffixed~74)

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

[**-ffixed-x14**](#ffixed~75)

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

[**-ffixed-x15**](#ffixed~76)

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

[**-ffixed-x16**](#ffixed~77)

Reserve the x16 register (RISC-V only)

[**-ffixed-x17**](#ffixed~78)

Reserve the x17 register (RISC-V only)

[**-ffixed-x18**](#ffixed~79)

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

[**-ffixed-x19**](#ffixed~80)

Reserve the x19 register (RISC-V only)

[**-ffixed-x1**](#ffixed~81)

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

[**-ffixed-x20**](#ffixed~82)

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

[**-ffixed-x21**](#ffixed~83)

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

[**-ffixed-x22**](#ffixed~84)

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

[**-ffixed-x23**](#ffixed~85)

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

[**-ffixed-x24**](#ffixed~86)

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

[**-ffixed-x25**](#ffixed~87)

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

[**-ffixed-x26**](#ffixed~88)

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

[**-ffixed-x27**](#ffixed~89)

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

[**-ffixed-x28**](#ffixed~90)

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

[**-ffixed-x29**](#ffixed~91)

Reserve the x29 register (RISC-V only)

[**-ffixed-x2**](#ffixed~92)

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

[**-ffixed-x30**](#ffixed~93)

Reserve the x30 register (RISC-V only)

[**-ffixed-x31**](#ffixed~94)

Reserve the x31 register (RISC-V only)

[**-ffixed-x3**](#ffixed~95)

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

[**-ffixed-x4**](#ffixed~96)

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

[**-ffixed-x5**](#ffixed~97)

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

[**-ffixed-x6**](#ffixed~98)

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

[**-ffixed-x7**](#ffixed~99)

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

[**-ffixed-x8**](#ffixed~100)

Reserve the x8 register (RISC-V only)

[**-ffixed-x9**](#ffixed~101)

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

[**-fforce-check-cxx20-modules-input-files**](#fforce)

Check the input source files from C++20 modules explicitly

[**-fforce-dwarf-frame**](#fforce~2)

Always emit a debug frame section

[**-fforce-emit-vtables**](#fforce~3)

Emits more virtual tables to improve devirtualization

[**-fforce-enable-int128**](#fforce~4)

Enable support for int128\_t type

[**-ffp-contract=**&lt;value&gt;](#ffp)

Form fused FP ops (e.g. FMAs)

[**-ffp-eval-method=**&lt;value&gt;](#ffp~2)

Specifies the evaluation method to use for floating-point arithmetic.

[**-ffp-exception-behavior=**&lt;value&gt;](#ffp~3)

Specifies the exception behavior of floating-point operations.

[**-ffp-model=**&lt;value&gt;](#ffp~4)

Controls the semantics of floating-point calculations.

[**-ffreestanding**](#ffreestanding)

Assert that the compilation takes place in a freestanding environment

[**-ffuchsia-api-level=**&lt;value&gt;](#ffuchsia)

Set Fuchsia API level

[**-ffunction-sections**](#ffunction)

Place each function in its own section

[**-fglobal-isel**](#fglobal)

Enables the global instruction selector

[**-fgnu-keywords**](#fgnu)

Allow GNU-extension keywords regardless of language standard

[**-fgnu-runtime**](#fgnu~2)

Generate output compatible with the standard GNU Objective-C runtime

[**-fgnu89-inline**](#fgnu89)

Use the gnu89 inline semantics

[**-fgnuc-version=**&lt;value&gt;](#fgnuc)

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

**-fgpu-allow-device-init** Allow device side init function in HIP (experimental)

[**-fgpu-approx-transcendentals**](#fgpu)

Use approximate transcendental functions

[**-fgpu-default-stream=**&lt;value&gt;](#fgpu~2)

Specify default stream. The default value is 'legacy'. (CUDA/HIP only)

[**-fgpu-defer-diag**](#fgpu~3)

Defer host/device related diagnostic messages for CUDA/HIP

[**-fgpu-exclude-wrong-side-overloads**](#fgpu~4)

Always exclude wrong side overloads in overloading resolution for CUDA/HIP

[**-fgpu-flush-denormals-to-zero**](#fgpu~5)

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

[**-fgpu-inline-threshold=**&lt;value&gt;](#fgpu~6)

Inline threshold for device compilation for CUDA/HIP

[**-fgpu-rdc**](#fgpu~7)

Generate relocatable device code, also known as separate compilation mode

[**-fgpu-sanitize**](#fgpu~8)

Enable sanitizer for supported offloading devices

[**-fheinous-gnu-extensions**](#fheinous)

(Deprecated) Controls whether '-Winvalid-gnu-asm-cast' defaults to an error or a warning

[**-fhip-emit-relocatable**](#fhip)

Compile HIP source to relocatable

[**-fhip-fp32-correctly-rounded-divide-sqrt**](#fhip~2)

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

[**-fhip-kernel-arg-name**](#fhip~3)

Specify that kernel argument names are preserved (HIP only)

[**-fhip-new-launch-api**](#fhip~4)

Use new kernel launching API for HIP

[**-fhlsl-strict-availability**](#fhlsl)

Enables strict availability diagnostic mode for HLSL built-in functions.

[**-fhonor-infinities**](#fhonor)

Specify that floating-point optimizations are not allowed that assume arguments and results are not +-inf.

[**-fhonor-nans**](#fhonor~2)

Specify that floating-point optimizations are not allowed that assume arguments and results are not NANs.

[**-fignore-exceptions**](#fignore)

Enable support for ignoring exception handling constructs

[**-fimplicit-module-maps**](#fimplicit)

Implicitly search the file system for module map files.

[**-fimplicit-modules-lock-timeout=**&lt;seconds&gt;](#fimplicit~2)

Time to wait on an implicit module lock before timing out

[**-fincremental-extensions**](#fincremental)

Enable incremental processing extensions such as processing statements on the global scope.

[**-finline-functions**](#finline)

Inline suitable functions

**-finline-hint-functions** Inline functions which are (explicitly or implicitly) marked inline

[**-finline-max-stacksize=**&lt;value&gt;](#finline~2)

Suppress inlining of functions whose stack size exceeds the given value

**-finput-charset=**&lt;value&gt; Specify the default character set for source files

[**-finstrument-function-entry-bare**](#finstrument)

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

[**-finstrument-functions-after-inlining**](#finstrument~2)

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

[**-finstrument-functions**](#finstrument~3)

Generate calls to instrument function entry and exit

[**-fintegrated-as**](#fintegrated)

Enable the integrated assembler

[**-fintegrated-cc1**](#fintegrated~2)

Run cc1 in-process

**-fintegrated-objemitter** Use internal machine object code emitter.

[**-fjmc**](#fjmc)

Enable just-my-code debugging

[**-fjump-tables**](#fjump)

Use jump tables for lowering switches

[**-fkeep-persistent-storage-variables**](#fkeep)

Enable keeping all variables that have a persistent storage duration, including global, static and thread-local variables, to guarantee that they can be directly addressed

[**-fkeep-static-consts**](#fkeep~2)

Keep static const variables even if unused

[**-fkeep-system-includes**](#fkeep~3)

Instead of expanding system headers when emitting preprocessor output, preserve the #include directive. Useful when producing preprocessed output for test case reduction. May produce incorrect output if preprocessor symbols that control the included content (e.g. \_XOPEN\_SOURCE) are defined in the including source file. The portability of the resulting source to other compilation environments is not guaranteed.

Only valid with **-E**.

[**-flax-vector-conversions=**&lt;value&gt;](#flax)

Enable implicit vector bit-casts

[**-flifetime-dse**](#flifetime)

Aggressively delete stores after the end of an object's lifetime

[**-floop-interchange**](#floop)

Enable the loop interchange pass

[**-flto-jobs=**&lt;value&gt;](#flto)

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-partitions=**&lt;value&gt;](#flto~2)

Number of partitions to use for parallel full LTO codegen, ld.lld only.

[**-flto**=*auto*](#flto~3)

Enable LTO in 'full' mode

[**-flto**=*jobserver*](#flto~4)

Enable LTO in 'full' mode

[**-flto=**&lt;value&gt;](#flto=)

Set LTO mode

[**-flto**](#flto~5)

Enable LTO in 'full' mode

[**-fmacro-backtrace-limit=**&lt;value&gt;](#fmacro)

Set the maximum number of entries to print in a macro expansion backtrace (0 = no limit)

[**-fmacro-prefix-map=**&lt;value&gt;](#fmacro~2)

remap file source paths in predefined preprocessor macros and \_\_builtin\_FILE(). Implies **-ffile-reproducible**.

[**-fmath-errno**](#fmath)

Require math functions to indicate errors by setting errno

[**-fmatrix-memory-layout=**&lt;value&gt;](#fmatrix)

Sets the matrix memory layout (row-major or column-major)

[**-fmax-tokens=**&lt;value&gt;](#fmax)

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

[**-fmax-type-align=**&lt;value&gt;](#fmax~2)

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

[**-fmemory-profile-use=**&lt;pathname&gt;](#fmemory)

Use memory profile for profile-guided memory optimization

[**-fmemory-profile=**&lt;directory&gt;](#fmemory~2)

Enable heap memory profiling and dump results into &lt;directory&gt;

[**-fmemory-profile**](#fmemory~3)

Enable heap memory profiling

[**-fmerge-all-constants**](#fmerge)

Allow merging of constants

[**-fmessage-length=**&lt;value&gt;](#fmessage)

Format message diagnostics so that they fit within N columns

[**-fminimize-whitespace**](#fminimize)

Ignore the whitespace from the input file when emitting preprocessor output. It will only contain whitespace when necessary, e.g. to keep two minus signs from merging into to an increment operator. Useful with the **-P** option to normalize whitespace such that two files with only formatting changes are equal.

Only valid with **-E** on C-like inputs and incompatible with **-traditional-cpp**.

[**-fmodule-file-deps=**&lt;value&gt;](#fmodule)

Include module files in dependency output

[**-fmodule-file**=*\[*&lt;name&gt;=\]&lt;file&gt;](#fmodule~2)

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

[**-fmodule-header=**&lt;kind&gt;](#fmodule~3)

Build a C++20 Header Unit from a header that should be found in the user (fmodule-header=user) or system (fmodule-header=system) search path.

[**-fmodule-header**](#fmodule~4)

Build a C++20 Header Unit from a header

[**-fmodule-map-file=**&lt;file&gt;](#fmodule~5)

Load this module map file

[**-fmodule-name=**&lt;name&gt;](#fmodule~6)

Specify the name of the module to build

**-fmodule-output=**&lt;value&gt; Save intermediate module file results when compiling a standard C++ module unit.

[**-fmodule-output**](#fmodule~7)

Save intermediate module file results when compiling a standard C++ module unit.

[**-fmodulemap-allow-subdirectory-search**](#fmodulemap)

Allow to search for module maps in subdirectories of search paths

[**-fmodules-cache-path=**&lt;directory&gt;](#fmodules)

Specify the module cache path

[**-fmodules-decluse**](#fmodules~2)

Require declaration of modules used within a module

[**-fmodules-disable-diagnostic-validation**](#fmodules~3)

Disable validation of the diagnostic options when loading the module

[**-fmodules-driver**](#fmodules~4)

Enable support for driver managed module builds (experimental)

[**-fmodules-embed-all-files**&lt;value&gt;](#fmodules~5)

Embed the contents of all files read by this compilation into the produced module file.

[**-fmodules-force-validate-user-headers**](#fmodules~6)

Force validation of user headers when repeatedly loading a module file within single build session

[**-fmodules-ignore-macro=**&lt;value&gt;](#fmodules~7)

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

[**-fmodules-prune-after=**&lt;seconds&gt;](#fmodules~8)

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

[**-fmodules-prune-interval=**&lt;seconds&gt;](#fmodules~9)

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

[**-fmodules-reduced-bmi**](#fmodules~10)

Generate the reduced BMI

[**-fmodules-search-all**](#fmodules~11)

Search even non-imported modules to resolve references

[**-fmodules-strict-decluse**](#fmodules~12)

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

[**-fmodules-user-build-path** &lt;directory&gt;](#fmodules~13)

Specify the module user build path

[**-fmodules-validate-input-files-content**](#fmodules~14)

Validate PCM input files based on content if mtime differs

[**-fmodules-validate-once-per-build-session**](#fmodules~15)

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**](#fmodules~16)

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

[**-fmodules**](#fmodules~17)

Enable the 'modules' language feature

[**-fms-anonymous-structs**](#fms)

Enable Microsoft anonymous struct/union extension.

[**-fms-compatibility-version=**&lt;value&gt;](#fms~2)

Dot-separated value representing the Microsoft compiler version number to report in \_MSC\_VER (0 = don't define it (default))

[**-fms-compatibility**](#fms~3)

Enable full Microsoft Visual C++ compatibility

[**-fms-define-stdc**](#fms~4)

Define '\_\_STDC\_\_' to '1' in MSVC Compatibility mode

[**-fms-extensions**](#fms~5)

Accept some non-standard constructs supported by the Microsoft compiler

[**-fms-hotpatch**](#fms~6)

Ensure that all functions can be hotpatched at runtime

[**-fms-runtime-lib=**&lt;value&gt;](#fms~7)

Select Windows run-time library

[**-fms-secure-hotpatch-functions-file=**&lt;value&gt;](#fms~8)

Path to a file that contains a list of mangled names of functions that should be hot-patched for Windows Secure Hot-Patching

[**-fms-secure-hotpatch-functions-list=**&lt;value&gt;](#fms~9)

List of mangled symbol names of functions that should be hot-patched for Windows Secure Hot-Patching

[**-fms-volatile**](#fms~10)

Volatile loads and stores have acquire and release semantics

[**-fmsc-version=**&lt;value&gt;](#fmsc)

Microsoft compiler version number to report in \_MSC\_VER (0 = don't define it (default))

[**-fnamed-loops**](#fnamed)

Enable support for named loops

**-fnew-alignment=**&lt;align&gt; Specifies the largest alignment guaranteed by '::operator new(size\_t)'

[**-fnew-infallible**](#fnew)

Enable treating throwing global C++ operator new as always returning valid memory (annotates with \_\_attribute\_\_((returns\_nonnull)) and throw()). This is detectable in source.

[**-fno-aapcs-bitfield-width**](#fno)

Do not follow the AAPCS standard requirement stating that volatile bit-field width is dictated by the field container type. (ARM only).

[**-fno-access-control**](#fno~2)

Disable C++ access control

[**-fno-addrsig**](#fno~3)

Don't emit an address-significance table

[**-fno-apinotes-modules**](#fno~4)

Disable module-based external API notes support

[**-fno-apinotes**](#fno~5)

Disable external API notes support

[**-fno-assume-sane-operator-new**](#fno~6)

Don't assume that C++'s global operator new can't alias any pointer

[**-fno-assume-unique-vtables**](#fno~7)

Disable optimizations based on vtable pointer identity

[**-fno-assumptions**](#fno~8)

Disable codegen and compile-time checks for C++23's \[\[assume]] attribute

[**-fno-atomic-fine-grained-memory**](#fno~9)

Assume no atomic operations on fine-grained memory

[**-fno-atomic-ignore-denormal-mode**](#fno~10)

Disallow atomic operations to ignore denormal mode

[**-fno-atomic-remote-memory**](#fno~11)

Assume no atomic operations on remote memory

[**-fno-auto-import**](#fno~12)

MinGW specific. Disable support for automatic dllimport in code generation and linking

[**-fno-autolink**](#fno~13)

Disable generation of linker directives for automatic library linking

[**-fno-builtin-**&lt;value&gt;](#fno~14)

Disable implicit builtin knowledge of a specific function

[**-fno-builtin**](#fno~15)

Disable implicit builtin knowledge of functions

[**-fno-c**++-static-destructors](#fno~16)

Disable C++ static destructor registration

[**-fno-char8\_t**](#fno~17)

Disable C++ builtin type char8\_t

[**-fno-clangir**](#fno~18)

Use the AST -&gt; LLVM pipeline to compile

[**-fno-color-diagnostics**](#fno~19)

Disable colors in diagnostics

[**-fno-common**](#fno~20)

Compile common globals like normal definitions

[**-fno-complete-member-pointers**](#fno~21)

Do not require member pointer base types to be complete if they would be significant under the Microsoft ABI

**-fno-constant-cfstrings** Disable creation of CodeFoundation-type constant strings

[**-fno-constant-nsarray-literals**](#fno~22)

Disable compile-time constant NSArray literals

[**-fno-constant-nsdictionary-literals**](#fno~23)

Disable compile-time constant NSDictionary literals

[**-fno-constant-nsnumber-literals**](#fno~24)

Disable compile-time constant NSNumber literals

[**-fno-coverage-mapping**](#fno~25)

Disable code coverage analysis

[**-fno-coverage-mcdc**](#fno~26)

Disable MC/DC coverage criteria

[**-fno-crash-diagnostics**](#fno~27)

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

[**-fno-cx-fortran-rules**](#fno~28)

Range reduction is disabled for complex arithmetic operations

[**-fno-cx-limited-range**](#fno~29)

Basic algebraic expansions of complex arithmetic operations involving are disabled.

[**-fno-cxx-modules**](#fno~30)

Disable modules for C++

[**-fno-debug-macro**](#fno~31)

Do not emit macro debug information

[**-fno-debug-record-sysroot**](#fno~32)

Scrub the sysroot path from debug info.

[**-fno-declspec**](#fno~33)

Disallow \_\_declspec as a keyword

[**-fno-define-target-os-macros**](#fno~34)

Disable predefined target OS macros

[**-fno-delayed-template-parsing**](#fno~35)

Disable delayed template parsing

[**-fno-delete-null-pointer-checks**](#fno~36)

Do not treat usage of null pointers as undefined behavior

[**-fno-diagnostics-fixit-info**](#fno~37)

Do not include fixit information in diagnostics

[**-fno-diagnostics-show-line-numbers**](#fno~38)

Show line numbers in diagnostic code snippets

[**-fno-digraphs**](#fno~39)

Disallow alternative token representations '&lt;:', ':&gt;', '&lt;%', '%&gt;', '%:', '%:%:'

[**-fno-direct-access-external-data**](#fno~40)

Use GOT indirection to reference external data symbols

[**-fno-disable-block-signature-string**](#fno~41)

Don't disable block signature string)

[**-fno-discard-value-names**](#fno~42)

Do not discard value names in LLVM IR

[**-fno-dollars-in-identifiers**](#fno~43)

Disallow '$' in identifiers

**-fno-elide-constructors** Disable C++ copy constructor elision

[**-fno-elide-type**](#fno~44)

Do not elide types when printing diagnostics

[**-fno-eliminate-unused-debug-types**](#fno~45)

Emit debug info for defined but unused types

[**-fno-enable-new-pm-codegen**](#fno~46)

Do not use the NewPM for the Codegen Pipeline

[**-fno-exceptions**](#fno~47)

Disable support for exception handling

[**-fno-experimental-allow-pointer-field-protection-attr**](#fno~48)

Disallow the use of the \[\[clang::pointer\_field\_protection]] attribute

[**-fno-experimental-late-parse-attributes**](#fno~49)

Disable experimental late parsing of attributes

[**-fno-experimental-loop-fusion**](#fno~50)

DisableEnable the loop fusion pass

[**-fno-experimental-pointer-field-protection-abi**](#fno~51)

Do not enable pointer field protection on all non-standard layout struct types

[**-fno-experimental-pointer-field-protection-tagged**](#fno~52)

Do not use pointer identity (tag) to discriminate pointers of non-trivially-copyable types

[**-fno-experimental-relative-c**++-abi-vtables](#fno~53)

Do not use the experimental C++ class ABI for classes with virtual tables

[**-fno-experimental-sanitize-metadata=**&lt;value&gt;](#fno~54)

Disable emitting metadata for binary analysis sanitizers

[**-fno-fat-lto-objects**](#fno~55)

Disable fat LTO object support

[**-fno-file-reproducible**](#fno~56)

Use the host's platform-specific path separator character when expanding the \_\_FILE\__ macro

[**-fno-fine-grained-bitfield-accesses**](#fno~57)

Use large-integer access for consecutive bitfield runs.

[**-fno-finite-loops**](#fno~58)

Do not assume that any loop is finite.

[**-fno-fixed-point**](#fno~59)

Disable fixed point types

[**-fno-force-enable-int128**](#fno~60)

Disable support for int128\_t type

[**-fno-global-isel**](#fno~61)

Disables the global instruction selector

[**-fno-gnu-inline-asm**](#fno~62)

Disable GNU style inline asm

[**-fno-gpu-allow-device-init**](#fno~63)

Don't allow device side init function in HIP (experimental)

[**-fno-gpu-approx-transcendentals**](#fno~64)

Don't use approximate transcendental functions

[**-fno-gpu-defer-diag**](#fno~65)

Don't defer host/device related diagnostic messages for CUDA/HIP

[**-fno-gpu-exclude-wrong-side-overloads**](#fno~66)

Exclude wrong side overloads only if there are same side overloads in overloading resolution for CUDA/HIP

[**-fno-hip-emit-relocatable**](#fno~67)

Do not override toolchain to compile HIP source to relocatable

[**-fno-hip-fp32-correctly-rounded-divide-sqrt**](#fno~68)

Don't specify that single precision floating-point divide and sqrt used in the program source are correctly rounded (HIP device compilation only)

[**-fno-hip-kernel-arg-name**](#fno~69)

Don't specify that kernel argument names are preserved (HIP only)

**-fno-hip-new-launch-api** Don't use new kernel launching API for HIP

[**-fno-integrated-as**](#fno~70)

Disable the integrated assembler

[**-fno-integrated-cc1**](#fno~71)

Spawn a separate process for each cc1

[**-fno-integrated-objemitter**](#fno~72)

Use external machine object code emitter.

[**-fno-jump-tables**](#fno~73)

Do not use jump tables for lowering switches

[**-fno-keep-persistent-storage-variables**](#fno~74)

Disable keeping all variables that have a persistent storage duration, including global, static and thread-local variables, to guarantee that they can be directly addressed

**-fno-keep-static-consts** Don't keep static const variables even if unused

[**-fno-knr-functions**](#fno~75)

Disable support for K&R C function declarations

[**-fno-lifetime-dse**](#fno~76)

Do not aggressively delete stores after the end of an object's lifetime

[**-fno-loop-interchange**](#fno~77)

Disable the loop interchange pass

[**-fno-lto**](#fno~78)

Disable LTO mode (default)

[**-fno-memory-profile**](#fno~79)

Disable heap memory profiling

[**-fno-merge-all-constants**](#fno~80)

Disallow merging of constants

[**-fno-modules-driver**](#fno~81)

Disable support for driver managed module builds (experimental)

[**-fno-modules-force-validate-user-headers**](#fno~82)

Do not force validation of user headers when repeatedly loading a module file within single build session

[**-fno-modules-validate-textual-header-includes**](#fno~83)

Do not enforce **-fmodules-decluse** and private header restrictions for textual headers. This flag will be removed in a future Clang release.

[**-fno-ms-anonymous-structs**](#fno~84)

Disable Microsoft anonymous struct/union extension.

[**-fno-ms-tls-guards**](#fno~85)

Do not emit code to perform on-demand initialization of thread-local variables

[**-fno-new-infallible**](#fno~86)

Disable treating throwing global C++ operator new as always returning valid memory (annotates with \_\_attribute\_\_((returns\_nonnull)) and throw()). This is detectable in source.

[**-fno-objc-avoid-heapify-local-blocks**](#fno~87)

Don't try to avoid heapifying local blocks

[**-fno-objc-constant-literals**](#fno~88)

Disable compile-time constant ObjC literals

[**-fno-objc-infer-related-result-type**](#fno~89)

do not infer Objective-C related result type based on method family

[**-fno-offload-lto**](#fno~90)

Disable LTO mode (default) for offload compilation

[**-fno-offload-uniform-block**](#fno~91)

Don't assume that kernels are launched with uniform block sizes (default true for CUDA/HIP and false otherwise)

[**-fno-offload-via-llvm**](#fno~92)

Don't use LLVM/Offload as portable offloading runtime.

[**-fno-openmp-assume-no-nested-parallelism**](#fno~93)

Assert that a nested parallel region may be used in the GPU

[**-fno-openmp-assume-no-thread-state**](#fno~94)

Assert that a thread in a parallel region may modify an ICV

[**-fno-openmp-assume-teams-oversubscription**](#fno~95)

Do not assume teams oversubscription.

[**-fno-openmp-assume-threads-oversubscription**](#fno~96)

Do not assume threads oversubscription.

[**-fno-openmp-extensions**](#fno~97)

Disable all Clang extensions for OpenMP directives and clauses

[**-fno-openmp-new-driver**](#fno~98)

Don't use the new driver for OpenMP offloading.

[**-fno-openmp-simd**](#fno~99)

Do not emit code for any OpenMP constructs.

[**-fno-operator-names**](#fno~100)

Do not treat C++ operator name keywords as synonyms for operators

[**-fno-optimize-sibling-calls**](#fno~101)

Disable tail call optimization, keeping the call stack accurate

[**-fno-partition-static-data-sections**](#fno~102)

Disable partition static data sections using PGO profile information and MemProf (x86 and aarch64 ELF). See LLVM MemProf doc for usage.

[**-fno-pch-codegen**](#fno~103)

Do not generate code for uses of this PCH that assumes an explicit object file will be built for the PCH

[**-fno-pch-debuginfo**](#fno~104)

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

[**-fno-plt**](#fno~105)

Use GOT indirection instead of PLT to make external function calls (x86 only)

[**-fno-preserve-as-comments**](#fno~106)

Do not preserve comments in inline assembly

[**-fno-profile-generate**](#fno~107)

Disable generation of profile instrumentation.

[**-fno-profile-instr-generate**](#fno~108)

Disable generation of profile instrumentation.

[**-fno-profile-instr-use**](#fno~109)

Disable using instrumentation data for profile-guided optimization

[**-fno-pseudo-probe-for-profiling**](#fno~110)

Do not emit pseudo probes for sample profiling

[**-fno-raw-string-literals**](#fno~111)

Disable raw string literals

[**-fno-register-global-dtors-with-atexit**](#fno~112)

Don't use atexit or \_\_cxa\_atexit to register global destructors

[**-fno-rtlib-add-rpath**](#fno~113)

Do not add **-rpath** with architecture-specific resource directory to the linker flags. When **--hip-link** is specified, do not add **-rpath** with HIP runtime library directory to the linker flags

[**-fno-rtlib-defaultlib**](#fno~114)

On Windows, do not emit /defaultlib: directives to link compiler-rt libraries

[**-fno-rtti-data**](#fno~115)

Disable generation of RTTI data

[**-fno-rtti**](#fno~116)

Disable generation of rtti information

[**-fno-sanitize-address-globals-dead-stripping**](#fno~117)

Disable linker dead stripping of globals in AddressSanitizer

[**-fno-sanitize-address-outline-instrumentation**](#fno~118)

Use default code inlining logic for the address sanitizer

[**-fno-sanitize-address-poison-custom-array-cookie**](#fno~119)

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

[**-fno-sanitize-address-use-after-scope**](#fno~120)

Disable use-after-scope detection in AddressSanitizer

[**-fno-sanitize-address-use-odr-indicator**](#fno~121)

Disable ODR indicator globals

[**-fno-sanitize-alloc-token-extended**](#fno~122)

Disable extended coverage to custom allocation functions

[**-fno-sanitize-alloc-token-fast-abi**](#fno~123)

Use the default AllocToken ABI

[**-fno-sanitize-annotate-debug-info=**&lt;value&gt;](#fno~124)

Do not allow compiler to annotate sanitizer instrumentation with extra debug info for the specified sanitizers

[**-fno-sanitize-annotate-debug-info**](#fno~125)

Do not allow compiler to annotate sanitizer instrumentation with extra debug info for any sanitizers

[**-fno-sanitize-cfi-canonical-jump-tables**](#fno~126)

Do not make the jump table addresses canonical in the symbol table

[**-fno-sanitize-cfi-cross-dso**](#fno~127)

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

[**-fno-sanitize-coverage=**&lt;value&gt;](#fno~128)

Disable features of coverage instrumentation for Sanitizers

[**-fno-sanitize-debug-trap-reasons**](#fno~129)

Alias for **-fsanitize-debug-trap-reasons**=*none*

[**-fno-sanitize-handler-preserve-all-regs**](#fno~130)

Disable handlers with preserve\_all calling convention

[**-fno-sanitize-hwaddress-experimental-aliasing**](#fno~131)

Disable aliasing mode in HWAddressSanitizer

[**-fno-sanitize-ignorelist**](#fno~132)

Don't use ignorelist file for sanitizers

[**-fno-sanitize-memory-param-retval**](#fno~133)

Disable detection of uninitialized parameters and return values

[**-fno-sanitize-memory-track-origins**](#fno~134)

Disable origins tracking in MemorySanitizer

[**-fno-sanitize-memory-use-after-dtor**](#fno~135)

Disable use-after-destroy detection in MemorySanitizer

[**-fno-sanitize-merge=**&lt;value&gt;](#fno~136)

Do not allow compiler to merge handlers for specified sanitizers

[**-fno-sanitize-merge**](#fno~137)

Do not allow compiler to merge handlers for any sanitizers

[**-fno-sanitize-recover=**&lt;value&gt;](#fno~138)

Disable recovery for specified sanitizers

[**-fno-sanitize-stable-abi**](#fno~139)

Conventional ABI instrumentation for sanitizer runtime. Default: Conventional

[**-fno-sanitize-stats**](#fno~140)

Disable sanitizer statistics gathering.

[**-fno-sanitize-thread-atomics**](#fno~141)

Disable atomic operations instrumentation in ThreadSanitizer

[**-fno-sanitize-thread-func-entry-exit**](#fno~142)

Disable function entry/exit instrumentation in ThreadSanitizer

[**-fno-sanitize-thread-memory-access**](#fno~143)

Disable memory access instrumentation in ThreadSanitizer

**-fno-sanitize-trap-loop** Do not use infinite loops in trapping mode

[**-fno-sanitize-trap=**&lt;value&gt;](#fno~144)

Disable trapping for specified sanitizers

[**-fno-sanitize-trap**](#fno~145)

Disable trapping for all sanitizers

[**-fno-sanitize-type-outline-instrumentation**](#fno~146)

Use code inlining logic for the type sanitizer

[**-fno-short-wchar**](#fno~147)

Force wchar\_t to be an unsigned int

[**-fno-show-column**](#fno~148)

Do not include column number on diagnostics

[**-fno-show-source-location**](#fno~149)

Do not include source location information with diagnostics

[**-fno-signed-char**](#fno~150)

char is unsigned

[**-fno-signed-zeros**](#fno~151)

Allow optimizations that ignore the sign of floating point zeros

[**-fno-skip-odr-check-in-gmf**](#fno~152)

Perform ODR checks for decls in the global module fragment.

[**-fno-spell-checking**](#fno~153)

Disable spell-checking

[**-fno-split-machine-functions**](#fno~154)

Disable late function splitting using profile information (x86 and aarch64 ELF)

[**-fno-split-stack**](#fno~155)

Wouldn't use segmented stack

[**-fno-stack-clash-protection**](#fno~156)

Disable stack clash protection

[**-fno-stack-protector**](#fno~157)

Disable the use of stack protectors

[**-fno-standalone-debug**](#fno~158)

Limit debug information produced to reduce size of debug binary

[**-fno-strict-aliasing**](#fno~159)

Disable optimizations based on strict aliasing rules

[**-fno-strict-bool=**&lt;value&gt;](#fno~160)

Disable optimizations based on bool bit patterns never being anything other than 0 or 1, specifying a conversion behavior.

[**-fno-strict-bool**](#fno~161)

Disable optimizations based on bool bit patterns never being anything other than 0 or 1

[**-fno-strict-float-cast-overflow**](#fno~162)

Relax language rules and try to match the behavior of the target's native float-to-int conversion instructions

[**-fno-strict-return**](#fno~163)

Don't treat control flow paths that fall off the end of a non-void function as unreachable

[**-fno-swift-version-independent-apinotes**](#fno~164)

Disable version-independent external API notes support

[**-fno-sycl**](#fno~165)

Disable SYCL C++ extensions

[**-fno-temp-file**](#fno~166)

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

**-fno-threadsafe-statics** Do not emit code to make initialization of local statics thread safe

[**-fno-trigraphs**](#fno~167)

Do not process trigraph sequences

[**-fno-unified-lto**](#fno~168)

Use distinct LTO pipelines

[**-fno-unique-section-names**](#fno~169)

Don't use unique names for text and data sections

[**-fno-unroll-loops**](#fno~170)

Turn off loop unroller

[**-fno-use-cxa-atexit**](#fno~171)

Don't use \_\_cxa\_atexit for calling destructors

[**-fno-use-init-array**](#fno~172)

Use .ctors/.dtors instead of .init\_array/.fini\_array

[**-fno-verify-intermediate-code**](#fno~173)

Disable verification of LLVM IR

[**-fno-visibility-inlines-hidden-static-local-var**](#fno~174)

Disables **-fvisibility-inlines-hidden-static-local-var** (this is the default on non-darwin targets)

[**-fno-xray-function-index**](#fno~175)

Omit function index section at the expense of single-function patching performance

[**-fno-zero-initialized-in-bss**](#fno~176)

Don't place zero initialized data in BSS

[**-fno-zos-extensions**](#fno~177)

Do not accept non-standard constructs supported by the z/OS compiler

[**-fobjc-arc-exceptions**](#fobjc)

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

[**-fobjc-arc**](#fobjc~2)

Synthesize retain and release calls for Objective-C pointers

[**-fobjc-avoid-heapify-local-blocks**](#fobjc~3)

Try to avoid heapifying local blocks

[**-fobjc-constant-literals**](#fobjc~4)

Enable compile-time constant ObjC literals

[**-fobjc-direct-precondition-thunk**](#fobjc~5)

Move precondition checks for direct methods into thunks

[**-fobjc-disable-direct-methods-for-testing**](#fobjc~6)

Ignore attribute objc\_direct so that direct methods can be tested

[**-fobjc-encode-cxx-class-template-spec**](#fobjc~7)

Fully encode c++ class template specialization

[**-fobjc-exceptions**](#fobjc~8)

Enable Objective-C exceptions

[**-fobjc-runtime=**&lt;value&gt;](#fobjc~9)

Specify the target Objective-C runtime kind and version

[**-fobjc-weak**](#fobjc~10)

Enable ARC-style weak references in Objective-C

[**-foffload-implicit-host-device-templates**](#foffload)

Template functions or specializations without host, device and global attributes have implicit host device attributes (CUDA/HIP only)

[**-foffload-lto=**&lt;value&gt;](#foffload~2)

Set LTO mode for offload compilation

[**-foffload-lto**](#foffload~3)

Enable LTO in 'full' mode for offload compilation

**-foffload-uniform-block** Assume that kernels are launched with uniform block sizes (default true for CUDA/HIP and false otherwise)

[**-foffload-via-llvm**](#foffload~4)

Use LLVM/Offload as portable offloading runtime.

[**-fomit-frame-pointer**](#fomit)

Omit the frame pointer from functions that don't need it. Some stack unwinding cases, such as profilers and sanitizers, may prefer specifying **-fno-omit-frame-pointer**. On many targets, **-O1** and higher omit the frame pointer by default. **-m\[no-]omit-leaf-frame-pointer** takes precedence for leaf functions

[**-fopenacc**](#fopenacc)

Enable OpenACC

[**-fopenmp-assume-no-nested-parallelism**](#fopenmp)

Assert no nested parallel regions in the GPU

[**-fopenmp-assume-no-thread-state**](#fopenmp~2)

Assert no thread in a parallel region modifies an ICV

[**-fopenmp-assume-teams-oversubscription**](#fopenmp~3)

Allow the optimizer to discretely increase the number of teams. May cause ignore environment variables that set the number of teams to be ignored. The combination of **-fopenmp-assume-teams-oversubscription** and **-fopenmp-assume-threads-oversubscription** may allow the conversion of loops into sequential code by ensuring that each team/thread executes at most one iteration.

[**-fopenmp-assume-threads-oversubscription**](#fopenmp~4)

Allow the optimizer to discretely increase the number of threads. May cause ignore environment variables that set the number of threads to be ignored. The combination of **-fopenmp-assume-teams-oversubscription** and **-fopenmp-assume-threads-oversubscription** may allow the conversion of loops into sequential code by ensuring that each team/thread executes at most one iteration.

[**-fopenmp-cuda-teams-reduction-recs-num=**&lt;value&gt;](#fopenmp~5)

Deprecated and ignored. The teams reduction buffer is sized automatically at kernel launch to match the actual number of teams; this flag is accepted for backwards compatibility only and emits a deprecation warning when used.

[**-fopenmp-enable-irbuilder**](#fopenmp~6)

Use the experimental OpenMP-IR-Builder codegen path.

[**-fopenmp-extensions**](#fopenmp~7)

Enable all Clang extensions for OpenMP directives and clauses

[**-fopenmp-force-usm**](#fopenmp~8)

Force behavior as if the user specified pragma omp requires unified\_shared\_memory.

[**-fopenmp-new-driver**](#fopenmp~9)

Use the new driver for OpenMP offloading.

[**-fopenmp-offload-mandatory**](#fopenmp~10)

Do not create a host fallback if offloading to the device fails.

[**-fopenmp-simd**](#fopenmp~11)

Emit OpenMP code only for SIMD-based constructs.

[**-fopenmp-target-debug**](#fopenmp~12)

Enable debugging in the OpenMP offloading device RTL

[**-fopenmp-target-fast**](#fopenmp~13)

Assert common GPU usage patterns to enable OpenMP runtime optimizations

[**-fopenmp-target-jit**](#fopenmp~14)

Emit code that can be JIT compiled for OpenMP offloading. Implies **-foffload-lto**=*full*

[**-fopenmp-targets=**&lt;value&gt;](#fopenmp~15)

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

[**-fopenmp-version=**&lt;value&gt;](#fopenmp~16)

Set OpenMP version (e.g. 45 for OpenMP 4.5, 51 for OpenMP 5.1). Default value is 51 for Clang

[**-fopenmp**](#fopenmp~17)

Parse OpenMP pragmas and generate parallel code.

[**-foperator-arrow-depth=**&lt;value&gt;](#foperator)

Maximum number of 'operator-&gt;'s to call for a member access

[**-foptimization-record-file=**&lt;file&gt;](#foptimization)

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** &lt;arch&gt; options.

[**-foptimization-record-passes=**&lt;regex&gt;](#foptimization~2)

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

[**-fpack-struct=**&lt;value&gt;](#fpack)

Specify the default maximum struct packing alignment

[**-fpartition-static-data-sections**](#fpartition)

Enable partition static data sections using PGO profile information and MemProf (x86 and aarch64 ELF). See LLVM MemProf doc for usage.

[**-fpascal-strings**](#fpascal)

Recognize and construct Pascal-style string literals

**-fpass-plugin=**&lt;dsopath&gt; Load pass plugin from a dynamic shared object file (only with new pass manager).

[**-fpatchable-function-entry=**&lt;N,M,Section&gt;](#fpatchable)

Generate M NOPs before function entry and N-M NOPs after function entry. If section is specified, use it instead of \_\_patchable\_function\_entries.

[**-fpcc-struct-return**](#fpcc)

Override the default ABI to return all structs on the stack

[**-fpch-codegen**](#fpch)

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

[**-fpch-debuginfo**](#fpch~2)

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

[**-fpch-instantiate-templates**](#fpch~3)

Instantiate templates already while building a PCH

[**-fpch-validate-input-files-content**](#fpch~4)

Validate PCH input files based on content if mtime differs

[**-fplugin-arg-**&lt;name&gt;-&lt;arg&gt;](#fplugin)

Pass &lt;arg&gt; to plugin &lt;name&gt;

[**-fplugin=**&lt;dsopath&gt;](#fplugin=)

Load the named plugin (dynamic shared object)

[**-fprebuilt-implicit-modules**](#fprebuilt)

Look up implicit modules in the prebuilt module path

[**-fprebuilt-module-path=**&lt;directory&gt;](#fprebuilt~2)

Specify the prebuilt module path

[**-fpreferred-function-alignment=**&lt;value&gt;](#fpreferred)

preferred alignment for functions

[**-fproc-stat-report=**&lt;value&gt;](#fproc)

Save subprocess statistics to the given file

[**-fproc-stat-report**&lt;value&gt;](#fproc~2)

Print subprocess statistics

[**-fprofile-arcs**](#fprofile)

Instrument code to produce gcov data files (\*.gcda)

[**-fprofile-continuous**](#fprofile~2)

Enable continuous instrumentation profiling mode

[**-fprofile-exclude-files=**&lt;value&gt;](#fprofile~3)

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

[**-fprofile-filter-files=**&lt;value&gt;](#fprofile~4)

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

[**-fprofile-function-groups=**&lt;N&gt;](#fprofile~5)

Partition functions into N groups and select only functions in group i to be instrumented using **-fprofile-selected-function-group**

[**-fprofile-generate-cold-function-coverage=**&lt;directory&gt;](#fprofile~6)

Generate instrumented code to collect coverage info for cold functions into &lt;directory&gt;/default.profraw (overridden by LLVM\_PROFILE\_FILE env var)

[**-fprofile-generate-cold-function-coverage**](#fprofile~7)

Generate instrumented code to collect coverage info for cold functions into default.profraw file (overridden by '=' form of option or LLVM\_PROFILE\_FILE env var)

[**-fprofile-generate=**&lt;directory&gt;](#fprofile~8)

Generate instrumented code to collect execution counts into &lt;directory&gt;/default.profraw (overridden by LLVM\_PROFILE\_FILE env var)

[**-fprofile-generate**](#fprofile~9)

Generate instrumented code to collect execution counts into default.profraw (overridden by LLVM\_PROFILE\_FILE env var)

[**-fprofile-instr-generate=**&lt;file&gt;](#fprofile~10)

Generate instrumented code to collect execution counts into &lt;file&gt; (overridden by LLVM\_PROFILE\_FILE env var)

[**-fprofile-instr-generate**](#fprofile~11)

Generate instrumented code to collect execution counts into default.profraw file (overridden by '=' form of option or LLVM\_PROFILE\_FILE env var)

[**-fprofile-instr-use=**&lt;value&gt;](#fprofile~12)

Use instrumentation data for profile-guided optimization

[**-fprofile-list=**&lt;value&gt;](#fprofile~13)

Filename defining the list of functions/files to instrument. The file uses the sanitizer special case list format.

[**-fprofile-remapping-file=**&lt;file&gt;](#fprofile~14)

Use the remappings described in &lt;file&gt; to match the profile data against names in the program

[**-fprofile-sample-accurate**](#fprofile~15)

Specifies that the sample profile is accurate

[**-fprofile-sample-use=**&lt;value&gt;](#fprofile~16)

Enable sample-based profile guided optimizations

[**-fprofile-selected-function-group=**&lt;i&gt;](#fprofile~17)

Partition functions into N groups using **-fprofile-function-groups** and select only functions in group i to be instrumented. The valid range is 0 to N-1 inclusive

[**-fprofile-update=**&lt;method&gt;](#fprofile~18)

Set update method of profile counters

[**-fprofile-use=**&lt;pathname&gt;](#fprofile~19)

Use instrumentation data for profile-guided optimization. If pathname is a directory, it reads from &lt;pathname&gt;/default.profdata. Otherwise, it reads from file &lt;pathname&gt;.

[**-fprotect-parens**](#fprotect)

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

[**-fpseudo-probe-for-profiling**](#fpseudo)

Emit pseudo probes for sample profiling

[**-fptrauth-auth-traps**](#fptrauth)

Enable traps on authentication failures

[**-fptrauth-block-descriptor-pointers**](#fptrauth~2)

Enable signing and authentication of block descriptors

[**-fptrauth-calls**](#fptrauth~3)

Enable signing and authentication of all indirect calls

[**-fptrauth-elf-got**](#fptrauth~4)

Enable authentication of pointers from GOT (ELF only)

[**-fptrauth-function-pointer-type-discrimination**](#fptrauth~5)

Enable type discrimination on C function pointers

[**-fptrauth-indirect-gotos**](#fptrauth~6)

Enable signing and authentication of indirect goto targets

[**-fptrauth-init-fini-address-discrimination**](#fptrauth~7)

Enable address discrimination of function pointers in init/fini arrays

[**-fptrauth-init-fini**](#fptrauth~8)

Enable signing of function pointers in init/fini arrays

[**-fptrauth-intrinsics**](#fptrauth~9)

Enable pointer authentication intrinsics

**-fptrauth-objc-class-ro** Enable signing and authentication for ObjC class\_ro pointers

[**-fptrauth-objc-interface-sel**](#fptrauth~10)

Enable signing and authentication of Objective-C object's 'SEL' fields

[**-fptrauth-objc-isa**](#fptrauth~11)

Enable signing and authentication of Objective-C object's 'isa' field

[**-fptrauth-returns**](#fptrauth~12)

Enable signing and authentication of return addresses

[**-fptrauth-type-info-vtable-pointer-discrimination**](#fptrauth~13)

Enable type and address discrimination of vtable pointer of std::type\_info

[**-fptrauth-vtable-pointer-address-discrimination**](#fptrauth~14)

Enable address discrimination of vtable pointers

[**-fptrauth-vtable-pointer-type-discrimination**](#fptrauth~15)

Enable type discrimination of vtable pointers

[**-frandomize-layout-seed-file=**&lt;file&gt;](#frandomize)

File holding the seed used by the randomize structure layout feature

[**-frandomize-layout-seed=**&lt;seed&gt;](#frandomize~2)

The seed used by the randomize structure layout feature

[**-fraw-string-literals**](#fraw)

Enable raw string literals

[**-freciprocal-math**](#freciprocal)

Allow division operations to be reassociated

[**-freflection**](#freflection)

Enable C++26 reflection

[**-freg-struct-return**](#freg)

Override the default ABI to return small structs in registers

[**-fregister-global-dtors-with-atexit**](#fregister)

Use atexit or \_\_cxa\_atexit to register global destructors

[**-fropi**](#fropi)

Generate read-only position independent code (ARM only)

[**-frtlib-add-rpath**](#frtlib)

Add **-rpath** with architecture-specific resource directory to the linker flags. When **--hip-link** is specified, also add **-rpath** with HIP runtime library directory to the linker flags

[**-frtlib-defaultlib**](#frtlib~2)

On Windows, emit /defaultlib: directives to link compiler-rt libraries (default)

[**-frwpi**](#frwpi)

Generate read-write position independent code (ARM only)

[**-fsafe-buffer-usage-suggestions**](#fsafe)

Display suggestions to update code associated with **-Wunsafe-buffer-usage** warnings

[**-fsample-profile-use-profi**](#fsample)

Use profi to infer block and edge counts

[**-fsanitize-address-destructor=**&lt;value&gt;](#fsanitize)

Set the kind of module destructors emitted by AddressSanitizer instrumentation. These destructors are emitted to unregister instrumented global variables when code is unloaded (e.g. via \`dlclose()\`).

[**-fsanitize-address-field-padding=**&lt;value&gt;](#fsanitize~2)

Level of field padding for AddressSanitizer

[**-fsanitize-address-globals-dead-stripping**](#fsanitize~3)

Enable linker dead stripping of globals in AddressSanitizer

[**-fsanitize-address-outline-instrumentation**](#fsanitize~4)

Always generate function calls for address sanitizer instrumentation

[**-fsanitize-address-poison-custom-array-cookie**](#fsanitize~5)

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

[**-fsanitize-address-use-after-return=**&lt;mode&gt;](#fsanitize~6)

Select the mode of detecting stack use-after-return in AddressSanitizer

[**-fsanitize-address-use-after-scope**](#fsanitize~7)

Enable use-after-scope detection in AddressSanitizer

[**-fsanitize-address-use-odr-indicator**](#fsanitize~8)

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

[**-fsanitize-alloc-token-extended**](#fsanitize~9)

Enable extended coverage to custom allocation functions

[**-fsanitize-alloc-token-fast-abi**](#fsanitize~10)

Use the AllocToken fast ABI

[**-fsanitize-annotate-debug-info=**&lt;value&gt;](#fsanitize~11)

Annotate sanitizer instrumentation with extra debug info for the specified sanitizers, if supported

[**-fsanitize-annotate-debug-info**](#fsanitize~12)

Allow compiler to annotate sanitizer instrumentation with extra debug info for all sanitizers, where supported

[**-fsanitize-blacklist=**&lt;value&gt;](#fsanitize~13)

Alias for **-fsanitize-ignorelist=**

[**-fsanitize-cfi-canonical-jump-tables**](#fsanitize~14)

Make the jump table addresses canonical in the symbol table

[**-fsanitize-cfi-cross-dso**](#fsanitize~15)

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

[**-fsanitize-cfi-icall-experimental-normalize-integers**](#fsanitize~16)

Normalize integers in CFI indirect call type signature checks

[**-fsanitize-cfi-icall-generalize-pointers**](#fsanitize~17)

Generalize pointers in CFI indirect call type signature checks

[**-fsanitize-coverage-allowlist=**&lt;value&gt;](#fsanitize~18)

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

[**-fsanitize-coverage-ignorelist=**&lt;value&gt;](#fsanitize~19)

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

[**-fsanitize-coverage-stack-depth-callback-min=**&lt;M&gt;](#fsanitize~20)

Use callback for max stack depth tracing with minimum stack depth M

[**-fsanitize-coverage=**&lt;value&gt;](#fsanitize~21)

Specify the type of coverage instrumentation for Sanitizers

[**-fsanitize-debug-trap-reasons=**&lt;value&gt;](#fsanitize~22)

Set how trap reasons are emitted. \`none\` - Not emitted. This gives the smallest debug info; \`basic\` - Emit a fixed trap message per check type. This increases the debug info size but not as much as \`detailed\`; \`detailed\` - Emit a more detailed trap message. This increases the debug info size the most. Default is \`detailed\`.

[**-fsanitize-debug-trap-reasons**](#fsanitize~23)

Alias for **-fsanitize-debug-trap-reasons**=*detailed*

[**-fsanitize-handler-preserve-all-regs**](#fsanitize~24)

Enable handlers with preserve\_all calling convention

[**-fsanitize-hwaddress-abi=**&lt;value&gt;](#fsanitize~25)

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

[**-fsanitize-hwaddress-experimental-aliasing**](#fsanitize~26)

Enable aliasing mode in HWAddressSanitizer

[**-fsanitize-ignore-for-ubsan-feature=**&lt;check&gt;](#fsanitize~27)

Prevents \`\_\_has\_feature(undefined\_behavior\_sanitizer)\` from evaluating true for these UBSan checks. See user manual for available UBSan checks

[**-fsanitize-ignorelist=**&lt;value&gt;](#fsanitize~28)

Path to ignorelist file for sanitizers

[**-fsanitize-kcfi-arity**](#fsanitize~29)

Embed function arity information into the KCFI patchable function prefix

[**-fsanitize-kcfi-hash=**&lt;value&gt;](#fsanitize~30)

Select hash algorithm for KCFI type IDs (xxHash64, FNV-1a)

[**-fsanitize-memory-param-retval**](#fsanitize~31)

Enable detection of uninitialized parameters and return values

[**-fsanitize-memory-track-origins=**&lt;value&gt;](#fsanitize~32)

Enable origins tracking in MemorySanitizer

[**-fsanitize-memory-track-origins**](#fsanitize~33)

Enable origins tracking in MemorySanitizer

[**-fsanitize-memory-use-after-dtor**](#fsanitize~34)

Enable use-after-destroy detection in MemorySanitizer

[**-fsanitize-memtag-mode=**&lt;value&gt;](#fsanitize~35)

Set default MTE mode to 'sync' (default) or 'async'

[**-fsanitize-merge=**&lt;value&gt;](#fsanitize~36)

Allow compiler to merge handlers for specified sanitizers

[**-fsanitize-merge**](#fsanitize~37)

Allow compiler to merge handlers for all sanitizers

[**-fsanitize-recover=**&lt;value&gt;](#fsanitize~38)

Enable recovery for specified sanitizers

[**-fsanitize-skip-hot-cutoff=**&lt;value&gt;](#fsanitize~39)

Exclude sanitization for the top hottest code responsible for the given fraction of PGO counters (0.0 \[default] = skip none; 1.0 = skip all). Argument format: &lt;sanitizer1&gt;=&lt;value1&gt;,&lt;sanitizer2&gt;=&lt;value2&gt;,...

[**-fsanitize-stable-abi**](#fsanitize~40)

Stable ABI instrumentation for sanitizer runtime. Default: Conventional

[**-fsanitize-stats**](#fsanitize~41)

Enable sanitizer statistics gathering.

[**-fsanitize-system-ignorelist=**&lt;value&gt;](#fsanitize~42)

Path to system ignorelist file for sanitizers

[**-fsanitize-thread-atomics**](#fsanitize~43)

Enable atomic operations instrumentation in ThreadSanitizer (default)

[**-fsanitize-thread-func-entry-exit**](#fsanitize~44)

Enable function entry/exit instrumentation in ThreadSanitizer (default)

[**-fsanitize-thread-memory-access**](#fsanitize~45)

Enable memory access instrumentation in ThreadSanitizer (default)

[**-fsanitize-trap-loop**](#fsanitize~46)

Use infinite loops in trapping mode

**-fsanitize-trap=**&lt;value&gt; Enable trapping for specified sanitizers

[**-fsanitize-trap**](#fsanitize~47)

Enable trapping for all sanitizers

[**-fsanitize-type-outline-instrumentation**](#fsanitize~48)

Always generate function calls for type sanitizer instrumentation

[**-fsanitize-undefined-ignore-overflow-pattern=**&lt;value&gt;](#fsanitize~49)

Specify the overflow patterns to exclude from arithmetic sanitizer instrumentation

[**-fsanitize-undefined-strip-path-components=**&lt;number&gt;](#fsanitize~50)

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

[**-fsanitize=**&lt;check&gt;](#fsanitize=)

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

[**-fsave-optimization-record=**&lt;format&gt;](#fsave)

Generate an optimization record file in a specific format

[**-fsave-optimization-record**](#fsave~2)

Generate a YAML optimization record file

[**-fseh-exceptions**](#fseh)

Use SEH style exceptions

[**-fseparate-named-sections**](#fseparate)

Use separate unique sections for named sections (ELF Only)

[**-fshort-enums**](#fshort)

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

[**-fshort-wchar**](#fshort~2)

Force wchar\_t to be a short unsigned int

[**-fshow-overloads=**&lt;value&gt;](#fshow)

Which overload candidates to show when overload resolution fails. Defaults to 'all'

**-fshow-skipped-includes** Show skipped includes in **-H** output.

[**-fsigned-char**](#fsigned)

char is signed

[**-fsized-deallocation**](#fsized)

Enable C++14 sized global deallocation functions

[**-fsjlj-exceptions**](#fsjlj)

Use SjLj style exceptions

**-fskip-odr-check-in-gmf** Skip ODR checks for decls in the global module fragment.

[**-fslp-vectorize**](#fslp)

Enable the superword-level parallelism vectorization passes

[**-fspell-checking-limit=**&lt;value&gt;](#fspell)

Set the maximum number of times to perform spell checking on unrecognized identifiers (0 = no limit)

[**-fsplit-dwarf-inlining**](#fsplit)

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**](#fsplit~2)

Enables splitting of the LTO unit

[**-fsplit-machine-functions**](#fsplit~3)

Enable late function splitting using profile information (x86 and aarch64 ELF)

[**-fsplit-stack**](#fsplit~4)

Use segmented stack

[**-fstack-clash-protection**](#fstack)

Enable stack clash protection

[**-fstack-protector-all**](#fstack~2)

Enable stack protectors for all functions

[**-fstack-protector-strong**](#fstack~3)

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-protector**](#fstack~4)

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-size-section**](#fstack~5)

Emit section containing metadata on function stack sizes

[**-fstack-usage**](#fstack~6)

Emit .su file containing information on function stack sizes

[**-fstandalone-debug**](#fstandalone)

Emit full debug info for all types used by the program

[**-fstrict-aliasing**](#fstrict)

Enable optimizations based on strict aliasing rules

[**-fstrict-bool**](#fstrict~2)

Enable optimizations based on bool bit patterns never being anything other than 0 or 1

[**-fstrict-enums**](#fstrict~3)

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

[**-fstrict-flex-arrays=**&lt;n&gt;](#fstrict~4)

Enable optimizations based on the strict definition of flexible arrays.

[**-fstrict-float-cast-overflow**](#fstrict~5)

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

[**-fstrict-vtable-pointers**](#fstrict~6)

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

[**-fswift-async-fp=**&lt;option&gt;](#fswift)

Control emission of Swift async extended frame info

[**-fswift-version-independent-apinotes**](#fswift~2)

Enable version-independent external API notes support

[**-fsycl-device-image-split=**&lt;value&gt;](#fsycl)

Select the granularity at which SYCL device code is grouped into device images: kernel (one device image per kernel) | translation\_unit (one device image per translation unit) | link\_unit (one device image per linking unit). Default is 'translation\_unit'.

[**-fsycl-device-image-split**](#fsycl~2)

Group SYCL device code into device images: one per 'translation\_unit'.

[**-fsycl-device-only**](#fsycl~3)

Compile SYCL code for device only

[**-fsycl-host-only**](#fsycl~4)

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

[**-fsycl**](#fsycl~5)

Enable SYCL C++ extensions

[**-fsyntax-only**](#fsyntax)

Run the preprocessor, parser and semantic analysis stages

[**-fsystem-module**](#fsystem)

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

[**-ftemplate-backtrace-limit=**&lt;value&gt;](#ftemplate)

Set the maximum number of entries to print in a template instantiation backtrace (0 = no limit)

[**-ftemplate-depth=**&lt;value&gt;](#ftemplate~2)

Set the maximum depth of recursive template instantiation

[**-ftemporal-profile**](#ftemporal)

Generate instrumented code to collect temporal information

[**-ftest-coverage**](#ftest)

Produce gcov notes files (\*.gcno)

[**-fthin-link-bitcode=**&lt;value&gt;](#fthin)

Write minimized bitcode to &lt;file&gt; for the ThinLTO thin link only

[**-fthinlto-distributor=**&lt;path&gt;](#fthinlto)

Path to the ThinLTO distributor process. If specified, ThinLTO backend compilations will be distributed by LLD

**-fthinlto-index=**&lt;value&gt; Perform ThinLTO importing using provided function summary index

[**-ftime-report=**&lt;value&gt;](#ftime)

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

[**-ftime-trace-granularity=**&lt;value&gt;](#ftime~2)

Minimum time granularity (in microseconds) traced by time profiler

[**-ftime-trace-verbose**&lt;value&gt;](#ftime~3)

Make time trace capture verbose event details (e.g. source filenames). This can increase the size of the output by 2-3 times

[**-ftime-trace=**&lt;value&gt;](#ftime~4)

Similar to **-ftime-trace**. Specify the JSON file or a directory which will contain the JSON file

[**-ftime-trace**](#ftime~5)

Turn on time profiler. Generates JSON file based on output filename.

**-ftrap-function=**&lt;value&gt; Issue call to specified function rather than a trap instruction

[**-ftrapv-handler=**&lt;function name&gt;](#ftrapv)

Specify the function to be called on overflow

[**-ftrapv**](#ftrapv~2)

Trap on integer overflow

[**-ftrigraphs**](#ftrigraphs)

Process trigraph sequences

[**-ftrivial-auto-var-init-max-size=**&lt;value&gt;](#ftrivial)

Stop initializing trivial automatic stack variables if var size exceeds the specified number of instances (in bytes)

[**-ftrivial-auto-var-init-stop-after=**&lt;value&gt;](#ftrivial~2)

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

[**-ftrivial-auto-var-init=**&lt;value&gt;](#ftrivial~3)

Initialize trivial automatic stack variables. Defaults to 'uninitialized'

[**-funified-lto**](#funified)

Use the unified LTO pipeline

[**-funique-basic-block-section-names**](#funique)

Use unique names for basic block sections (ELF Only)

[**-funique-internal-linkage-names**](#funique~2)

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

[**-funique-source-file-identifier=**&lt;value&gt;](#funique~3)

Specify the source file identifier for **-funique-source-file-names**; uses the source file path if not specified

[**-funique-source-file-names**](#funique~4)

Allow the compiler to assume that each translation unit has a unique source file identifier (see **-funique-source-file-identifier**) at link time

[**-funroll-loops**](#funroll)

Turn on loop unroller

[**-funsafe-math-optimizations**](#funsafe)

Allow unsafe floating-point math optimizations which may decrease precision

[**-fuse-cuid=**&lt;value&gt;](#fuse)

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 overridden by option '-cuid=\[ID]' if it is specified.

[**-fuse-line-directives**](#fuse~2)

Use #line in preprocessed output

[**-fvalidate-ast-input-files-content**](#fvalidate)

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=**&lt;value&gt;](#fveclib=)

Use the given vector functions library. Note: **-fveclib=**{ArmPL,SLEEF,libmvec} implies **-fno-math-errno**. Note: **-fveclib**=*libmvec* on AArch64 requires GLIBC 2.40 or newer.

[**-fvectorize**](#fvectorize)

Enable the loop vectorization passes

[**-fverbose-asm**](#fverbose)

Generate verbose assembly output

[**-fverify-intermediate-code**](#fverify)

Enable verification of LLVM IR

[**-fvirtual-function-elimination**](#fvirtual)

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

[**-fvisibility-dllexport=**&lt;value&gt;](#fvisibility)

The visibility for dllexport definitions. If Keep is specified the visibility is not adjusted \[-fvisibility-from-dllstorageclass]

[**-fvisibility-externs-dllimport=**&lt;value&gt;](#fvisibility~2)

The visibility for dllimport external declarations. If Keep is specified the visibility is not adjusted \[-fvisibility-from-dllstorageclass]

[**-fvisibility-externs-nodllstorageclass=**&lt;value&gt;](#fvisibility~3)

The visibility for external declarations without an explicit DLL storage class. If Keep is specified the visibility is not adjusted \[-fvisibility-from-dllstorageclass]

[**-fvisibility-from-dllstorageclass**](#fvisibility~4)

Override the visibility of globals based on their final DLL storage class.

[**-fvisibility-global-new-delete-hidden**](#fvisibility~5)

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

[**-fvisibility-global-new-delete=**&lt;value&gt;](#fvisibility~6)

The visibility for global C++ operator new and delete declarations. If 'source' is specified the visibility is not adjusted

[**-fvisibility-inlines-hidden-static-local-var**](#fvisibility~7)

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

[**-fvisibility-inlines-hidden**](#fvisibility~8)

Give inline C++ member functions hidden visibility by default

[**-fvisibility-ms-compat**](#fvisibility~9)

Give global types 'default' visibility and global functions and variables 'hidden' visibility by default

[**-fvisibility-nodllstorageclass=**&lt;value&gt;](#fvisibility~10)

The visibility for definitions without an explicit DLL storage class. If Keep is specified the visibility is not adjusted \[-fvisibility-from-dllstorageclass]

[**-fvisibility=**&lt;value&gt;](#fvisibility=)

Set the default symbol visibility for all global definitions

[**-fwasm-exceptions**](#fwasm)

Use WebAssembly style exceptions

**-fwhole-program-vtables** Enables whole-program vtable optimization. Requires **-flto**

[**-fwin-cfg-mechanism=**&lt;value&gt;](#fwin)

Sets the mechanism to use for Control Flow Guard on Windows

[**-fwinx64-eh-unwind=**&lt;value&gt;](#fwinx64)

Set x64 Windows unwind info version

[**-fwinx64-eh-unwindv2=**&lt;value&gt;](#fwinx64~2)

(Legacy) Generate unwind v2 (epilog) information for x64 Windows

[**-fwrapv-pointer**](#fwrapv)

Treat pointer overflow as two's complement

[**-fwrapv**](#fwrapv~2)

Treat signed integer overflow as two's complement

[**-fwritable-strings**](#fwritable)

Store string literals as writable data

[**-fxl-pragma-pack**](#fxl)

Enable IBM XL #pragma pack handling

[**-fxray-always-emit-customevents**](#fxray)

Always emit \_\_xray\_customevent(...) calls even if the containing function is not always instrumented

[**-fxray-always-emit-typedevents**](#fxray~2)

Always emit \_\_xray\_typedevent(...) calls even if the containing function is not always instrumented

[**-fxray-always-instrument=**&lt;value&gt;](#fxray~3)

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

[**-fxray-attr-list=**&lt;value&gt;](#fxray~4)

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

[**-fxray-function-groups=**&lt;value&gt;](#fxray~5)

Only instrument 1 of N groups

[**-fxray-ignore-loops**](#fxray~6)

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

[**-fxray-instruction-threshold=**&lt;value&gt;](#fxray~7)

Sets the minimum function size to instrument with XRay

[**-fxray-instrumentation-bundle=**&lt;value&gt;](#fxray~8)

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-instrument**](#fxray~9)

Generate XRay instrumentation sleds on function entry and exit

[**-fxray-link-deps**](#fxray~10)

Link XRay runtime library when **-fxray-instrument** is specified (default)

[**-fxray-modes=**&lt;value&gt;](#fxray~11)

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

[**-fxray-never-instrument=**&lt;value&gt;](#fxray~12)

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

[**-fxray-selected-function-group=**&lt;value&gt;](#fxray~13)

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

[**-fxray-shared**](#fxray~14)

Enable shared library instrumentation with XRay

[**-fzero-call-used-regs=**&lt;value&gt;](#fzero)

Clear call-used registers upon function return (AArch64/RISC-V/x86 only)

[**-fzos-extensions**](#fzos)

Accept some non-standard constructs supported by the z/OS compiler

[**-fzvector**](#fzvector)

Enable System z vector language extension

[**-F** &lt;value&gt;](#F)

Add directory to framework include search path

[**-gcall-site-info**](#gcall)

Enable call site debug info

[**--gcc-install-dir=**&lt;value&gt;](#gcc)

Use GCC installation in the specified directory. The directory ends with path components like 'lib{,32,64}/gcc{,-cross}/$triple/$version'. Note: executables (e.g. ld) used by the compiler are not overridden by the selected GCC installation

**--gcc-toolchain=**&lt;value&gt; Specify a directory where Clang can find 'include' and 'lib{,32,64}/gcc{,-cross}/$triple/$version'. Clang will use the GCC installation with the largest version

[**--gcc-triple=**&lt;value&gt;](#gcc~2)

Search for the GCC installation with the specified triple.

**-gcodeview-command-line** Emit compiler path and command line into CodeView debug information

[**-gcodeview-ghash**](#gcodeview)

Emit type record hashes in a .debug$H section

[**-gcodeview**](#gcodeview~2)

Generate CodeView debug information

[**-gdwarf-2**](#gdwarf)

Generate source-level debug information with DWARF version 2

[**-gdwarf-3**](#gdwarf~2)

Generate source-level debug information with DWARF version 3

[**-gdwarf-4**](#gdwarf~3)

Generate source-level debug information with DWARF version 4

[**-gdwarf-5**](#gdwarf~4)

Generate source-level debug information with DWARF version 5

[**-gdwarf-6**](#gdwarf~5)

Generate source-level debug information with DWARF version 6

[**-gdwarf32**](#gdwarf32)

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

[**-gdwarf64**](#gdwarf64)

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

[**-gdwarf**](#gdwarf~6)

Generate source-level debug information with the default DWARF version

[**-gembed-source**](#gembed)

Embed source text in DWARF debug sections

**-gen-reproducer=**&lt;value&gt; Emit reproducer on (option: off, crash (default), error, always)

[**-gkey-instructions**](#gkey)

Enable Key Instructions, which reduces the jumpiness of debug stepping in optimized C/C++ code in some debuggers. DWARF only.

[**-gline-directives-only**](#gline)

Emit debug line info directives only

[**-gline-tables-only**](#gline~2)

Emit debug line number tables only

[**-gmodules**](#gmodules)

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

[**-gno-call-site-info**](#gno)

Disable call site debug info

[**-gno-codeview-command-line**](#gno~2)

Don't emit compiler path and command line into CodeView debug information

[**-gno-embed-source**](#gno~3)

Restore the default behavior of not embedding source text in DWARF debug sections

**-gno-inline-line-tables** Don't emit inline line tables.

[**--gpu-bundle-output**](#gpu)

Bundle output files of HIP device compilation

[**--gpu-instrument-lib=**&lt;value&gt;](#gpu~2)

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=**&lt;value&gt;](#gpu~3)

Default max threads per block for kernel launch bounds for HIP

[**--gpu-use-aux-triple-only**](#gpu~4)

Prepare '-aux-triple' only without populating '-aux-target-cpu' and '-aux-target-feature'.

[**-gpulibc**](#gpulibc)

Link the LLVM C Library for GPUs

[**-gsplit-dwarf=**&lt;value&gt;](#gsplit)

Set DWARF fission mode

[**-gstrict-dwarf**](#gstrict)

Restrict DWARF features to those defined in the specified version, avoiding features from later versions.

[**-gstructor-decl-linkage-names**](#gstructor)

Attach linkage names to C++ constructor/destructor declarations in DWARF.

[**-gz=**&lt;value&gt;](#gz=)

DWARF debug sections compression type

[**-G** &lt;size&gt;](#G)

Put objects of at most &lt;size&gt; bytes into small data section (MIPS / Hexagon)

[**-g**](#g)

Generate source-level debug information

[**--help-hidden**](#help)

Display help for hidden options

[**-help**](#help~2)

Display available options

[**--hip-device-lib=**&lt;value&gt;](#hip)

HIP device library

[**--hip-link**](#hip~2)

Link clang-offload-bundler bundles for HIP

[**--hip-path=**&lt;value&gt;](#hip~3)

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

[**--hip-version=**&lt;value&gt;](#hip~4)

HIP version in the format of major.minor.patch

[**--hipspv-pass-plugin=**&lt;dsopath&gt;](#hipspv)

path to a pass plugin for HIP to SPIR-V passes.

[**--hipstdpar-interpose-alloc**](#hipstdpar)

Replace all memory allocation / deallocation calls with hipManagedMalloc / hipFree equivalents

[**--hipstdpar-path=**&lt;value&gt;](#hipstdpar~2)

HIP Standard Parallel Algorithm Acceleration library path, used for finding and implicitly including the library header

[**--hipstdpar-prim-path=**&lt;value&gt;](#hipstdpar~3)

rocPrim path, required by the HIP Standard Parallel Algorithm Acceleration library, used to implicitly include the rocPrim library

[**--hipstdpar-thrust-path=**&lt;value&gt;](#hipstdpar~4)

rocThrust path, required by the HIP Standard Parallel Algorithm Acceleration library, used to implicitly include the rocThrust library

[**--hipstdpar**](#hipstdpar~5)

Enable HIP acceleration for standard parallel algorithms

[**-H**](#H)

Show header includes and nesting depth

[**-I-**](#I)

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

[**-iapinotes-modules** &lt;directory&gt;](#iapinotes)

Add directory to the API notes search path referenced by module name

[**-ibuiltininc**](#ibuiltininc)

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

[**-idirafter** &lt;value&gt;](#idirafter)

Add directory to AFTER include search path

[**-iframeworkwithsysroot** &lt;directory&gt;](#iframeworkwithsysroot)

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

[**-iframework** &lt;value&gt;](#iframework)

Add directory to SYSTEM framework search path

[**-ignore-pch**](#ignore)

Disable precompiled headers, overrides **-emit-pch** and **-include-pch**

[**-imacros** &lt;file&gt;](#imacros)

Include macros from file before parsing

[**-include-pch** &lt;file&gt;](#include)

Include precompiled header file

[**-include** &lt;file&gt;](#include~2)

Include file before parsing

[**-iprefix** &lt;dir&gt;](#iprefix)

Set the **-iwithprefix**/-iwithprefixbefore prefix

[**-iquote** &lt;directory&gt;](#iquote)

Add directory to QUOTE include search path

[**-isysroot** &lt;dir&gt;](#isysroot)

Set the system root directory (usually /)

[**-isystem-after** &lt;directory&gt;](#isystem)

Add directory to end of the SYSTEM include search path

[**-isystem** &lt;directory&gt;](#isystem~2)

Add directory to SYSTEM include search path

[**-ivfsoverlay** &lt;value&gt;](#ivfsoverlay)

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

[**-iwithprefixbefore** &lt;dir&gt;](#iwithprefixbefore)

Set directory to include search path with prefix

[**-iwithprefix** &lt;dir&gt;](#iwithprefix)

Set directory to SYSTEM include search path with prefix

[**-iwithsysroot** &lt;directory&gt;](#iwithsysroot)

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

[**-I** &lt;dir&gt;](#I~2)

Add directory to the end of the list of include search paths

[**--libclc-lib=**&lt;value&gt;](#libclc)

Namespec of libclc OpenCL bitcode library to link

[**--libomptarget-amdgcn-bc-path=**&lt;value&gt;](#libomptarget)

Path to libomptarget-amdgcn bitcode library

[**--libomptarget-amdgpu-bc-path=**&lt;value&gt;](#libomptarget~2)

Path to libomptarget-amdgcn bitcode library

[**--libomptarget-nvptx-bc-path=**&lt;value&gt;](#libomptarget~3)

Path to libomptarget-nvptx bitcode library

[**--libomptarget-spirv-bc-path=**&lt;value&gt;](#libomptarget~4)

Path to libomptarget-spirv bitcode library

[**-lifetime-safety-lifetimebound-macro=**&lt;macro&gt;](#lifetime)

Use the given macro name when suggesting lifetimebound attributes

[**-lifetime-safety-max-cfg-blocks=**&lt;value&gt;](#lifetime~2)

Skip LifetimeSafety analysis for functions with CFG block count exceeding this threshold. Specify 0 for no limit.

[**-L** &lt;dir&gt;](#L)

Add directory to library search path

[**-mabi**=*quadword-atomics*](#mabi)

Enable quadword atomics ABI on AIX (AIX PPC64 only). Uses lqarx/stqcx. instructions.

[**-mabicalls**](#mabicalls)

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

[**-maix-shared-lib-tls-model-opt**](#maix)

For shared library loaded with the main program, change local-dynamic access(es) to initial-exec access(es) at the function level (AIX 64-bit only).

[**-maix-small-local-dynamic-tls**](#maix~2)

Produce a faster access sequence for local-dynamic TLS variables where the offset from the TLS base is encoded as an immediate operand (AIX 64-bit only). This access sequence is not used for variables larger than 32KB.

[**-maix-small-local-exec-tls**](#maix~3)

Produce a faster access sequence for local-exec TLS variables where the offset from the TLS base is encoded as an immediate operand (AIX 64-bit only). This access sequence is not used for variables larger than 32KB.

[**-maix-struct-return**](#maix~4)

Return all structs in memory (PPC32 only)

[**-maix-use-ptrgl**](#maix~5)

Use .\_ptrgl routine for indirect calls (AIX only)

[**-malign-branch-boundary=**&lt;value&gt;](#malign)

Specify the boundary's size to align branches

[**-malign-branch=**&lt;value&gt;](#malign~2)

Specify types of branches to align

[**-malign-double**](#malign~3)

Align doubles to two words in structs (x86 only)

[**-maltivec**](#maltivec)

Enable AltiVec vector initializer syntax

[**-mamdgpu-expand-waitcnt-profiling**](#mamdgpu)

Expand s\_waitcnt instructions to help PC-sampling profilers identify memory stalls. Instead of a single waitcnt(target), emits waitcnt(N-1), waitcnt(N-2), ..., waitcnt(target). (AMDGPU only)

[**-mamdgpu-ieee**](#mamdgpu~2)

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)

[**-mamdgpu-precise-memory-op**](#mamdgpu~3)

Enable precise memory mode (AMDGPU only)

[**-mannotate-tablejump**](#mannotate)

Enable annotate table jump instruction to correlate it with the jump table.

[**-mapx-features=**&lt;value&gt;](#mapx)

Enable features of APX

[**-mapx-inline-asm-use-gpr32**](#mapx~2)

Enable use of GPR32 in inline assembly for APX

[**-march=**&lt;value&gt;](#march=)

For a list of available architectures for the target use '-mcpu=help'

[**-marm64x**&lt;value&gt;](#marm64x)

Build as a hybrid ARM64X object/image

[**-mbackchain**](#mbackchain)

Link stack frames through backchain on System Z

[**-mbranch-protection=**&lt;value&gt;](#mbranch)

Enforce targets of indirect branches and function returns

[**-mbranches-within-32B-boundaries**](#mbranches)

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

[**-mcabac**](#mcabac)

Enable CABAC instructions

[**-mcf-branch-label-scheme=**&lt;value&gt;](#mcf)

Select label scheme for branch control-flow architecture protection

[**-mcmse**](#mcmse)

Allow use of CMSE (Armv8-M Security Extensions)

[**-mcode-object-version=**&lt;value&gt;](#mcode)

Specify code object ABI version. Defaults to 6. (AMDGPU only)

[**-mconstructor-aliases**](#mconstructor)

Enable emitting complete constructors and destructors as aliases when possible

[**-mcpu=**&lt;value&gt;](#mcpu=)

For a list of available CPUs for the target use '-mcpu=help'

[**-mcrbits**](#mcrbits)

Control the CR-bit tracking feature on PowerPC. \`\`-mcrbits\`\` (the enablement of CR-bit tracking support) is the default for POWER8 and above, as well as for all other CPUs when optimization is applied (**-O2** and above).

[**-mcrc**](#mcrc)

Allow use of CRC instructions (ARM/Mips only)

[**-mcumode**](#mcumode)

Specify CU wavefront execution mode (AMDGPU only)

[**-mdaz-ftz**](#mdaz)

Globally set the denormals-are-zero (DAZ) and flush-to-zero (FTZ) bits in the floating-point control register on program startup

[**-mdefault-visibility-export-mapping=**&lt;value&gt;](#mdefault)

Mapping between default visibility and export

[**-mdiv32**](#mdiv32)

Use div.w\[u] and mod.w\[u] instructions with input not sign-extended.

[**-mdouble=**&lt;n](#mdouble=)

Force double to be &lt;n&gt; bits

[**-MD**](#MD)

Write a depfile containing user and system headers

[**-meabi** &lt;value&gt;](#meabi)

Set EABI type. Default depends on triple)

[**-membedded-data**](#membedded)

Place constants in the .rodata section instead of the .sdata section even if they meet the **-G** &lt;size&gt; threshold (MIPS)

[**-menable-experimental-extensions**](#menable)

Enable use of experimental RISC-V extensions.

[**-mexec-model=**&lt;value&gt;](#mexec)

Execution model (WebAssembly only)

[**-mexecute-only**](#mexecute)

Disallow generation of data access to code sections (AArch64/ARM only)

[**-mexperimental-mtune-syntax**](#mexperimental)

Enable experimental syntax for the '-mtune' flag (RISC-V only)

[**-mextern-sdata**](#mextern)

Assume that externally defined data is in the small data if it meets the **-G** &lt;size&gt; threshold (MIPS)

[**-mfentry**](#mfentry)

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

[**-mfix-cmse-cve-2021-35465**](#mfix)

Work around VLLDM erratum CVE-2021-35465 (ARM only)

**-mfix-cortex-a53-835769** Work around Cortex-A53 erratum 835769 (AArch64 only)

**-mfix-cortex-a53-843419** Work around Cortex-A53 erratum 843419 (AArch64 only)

[**-mfix-cortex-a57-aes-1742098**](#mfix~2)

Work around Cortex-A57 Erratum 1742098 (ARM only)

[**-mfix-cortex-a72-aes-1655431**](#mfix~3)

Work around Cortex-A72 Erratum 1655431 (ARM only)

[**-mfix-gr712rc**](#mfix~4)

Enable workarounds for GR712RC errata

[**-mfix-r5900**](#mfix~5)

Enable R5900 short loop erratum fix

[**-mfix-ut700**](#mfix~6)

Enable workarounds for UT700 errata

[**-mfp32**](#mfp32)

Use 32-bit floating point registers (MIPS only)

[**-mfp64**](#mfp64)

Use 64-bit floating point registers (MIPS only)

[**-mfpxx**](#mfpxx)

Avoid FPU mode dependent operations when used with the O32 ABI

[**-mframe-chain=**&lt;value&gt;](#mframe)

Select the frame chain model used to emit frame records (Arm only).

[**-mfrecipe**](#mfrecipe)

Enable frecipe.{s/d} and frsqrte.{s/d}

[**-mfunction-return=**&lt;value&gt;](#mfunction)

Replace returns with jumps to \`\`\_\_x86\_return\_thunk\`\` (x86 only, error otherwise)

[**-MF** &lt;file&gt;](#MF)

Write depfile output from **-MMD**, **-MD**, **-MM**, or **-M** to &lt;file&gt;

[**-mgeneral-regs-only**](#mgeneral)

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

[**-mglobal-merge**](#mglobal)

Enable merging of globals

[**-mgpopt**](#mgpopt)

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

[**-mguard=**&lt;value&gt;](#mguard=)

Enable or disable Control Flow Guard checks and guard tables emission

[**-MG**](#MG)

Add missing headers to depfile

[**-mharden-sls=**&lt;value&gt;](#mharden)

Select straight-line speculation hardening scope (ARM/AArch64/X86 only). &lt;arg&gt; must be: all, none, retbr(ARM/AArch64), blr(ARM/AArch64), comdat(ARM/AArch64), nocomdat(ARM/AArch64), return(X86), indirect-jmp(X86)

[**-mhvx-ieee-fp**](#mhvx)

Enable Hexagon HVX IEEE floating-point

[**-mhvx-length=**&lt;value&gt;](#mhvx~2)

Set Hexagon Vector Length

[**-mhvx-qfloat=**&lt;value&gt;](#mhvx~3)

Enable Hexagon HVX QFloat instructions with mode: strict-ieee, ieee, lossy, legacy (v79+)

[**-mhvx-qfloat**](#mhvx~4)

Enable Hexagon HVX QFloat instructions

[**-mhvx=**&lt;value&gt;](#mhvx=)

Enable Hexagon Vector eXtensions

[**-mhvx**](#mhvx~5)

Enable Hexagon Vector eXtensions

[**-miamcu**](#miamcu)

Use Intel MCU ABI

[**-mignore-xcoff-visibility**](#mignore)

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

[**-mimplicit-float**](#mimplicit)

Generate implicit floating point or vector instructions

[**-mincremental-linker-compatible**](#mincremental)

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

[**-mindirect-branch-cs-prefix**](#mindirect)

Add cs prefix to call and jmp to indirect thunk

**-mindirect-jump=**&lt;value&gt; Change indirect jump instructions to inhibit speculation

[**-mios-version-min=**&lt;value&gt;](#mios)

Set iOS deployment target

[**-mips1**](#mips1)

Equivalent to **-march**=*mips1*

[**-mips2**](#mips2)

Equivalent to **-march**=*mips2*

[**-mips32r2**](#mips32r2)

Equivalent to **-march**=*mips32r2*

[**-mips32r3**](#mips32r3)

Equivalent to **-march**=*mips32r3*

[**-mips32r5**](#mips32r5)

Equivalent to **-march**=*mips32r5*

[**-mips32r6**](#mips32r6)

Equivalent to **-march**=*mips32r6*

[**-mips32**](#mips32)

Equivalent to **-march**=*mips32*

[**-mips3**](#mips3)

Equivalent to **-march**=*mips3*

[**-mips4**](#mips4)

Equivalent to **-march**=*mips4*

[**-mips5**](#mips5)

Equivalent to **-march**=*mips5*

[**-mips64r2**](#mips64r2)

Equivalent to **-march**=*mips64r2*

[**-mips64r3**](#mips64r3)

Equivalent to **-march**=*mips64r3*

[**-mips64r5**](#mips64r5)

Equivalent to **-march**=*mips64r5*

[**-mips64r6**](#mips64r6)

Equivalent to **-march**=*mips64r6*

[**-mips64**](#mips64)

Equivalent to **-march**=*mips64*

[**-MJ** &lt;value&gt;](#MJ)

Write a compilation database entry per input

[**-mlam-bh**](#mlam)

Enable amswap\[\_db].{b/h} and amadd\[\_db].{b/h}

[**-mlamcas**](#mlamcas)

Enable amcas\[\_db].{b/h/w/d}

[**-mlasx**](#mlasx)

Enable Loongson Advanced SIMD Extension (LASX).

[**-mld-seq-sa**](#mld)

Do not generate same-address load-load barrier instructions (dbar 0x700)

[**-mlink-builtin-bitcode-postopt**](#mlink)

Link builtin bitcodes after the optimization pipeline

[**-mllvm=**&lt;arg&gt;](#mllvm=)

Alias for **-mllvm**

[**-mllvm** &lt;value&gt;](#mllvm)

Additional arguments to forward to LLVM's option processing

[**-mlocal-sdata**](#mlocal)

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

[**-mlong-calls**](#mlong)

Generate branches with extended addressability, usually via indirect jumps.

[**-mlong-double-128**](#mlong~2)

Force long double to be 128 bits

[**-mlong-double-64**](#mlong~3)

Force long double to be 64 bits

[**-mlong-double-80**](#mlong~4)

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

[**-mlr-for-calls-only**](#mlr)

Do not allocate the LR register for general purpose usage, only for calls. (AArch64 only)

[**-mlsx**](#mlsx)

Enable Loongson SIMD Extension (LSX).

[**-mlvi-cfi**](#mlvi)

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

[**-mlvi-hardening**](#mlvi~2)

Enable all mitigations for Load Value Injection (LVI)

[**-mmacos-version-min=**&lt;value&gt;](#mmacos)

Set macOS deployment target

[**-mmadd4**](#mmadd4)

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

[**-mmark-bti-property**](#mmark)

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

[**-MMD**](#MMD)

Write a depfile containing user headers

[**-mmemops**](#mmemops)

Enable generation of memop instructions

[**-mmlir** &lt;value&gt;](#mmlir)

Additional arguments to forward to MLIR's option processing

[**-mms-bitfields**](#mms)

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

[**-mmsa**](#mmsa)

Enable MSA ASE (MIPS only)

[**-mmt**](#mmt)

Enable MT ASE (MIPS only)

[**-MM**](#MM)

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

[**-mno-abicalls**](#mno)

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

[**-mno-aix-use-ptrgl**](#mno~2)

Emit indirect calls inline (AIX only) (default)

[**-mno-amdgpu-precise-memory-op**](#mno~3)

Disable precise memory mode (AMDGPU only)

**-mno-annotate-tablejump** Disable annotate table jump instruction to correlate it with the jump table.

[**-mno-apx-features=**&lt;value&gt;](#mno~4)

Disable features of APX

[**-mno-bti-at-return-twice**](#mno~5)

Do not add a BTI instruction after a setjmp or other return-twice construct (Arm/AArch64 only)

[**-mno-constructor-aliases**](#mno~6)

Disable emitting complete constructors and destructors as aliases when possible

[**-mno-crc**](#mno~7)

Disallow use of CRC instructions (Mips only)

[**-mno-cumode**](#mno~8)

Specify WGP wavefront execution mode (AMDGPU only)

[**-mno-daz-ftz**](#mno~9)

Do not globally set the denormals-are-zero (DAZ) and flush-to-zero (FTZ) bits in the floating-point control register on program startup

[**-mno-div32**](#mno~10)

Do not use div.w\[u] and mod.w\[u] instructions with input not sign-extended.

[**-mno-embedded-data**](#mno~11)

Do not place constants in the .rodata section instead of the .sdata if they meet the **-G** &lt;size&gt; threshold (MIPS)

[**-mno-execute-only**](#mno~12)

Allow generation of data access to code sections (AArch64/ARM only)

[**-mno-experimental-mtune-syntax**](#mno~13)

Disable experimental syntax for the '-mtune' flag (RISC-V only)

[**-mno-extern-sdata**](#mno~14)

Do not assume that externally defined data is in the small data if it meets the **-G** &lt;size&gt; threshold (MIPS)

[**-mno-fix-cmse-cve-2021-35465**](#mno~15)

Don't work around VLLDM erratum CVE-2021-35465 (ARM only)

[**-mno-fix-cortex-a53-835769**](#mno~16)

Don't work around Cortex-A53 erratum 835769 (AArch64 only)

[**-mno-fix-cortex-a53-843419**](#mno~17)

Don't work around Cortex-A53 erratum 843419 (AArch64 only)

[**-mno-fix-cortex-a57-aes-1742098**](#mno~18)

Don't work around Cortex-A57 Erratum 1742098 (ARM only)

[**-mno-fix-cortex-a72-aes-1655431**](#mno~19)

Don't work around Cortex-A72 Erratum 1655431 (ARM only)

[**-mno-fix-r5900**](#mno~20)

Disable R5900 short loop erratum fix

[**-mno-fmv**](#mno~21)

Disable function multiversioning

[**-mno-frecipe**](#mno~22)

Disable frecipe.{s/d} and frsqrte.{s/d}

[**-mno-gather**](#mno~23)

Disable generation of gather instructions in auto-vectorization(x86 only)

[**-mno-global-merge**](#mno~24)

Disable merging of globals

[**-mno-gpopt**](#mno~25)

Do not use GP relative accesses for symbols known to be in a small data section (MIPS)

[**-mno-hvx-ieee-fp**](#mno~26)

Disable Hexagon HVX IEEE floating-point

[**-mno-hvx-qfloat**](#mno~27)

Disable Hexagon HVX QFloat instructions

[**-mno-hvx**](#mno~28)

Disable Hexagon Vector eXtensions

[**-mno-implicit-float**](#mno~29)

Don't generate implicit floating point or vector instructions

[**-mno-incremental-linker-compatible**](#mno~30)

(integrated-as) Emit an object file which cannot be used with an incremental linker

[**-mno-lam-bh**](#mno~31)

Disable amswap\[\_db].{b/h} and amadd\[\_db].{b/h}

[**-mno-lamcas**](#mno~32)

Disable amcas\[\_db].{b/h/w/d}

[**-mno-lasx**](#mno~33)

Disable Loongson Advanced SIMD Extension (LASX).

[**-mno-ld-seq-sa**](#mno~34)

Generate same-address load-load barrier instructions (dbar 0x700)

[**-mno-local-sdata**](#mno~35)

Do not extend the **-G** behaviour to object local data (MIPS)

[**-mno-long-calls**](#mno~36)

Restore the default behaviour of not generating long calls

[**-mno-lsx**](#mno~37)

Disable Loongson SIMD Extension (LSX).

[**-mno-lvi-cfi**](#mno~38)

Disable control-flow mitigations for Load Value Injection (LVI)

[**-mno-lvi-hardening**](#mno~39)

Disable mitigations for Load Value Injection (LVI)

[**-mno-madd4**](#mno~40)

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

[**-mno-memops**](#mno~41)

Disable generation of memop instructions

[**-mno-movt**](#mno~42)

Disallow use of movt/movw pairs (ARM only)

[**-mno-ms-bitfields**](#mno~43)

Do not set the default structure layout to be compatible with the Microsoft compiler standard

[**-mno-msa**](#mno~44)

Disable MSA ASE (MIPS only)

[**-mno-mt**](#mno~45)

Disable MT ASE (MIPS only)

[**-mno-neg-immediates**](#mno~46)

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

[**-mno-nvj**](#mno~47)

Disable generation of new-value jumps

[**-mno-nvs**](#mno~48)

Disable generation of new-value stores

[**-mno-odd-spreg**](#mno~49)

Disable odd single-precision floating point registers

[**-mno-outline-atomics**](#mno~50)

Don't generate local calls to out-of-line atomic operations

[**-mno-outline**](#mno~51)

Disable function outlining (AArch64,Arm,RISC-V,X86 only)

[**-mno-packets**](#mno~52)

Disable generation of instruction packets

[**-mno-pic-data-is-text-relative**](#mno~53)

Don't assume data segments are relative to text segment

[**-mno-regnames**](#mno~54)

Use only register numbers when writing assembly output

[**-mno-relax-pic-calls**](#mno~55)

Do not produce relaxation hints for linkers to try optimizing PIC call sequences into direct calls (MIPS only)

[**-mno-relax**](#mno~56)

Disable linker relaxation

[**-mno-restrict-it**](#mno~57)

Allow generation of complex IT blocks.

[**-mno-save-restore**](#mno~58)

Disable using library calls for save and restore

[**-mno-scalar-strict-align**](#mno~59)

Allow scalar memory accesses to be unaligned (RISC-V only)

[**-mno-scatter**](#mno~60)

Disable generation of scatter instructions in auto-vectorization(x86 only)

[**-mno-scq**](#mno~61)

Disable sc.q instruction.

[**-mno-seses**](#mno~62)

Disable speculative execution side effect suppression (SESES)

[**-mno-sramecc**](#mno~63)

Disable SRAMECC (AMDGPU only)

[**-mno-stack-arg-probe**](#mno~64)

Disable stack probes which are enabled by default

[**-mno-strict-align**](#mno~65)

Allow memory accesses to be unaligned (AArch64/LoongArch/RISC-V only)

[**-mno-tgsplit**](#mno~66)

Disable threadgroup split execution mode (AMDGPU only)

[**-mno-tls-direct-seg-refs**](#mno~67)

Disable direct TLS access through segment registers

[**-mno-tocdata=**&lt;value&gt;](#mno~68)

Specifies a list of variables to be exempt from the TOC data transformation.

[**-mno-tocdata**](#mno~69)

This is the default. TOC data transformation is not applied to any variables. Only variables specified explicitly in **-mtocdata=** will have the TOC data transformation.

[**-mno-unaligned-access**](#mno~70)

Force all memory accesses to be aligned (AArch32/MIPSr6 only)

[**-mno-unaligned-symbols**](#mno~71)

Expect external char-aligned symbols to be without ABI alignment (SystemZ only)

[**-mno-v8plus**](#mno~72)

Disable V8+ mode

[**-mno-vector-strict-align**](#mno~73)

Allow vector memory accesses to be unaligned (RISC-V only)

[**-mno-wavefrontsize64**](#mno~74)

Specify wavefront size 32 mode (AMDGPU only)

[**-mno-xnack**](#mno~75)

Disable XNACK (AMDGPU only)

[**-mno-zos-ppa1-name**](#mno~76)

Do not emit the function name in PPA1 on z/OS

[**-mnocrc**](#mnocrc)

Disallow use of CRC instructions (ARM only)

[**-mnop-mcount**](#mnop)

Generate mcount/\_\_fentry\__ calls as nops. To activate they need to be patched in.

[**-mnvj**](#mnvj)

Enable generation of new-value jumps

[**-mnvs**](#mnvs)

Enable generation of new-value stores

[**-modd-spreg**](#modd)

Enable odd single-precision floating point registers

[**-module-dependency-dir** &lt;value&gt;](#module)

Directory to dump module dependencies to

[**-module-file-info**](#module~2)

Provide information about a particular module file

[**-momit-leaf-frame-pointer**](#momit)

Omit frame pointer setup for leaf functions

[**-moutline-atomics**](#moutline)

Generate local calls to out-of-line atomic operations

[**-moutline**](#moutline~2)

Enable function outlining (AArch64,Arm,RISC-V,X86 only)

[**-mpacked-stack**](#mpacked)

Use packed stack layout (SystemZ only).

[**-mpackets**](#mpackets)

Enable generation of instruction packets

[**-mpad-max-prefix-size=**&lt;value&gt;](#mpad)

Specify maximum number of prefixes to use for padding

[**-mpic-data-is-text-relative**](#mpic)

Assume data segments are relative to text segment

[**-mprefer-vector-width=**&lt;value&gt;](#mprefer)

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

[**-mprintf-kind=**&lt;value&gt;](#mprintf)

Specify the printf lowering scheme (AMDGPU only), allowed values are "hostcall"(printing happens during kernel execution, this scheme relies on hostcalls which require system to support pcie atomics) and "buffered"(printing happens after all kernel threads exit, this uses a printf buffer and does not rely on pcie atomic support)

[**-MP**](#MP)

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

[**-mqdsp6-compat**](#mqdsp6)

Enable hexagon-qdsp6 backward compatibility

[**-MQ** &lt;value&gt;](#MQ)

Specify name of main file output to quote in depfile

[**-mrecip=**&lt;value&gt;](#mrecip=)

Control use of approximate reciprocal and reciprocal square root instructions followed by &lt;n&gt; iterations of Newton-Raphson refinement. &lt;value&gt; = ( \['!'] \['vec-'] ('rcp'|'sqrt') \[('h'|'s'|'d')] \[':'&lt;n&gt;] ) | 'all' | 'default' | 'none'

[**-mrecip**](#mrecip)

Equivalent to '-mrecip=all'

[**-mrecord-mcount**](#mrecord)

Generate a \_\_mcount\_loc section entry for each \_\_fentry\__ call.

[**-mregnames**](#mregnames)

Use full register names when writing assembly output

[**-mrelax-all**](#mrelax)

(integrated-as) Relax all machine instructions

[**-mrelax-pic-calls**](#mrelax~2)

Produce relaxation hints for linkers to try optimizing PIC call sequences into direct calls (MIPS only)

[**-mrelax**](#mrelax~3)

Enable linker relaxation

[**-mreserve-frame-pointer-reg**](#mreserve)

Reserve the frame pointer register even if the function doesn't have a frame

[**-mrestrict-it**](#mrestrict)

Disallow generation of complex IT blocks. It is off by default.

[**-mrtd**](#mrtd)

Make StdCall calling convention the default

[**-mrvv-vector-bits=**&lt;value&gt;](#mrvv)

Specify the size in bits of an RVV vector register

[**-msave-reg-params**](#msave)

Save arguments passed by registers to ABI-defined stack positions

[**-msave-restore**](#msave~2)

Enable using library calls for save and restore

[**-mscalar-strict-align**](#mscalar)

Force all scalar memory accesses to be aligned (RISC-V only)

[**-mscq**](#mscq)

Enable sc.q instruction.

[**-mscs-reg=**&lt;value&gt;](#mscs)

Register holding the shadow call stack pointer: one of the callee-saved registers r16-r27, default r18 (Hexagon only). The chosen register must also be reserved with **-ffixed-**&lt;reg&gt;.

[**-mseses**](#mseses)

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

[**-msign-return-address=**&lt;value&gt;](#msign)

Select return address signing scope

[**-msimd=**&lt;value&gt;](#msimd=)

Select the SIMD extension(s) to be enabled in LoongArch either 'none', 'lsx', 'lasx'.

[**-mskip-rax-setup**](#mskip)

Skip setting up RAX register when passing variable arguments (x86 only)

[**-msmall-data-limit=**&lt;value&gt;](#msmall)

Put global and static data smaller than the limit into a special section

[**-msoft-float**](#msoft)

Use software floating point

[**-msramecc**](#msramecc)

Enable SRAMECC (AMDGPU only)

[**-msse2avx**](#msse2avx)

Specify that the assembler should encode SSE instructions with VEX prefix

[**-mstack-alignment=**&lt;value&gt;](#mstack)

Set the stack alignment

[**-mstack-arg-probe**](#mstack~2)

Enable stack probes

[**-mstack-probe-size=**&lt;value&gt;](#mstack~3)

Set the stack probe size

[**-mstack-protector-guard-offset=**&lt;value&gt;](#mstack~4)

Use the given offset for addressing the stack-protector guard

[**-mstack-protector-guard-record**](#mstack~5)

Generate a \_\_stack\_protector\_loc section entry for each load of the stack-protector guard address.

[**-mstack-protector-guard-reg=**&lt;value&gt;](#mstack~6)

Use the given reg for addressing the stack-protector guard

[**-mstack-protector-guard-symbol=**&lt;value&gt;](#mstack~7)

Use the given symbol for addressing the stack-protector guard

[**-mstack-protector-guard-value-width=**&lt;value&gt;](#mstack~8)

Use the given width for the value of the the stack-protector guard (AArch64-only)

[**-mstack-protector-guard=**&lt;value&gt;](#mstack~9)

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

[**-mstackrealign**](#mstackrealign)

Force realign the stack at entry to every function

[**-mstrict-align**](#mstrict)

Force all memory accesses to be aligned (AArch64/LoongArch/RISC-V only)

[**-msve-streaming-vector-bits=**&lt;value&gt;](#msve)

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

[**-msve-vector-bits=**&lt;value&gt;](#msve~2)

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

[**-msvr4-struct-return**](#msvr4)

Return small structs in registers (PPC32 only)

[**-mtargetos=**&lt;value&gt;](#mtargetos=)

Set the deployment target to be the specified OS and OS version

[**-mtgsplit**](#mtgsplit)

Enable threadgroup split execution mode (AMDGPU only)

[**-mthread-model** &lt;value&gt;](#mthread)

The thread model to use. Defaults to 'posix')

[**-mtls-dialect=**&lt;value&gt;](#mtls)

Which thread-local storage dialect to use for dynamic accesses of TLS variables

[**-mtls-direct-seg-refs**](#mtls~2)

Enable direct TLS access through segment registers (default)

[**-mtls-size=**&lt;value&gt;](#mtls~3)

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*)

[**-mtocdata=**&lt;value&gt;](#mtocdata=)

Specifies a list of variables to which the TOC data transformation will be applied.

[**-mtocdata**](#mtocdata)

All suitable variables will have the TOC data transformation applied

[**-mtp=**&lt;value&gt;](#mtp=)

Thread pointer access method. For AArch32: 'soft' uses a function call, or 'tpidrurw', 'tpidruro' or 'tpidrprw' use the three CP15 registers. 'cp15' is an alias for 'tpidruro'. For AArch64: 'tpidr\_el0', 'tpidr\_el1', 'tpidr\_el2', 'tpidr\_el3' or 'tpidrro\_el0' use the five system registers. 'elN' is an alias for 'tpidr\_elN'.

[**-mtune=**&lt;value&gt;](#mtune=)

Only supported on AArch64, PowerPC, RISC-V, SPARC, SystemZ, and X86

[**-MT** &lt;value&gt;](#MT)

Specify name of main file output in depfile

[**-multi-lib-config=**&lt;file&gt;](#multi)

Path to the YAML configuration file to be used for multilib selection

[**-munaligned-access**](#munaligned)

Allow memory accesses to be unaligned (AArch32/MIPSr6 only)

[**-munaligned-symbols**](#munaligned~2)

Expect external char-aligned symbols to be without ABI alignment (SystemZ only)

[**-mv8plus**](#mv8plus)

Enable V8+ mode, allowing use of 64-bit V9 instructions in 32-bit code

[**-mvector-strict-align**](#mvector)

Force all vector memory accesses to be aligned (RISC-V only)

[**-mvevpu**](#mvevpu)

Emit VPU instructions for VE

[**-MV**](#MV)

Use NMake/Jom format for the depfile

[**-mwavefrontsize64**](#mwavefrontsize64)

Specify wavefront size 64 mode (AMDGPU only)

[**-mxcoff-build-id=**&lt;0xHEXSTRING&gt;](#mxcoff)

On AIX, request creation of a build-id string, "0xHEXSTRING", in the string table of the loader section inside the linked binary

[**-mxcoff-roptr**](#mxcoff~2)

Place constant objects with relocatable address values in the RO data section and add **-bforceimprw** to the linker flags (AIX only)

[**-mxnack**](#mxnack)

Enable XNACK (AMDGPU only)

[**-mzilsd-strict-align**](#mzilsd)

Force all Zilsd/Zclsd memory accesses to be 8-byte aligned (RISC-V only)

[**-mzilsd-word-align**](#mzilsd~2)

Allow Zilsd/Zclsd memory accesses to be 4-byte aligned (RISC-V only)

[**-mzos-hlq-clang=**&lt;ClangHLQ&gt;](#mzos)

High level qualifier for z/OS C++RT side deck datasets

[**-mzos-hlq-csslib=**&lt;CsslibHLQ&gt;](#mzos~2)

High level qualifier for z/OS CSSLIB dataset

[**-mzos-hlq-le=**&lt;LeHLQ&gt;](#mzos~3)

High level qualifier for z/OS Language Environment datasets

[**-mzos-ppa1-name**](#mzos~4)

Emit the function name in PPA1 on z/OS

[**-mzos-sys-include=**&lt;SysInclude&gt;](#mzos~5)

Path to system headers on z/OS

[**-mzos-target=**&lt;value&gt;](#mzos~6)

Set the z/OS release of the runtime environment

[**-M**](#M)

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

[**-no-canonical-prefixes**](#no)

Use relative paths for invoking subcommands

[**--no-cuda-include-ptx=**&lt;value&gt;](#no~2)

Do not include PTX for the following GPU architecture (e.g. sm\_35) or 'all'. May be specified more than once.

**--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-default-config**](#no~3)

Disable loading default configuration files

[**--no-gpu-bundle-output**](#no~4)

Do not bundle output files of HIP device compilation

[**-no-hip-rt**](#no~5)

Do not link against HIP runtime libraries

[**--no-offload-arch=**&lt;value&gt;](#no~6)

Remove CUDA/HIP offloading device architecture (e.g. sm\_35, gfx906) from the list of devices to compile for. 'all' resets the list to its default value.

[**--no-offload-inc**](#no~7)

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

**--no-offload-new-driver** Don't Use the new driver for offloading compilation.

[**--no-offloadlib**](#no~8)

Do not link device library for CUDA/HIP/SYCL device compilation

[**--no-system-header-prefix=**&lt;prefix&gt;](#no~9)

Treat all #include paths starting with &lt;prefix&gt; as not including a system header.

[**-no-use-spirv-backend**](#no~10)

Do not use the SPIRV backend for compilation

[**--no-wasm-opt**](#no~11)

Disable the wasm-opt optimizer

[**--no-xcselect**](#no~12)

Disable xcselect SDK injection on macOS

[**-nobuiltininc**](#nobuiltininc)

Disable builtin #include directories only

[**-nohipwrapperinc**](#nohipwrapperinc)

Do not include the default HIP wrapper headers and include paths

[**-nolibsycl**](#nolibsycl)

Do not link the SYCL runtime library

[**-nostdinc**++](#nostdinc)

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

[**-nostdinc**](#nostdinc~2)

Disable both standard system #include directories and builtin #include directories

[**-nostdlibinc**](#nostdlibinc)

Disable standard system #include directories only

[**-ObjC**++](#ObjC)

Treat source input files as Objective-C++ inputs

[**-ObjC**](#ObjC~2)

Treat source input files as Objective-C inputs

[**-object-file-name=**&lt;file&gt;](#object)

Set the output &lt;file&gt; for debug infos

[**-Ofast**&lt;value&gt;](#Ofast)

Deprecated; use '-O3 **-ffast-math**' for the same behavior, or '-O3' to enable only conforming optimizations

[**--offload-arch-tool=**&lt;value&gt;](#offload)

Tool used for detecting offloading architectures in the system.

[**--offload-arch=**&lt;value&gt;](#offload~2)

Specify an offloading device architecture for CUDA, HIP, or OpenMP. (e.g. sm\_35). If 'native' is used the compiler will detect locally installed architectures. For HIP offloading, the device architecture can be followed by target ID features delimited by a colon (e.g. gfx908:xnack+:sramecc-). May be specified more than once.

[**--offload-compression-level=**&lt;value&gt;](#offload~3)

Compression level for offload device binaries (HIP only)

[**--offload-compress**](#offload~4)

Compress offload device binaries (HIP only)

[**--offload-device-only**](#offload~5)

Only compile for the offloading device.

[**--offload-host-device**](#offload~6)

Compile for both the offloading host and device (default).

[**--offload-host-only**](#offload~7)

Only compile for the offloading host.

[**--offload-inc**](#offload~8)

Add include paths for CUDA/HIP and include the default CUDA/HIP wrapper headers (default)

[**--offload-jobs=**&lt;value&gt;](#offload~9)

Specify the number of threads to use for device offloading tasks during compilation. Can be a positive integer or the string 'jobserver' to use the make-style jobserver from the environment.

[**--offload-link**](#offload~10)

Use the new offloading linker to perform the link job.

[**--offload-new-driver**](#offload~11)

Use the new driver for offloading compilation.

[**--offload-targets=**&lt;value&gt;](#offload~12)

Specify a list of target architectures to use for offloading.

[**--offload=**&lt;value&gt;](#offload=)

Specify comma-separated list of offloading target triples (CUDA and HIP only)

[**--offloadlib**](#offloadlib)

Link device libraries for GPU device compilation

[**-o** &lt;file&gt;](#o)

Write output to &lt;file&gt;

[**-pedantic**](#pedantic)

Warn on language extensions

[**-pg**](#pg)

Enable mcount instrumentation

[**-pipe**](#pipe)

Use pipes between commands, when possible

[**--precompile-reduced-bmi**](#precompile)

Only precompile the input as a reduced BMI

[**--precompile**](#precompile~2)

Only precompile the input

[**--pretty-sgf**](#pretty)

Emit pretty printed symbol graphs

[**-print-diagnostic-options**](#print)

Print all of Clang's warning options

**-print-effective-triple** Print the effective target triple

[**-print-enabled-extensions**](#print~2)

Print the extensions enabled by the given target and **-march**/-mcpu options. (AArch64 and RISC-V only)

**-print-file-name=**&lt;file&gt; Print the full library path of &lt;file&gt;

[**-print-ivar-layout**](#print~3)

Enable Objective-C Ivar layout bitmap print trace

**-print-libgcc-file-name** Print the library path for the currently used compiler runtime library ("libgcc.a" or "libclang\_rt.builtins.\*.a")

[**-print-library-module-manifest-path**](#print~4)

Print the path for the C++ Standard library module manifest

[**-print-multi-flags-experimental**](#print~5)

Print the flags used for selecting multilibs (experimental)

**-print-prog-name=**&lt;name&gt; Print the full program path of &lt;name&gt;

[**-print-resource-dir**](#print~6)

Print the resource directory pathname

**-print-rocm-search-dirs** Print the paths used for finding ROCm installation

[**-print-runtime-dir**](#print~7)

Print the directory pathname containing Clang's runtime libraries

[**-print-search-dirs**](#print~8)

Print the paths used for finding libraries and programs

[**-print-supported-cpus**](#print~9)

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

[**-print-supported-extensions**](#print~10)

Print supported **-march** extensions (RISC-V, AArch64 and ARM only)

[**-print-target-triple**](#print~11)

Print the normalized target triple

[**-print-targets**](#print~12)

Print the registered targets

[**-pthread**](#pthread)

Support POSIX threads in generated code

[**--ptxas-path=**&lt;value&gt;](#ptxas)

Path to ptxas (used for compiling CUDA code)

[**-P**](#P)

Disable linemarker output in **-E** mode

[**-p**](#p)

Enable mcount instrumentation with prof

[**-Qn**](#Qn)

Do not emit metadata containing compiler name and version

[**-Qunused-arguments**](#Qunused)

Don't emit warning for unused driver arguments

[**-Qy**](#Qy)

Emit metadata containing compiler name and version

[**-regcall4**](#regcall4)

Set \_\_regcall4 as a default calling convention to respect \_\_regcall ABI v.4

[**-relocatable-pch**](#relocatable)

Whether to build a relocatable precompiled header

[**-resource-dir** &lt;value&gt;](#resource)

The directory which holds the compiler resource files

[**-rewrite-legacy-objc**](#rewrite)

Rewrite Legacy Objective-C source to C++

[**-rewrite-objc**](#rewrite~2)

Rewrite Objective-C source to C++

[**--rocm-device-lib-path=**&lt;value&gt;](#rocm)

ROCm device library path. Alternative to rocm-path.

[**--rocm-path=**&lt;value&gt;](#rocm~2)

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

**-Rpass-analysis=**&lt;value&gt; Report transformation analysis from optimization passes whose name matches the given POSIX regular expression

[**-Rpass-missed=**&lt;value&gt;](#Rpass)

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

[**-Rpass=**&lt;value&gt;](#Rpass=)

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

[**-rtlib=**&lt;value&gt;](#rtlib=)

Compiler runtime library to use

[**-R**&lt;remark&gt;](#R)

Enable the specified remark

[**-save-stats=**&lt;value&gt;](#save)

Save llvm statistics.

[**-save-stats**](#save~2)

Save llvm statistics.

[**-save-temps=**&lt;value&gt;](#save~3)

Save intermediate compilation results. &lt;arg&gt; can be set to 'cwd' for current working directory, or 'obj' which will save temporary files in the same directory as the final output file

[**-save-temps**](#save~4)

Alias for **--save-temps**=*cwd*

[**-serialize-diagnostics** &lt;value&gt;](#serialize)

Serialize compiler diagnostics to a file

[**-shared-libsan**](#shared)

Dynamically link the sanitizer runtime

[**--ssaf-compilation-unit-id=**&lt;id&gt;](#ssaf)

Stable identifier used as the CompilationUnit namespace name of every produced SSAF TU summary. Required when '--ssaf-tu-summary-file=' is set.

[**--ssaf-extract-summaries=**&lt;summary-names&gt;](#ssaf~2)

Comma-separated list of summary names to extract

[**--ssaf-include-local-entities**](#ssaf~3)

Include block-scope (function-local) declarations in extracted SSAF summaries. By default they are omitted.

[**--ssaf-list-extractors**](#ssaf~4)

Display the list of available SSAF summary extractors

[**--ssaf-list-formats**](#ssaf~5)

Display the list of available SSAF serialization formats

[**--ssaf-tu-summary-file=**&lt;path&gt;.&lt;format&gt;](#ssaf~6)

The output file for the extracted summaries. The extension selects which file format to use.

[**--start-no-unused-arguments**](#start)

Don't emit warnings about unused arguments for the following arguments

[**-static-libclosure**](#static)

Generate code for statically linking libclosure (BlocksRuntime)

[**-static-libsan**](#static~2)

Statically link the sanitizer runtime (Not supported for ASan, TSan or UBSan on darwin)

[**-static-openmp**](#static~3)

Use the static host OpenMP runtime while linking.

[**-std=**&lt;value&gt;](#std=)

Language standard to compile for

[**-stdlib**++-isystem &lt;directory&gt;](#stdlib)

Use directory as the C++ standard library include path

[**-stdlib=**&lt;value&gt;](#stdlib=)

C++ standard library to use

[**--sycl-link**](#sycl)

Perform link through clang-sycl-linker via the target offloading toolchain.

[**-sycl-std=**&lt;value&gt;](#sycl~2)

SYCL language standard to compile for.

[**--symbol-graph-dir=**&lt;value&gt;](#symbol)

Directory in which to emit symbol graphs.

[**--system-header-prefix=**&lt;prefix&gt;](#system)

Treat all #include paths starting with &lt;prefix&gt; as including a system header.

[**-S**](#S)

Only run preprocess and compilation steps

[**--target=**&lt;value&gt;](#target=)

Generate code for the given target

[**-time**](#time)

Time individual commands

[**-traditional-cpp**](#traditional)

Enable some traditional CPP emulation

[**-trigraphs**](#trigraphs)

Process trigraph sequences

[**-T** &lt;script&gt;](#T)

Specify &lt;script&gt; as linker script

[**-undef**](#undef)

undef all system defines

[**-unwindlib=**&lt;value&gt;](#unwindlib=)

Unwind library to use

[**-use-spirv-backend**](#use)

Use the SPIRV backend for compilation

[**-U** &lt;macro&gt;](#U)

Undefine macro &lt;macro&gt;

[**--verify-debug-info**](#verify)

Verify the binary representation of debug output

[**-verify-pch**](#verify~2)

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

[**--version**](#version)

Print version information

[**-vfsoverlay** &lt;value&gt;](#vfsoverlay)

Overlay the virtual filesystem described by file over the real file system. Additionally, pass this overlay file to the linker if it supports it

[**-v**](#v)

Show commands to run and use verbose output

[**-Wa**,&lt;arg&gt;](#Wa)

Pass the comma separated arguments in &lt;arg&gt; to the assembler

[**--warning-suppression-mappings=**&lt;value&gt;](#warning)

File containing diagnostic suppression mappings. See user manual for file format.

[**--wasm-opt**](#wasm)

Enable the wasm-opt optimizer (default)

[**-Wdeprecated**](#Wdeprecated)

Enable warnings for deprecated constructs and define \_\_DEPRECATED

[**-Winvalid-constexpr**](#Winvalid)

Enable checking of constexpr function bodies for validity within a constant expression context

[**-Wl**,&lt;arg&gt;](#Wl)

Pass the comma separated arguments in &lt;arg&gt; to the linker

[**-Wlarge-by-value-copy**](#Wlarge)

Warn if a function definition returns or accepts an object larger in bytes than a given value

[**-Wno-invalid-constexpr**](#Wno)

Disable checking of constexpr function bodies for validity within a constant expression context

[**-working-directory** &lt;value&gt;](#working)

Resolve file paths relative to the specified directory

[**-Wp**,&lt;arg&gt;](#Wp)

Pass the comma separated arguments in &lt;arg&gt; to the preprocessor

[**-Wsystem-headers-in-module=**&lt;module&gt;](#Wsystem)

Enable **-Wsystem-headers** when building &lt;module&gt;

[**-Wundef-prefix=**&lt;arg&gt;](#Wundef)

Enable warnings for undefined macros with a prefix in the comma separated list &lt;arg&gt;

[**-W**&lt;warning&gt;](#W)

Enable the specified warning

[**-w**](#w)

Suppress all warnings

[**-Xanalyzer** &lt;arg&gt;](#Xanalyzer)

Pass &lt;arg&gt; to the static analyzer

[**-Xarch\_device** &lt;arg&gt;](#Xarch_device)

Pass &lt;arg&gt; to device compilation in the offloading toolchain

[**-Xarch\_host** &lt;arg&gt;](#Xarch_host)

Pass &lt;arg&gt; to host compilation in the offloading toolchain

[**-Xarch\_**&lt;arch&gt; &lt;arg&gt;](#Xarch_)

Pass &lt;arg&gt; to the compilation if the target matches &lt;arch&gt;

[**-Xassembler** &lt;arg&gt;](#Xassembler)

Pass &lt;arg&gt; to the assembler

[**-Xclang=**&lt;arg&gt;](#Xclang=)

Alias for **-Xclang**

[**-Xclangas=**&lt;arg&gt;](#Xclangas=)

Alias for **-Xclangas**

[**-Xclangas** &lt;arg&gt;](#Xclangas)

Pass &lt;arg&gt; to clang **-cc1as**

[**-Xclang** &lt;arg&gt;](#Xclang)

Pass &lt;arg&gt; to clang **-cc1**

[**-Xcuda-fatbinary** &lt;arg&gt;](#Xcuda)

Pass &lt;arg&gt; to fatbinary invocation

[**-Xcuda-ptxas** &lt;arg&gt;](#Xcuda~2)

Pass &lt;arg&gt; to the ptxas assembler

[**-Xlinker** &lt;arg&gt;](#Xlinker)

Pass &lt;arg&gt; to the linker

[**-Xoffload-compiler**&lt;triple&gt; &lt;arg&gt;](#Xoffload)

Pass &lt;arg&gt; to the offload compilers or the ones identified by -&lt;triple&gt;

[**-Xoffload-linker**&lt;triple&gt; &lt;arg&gt;](#Xoffload~2)

Pass &lt;arg&gt; to the offload linkers or the ones identified by -&lt;triple&gt;

[**-Xopenmp-target=**&lt;triple&gt; &lt;arg&gt;](#Xopenmp)

Pass &lt;arg&gt; to the target offloading toolchain identified by &lt;triple&gt;.

[**-Xopenmp-target** &lt;arg&gt;](#Xopenmp~2)

Pass &lt;arg&gt; to the target offloading toolchain.

[**-Xpreprocessor** &lt;arg&gt;](#Xpreprocessor)

Pass &lt;arg&gt; to the preprocessor

[**-Xthinlto-distributor=**&lt;arg&gt;](#Xthinlto)

Pass &lt;arg&gt; to the ThinLTO distributor process. Can be specified multiple times or with comma-separated values.

[**-x** &lt;language&gt;](#x)

Treat subsequent input files as having type &lt;language&gt;

[**-z** &lt;arg&gt;](#z)

Pass **-z** &lt;arg&gt; to the linker

## [dxc compatibility options.:](#dxc_compatibility_options.:)

[**--dxv-path=**&lt;value&gt;](#dxv)

DXIL validator installation path

[**-fdx-rootsignature-define=**&lt;value&gt;](#fdx)

Override entry function root signature with root signature at given macro name.

[**-fdx-rootsignature-version=**&lt;value&gt;](#fdx~2)

Root Signature Version

**-fspv-extension=**&lt;value&gt; Specify the available SPIR-V extensions. If this option is not specified, then all extensions are available. If KHR is specified, then all KHR extensions will be available. If DXC is specifided, then all extensions implemented by the DirectX Shader compiler will be available. This option is useful for moving from DXC to Clang.

[**-fspv-target-env=**&lt;value&gt;](#fspv)

Specify the target environment

*/hlsl-all-resources-bound*

Enables agressive flattening

[**-hlsl-entry** &lt;value&gt;](#hlsl)

Entry point name for hlsl

*/res-may-alias*

Assume that UAVs/SRVs may alias

[**--spirv-val-path=**&lt;value&gt;](#spirv)

SPIRV validator installation path

*/validator-version* &lt;value&gt;

Override validator version for module. Format: &lt;major.minor&gt;;Default: DXIL.dll version or current internal version

October 2026 Wolfi

---

Document: clang(1) from clang-23 23.1.4-r1 (Chainguard OS)