ccache(1) ccache(1)
NAME
ccache - a fast C/C++ compiler cache
SYNOPSIS
ccache [ccache options]
ccache [KEY=VALUE ...] compiler [compiler options]
compiler [compiler options]
The first form takes options described in COMMAND LINE OPTIONS below.
The second form invokes the compiler, optionally using configuration
options as KEY=VALUE arguments. In the third form, ccache is
masquerading as the compiler as described in RUN MODES.
DESCRIPTION
Ccache is a compiler cache. It speeds up recompilation by caching the
result of previous compilations and detecting when the same compilation
is being done again.
Ccache has been carefully written to always produce exactly the same
compiler output that you would get without the cache. The only way you
should be able to tell that you are using ccache is the speed.
Currently known exceptions to this goal are listed under CAVEATS. If
you discover an undocumented case where ccache changes the output of
your compiler, please let us know.
RUN MODES
There are two different ways to use ccache to cache a compilation:
1. Prefix your compilation command with ccache. This method is most
convenient if you just want to try out ccache or wish to use it for
some specific projects. Example:
ccache gcc -c example.c
2. Let ccache masquerade as the compiler. This method is most useful
when you wish to use ccache for all your compilations. To do this,
create a symbolic link to ccache named as the compiler. For
example, here is set up ccache to masquerade as gcc and g++:
cp ccache /usr/local/bin/
ln -s ccache /usr/local/bin/gcc
ln -s ccache /usr/local/bin/g++
On platforms that don't support symbolic links you can simply copy
ccache to the compiler name instead for a similar effect:
cp ccache /usr/local/bin/gcc
cp ccache /usr/local/bin/g++
And so forth. This will work as long as the directory with symbolic
links or ccache copies comes before the directory with the compiler
(typically /usr/bin) in PATH.
Warning
The technique of letting ccache masquerade as the compiler
works well,
but currently doesn't interact well with other tools that do
the same thing. See USING CCACHE WITH OTHER COMPILER WRAPPERS.
COMMAND LINE OPTIONS
These command line options only apply when you invoke ccache as
"ccache". When ccache masquerades as a compiler (as described in the
previous section), the normal compiler options apply and you should
refer to the compiler's documentation.
Common options
-c, --cleanup
Clean up the cache by removing not recently used cached files until
the specified file number and cache size limits are not exceeded.
This also recalculates the cache file count and size totals.
Normally, there is no need to initiate cleanup manually as ccache
keeps the cache below the specified limits at runtime and keeps
statistics up to date on each compilation. Forcing a cleanup is
mostly useful if you have modified the cache contents manually or
believe that the cache size statistics may be inaccurate.
-C, --clear
Clear the entire cache, removing all cached files, but keeping the
configuration file.
--config-path PATH
Let the command line options operate on configuration file PATH
instead of the default. Using this option has the same effect as
setting (overriding) the environment variable CCACHE_CONFIGPATH
temporarily.
-d, --dir PATH
Let the command line options operate on cache directory PATH
instead of the default. For example, to show statistics for a cache
directory at /shared/ccache you can run ccache -d /shared/ccache
-s. Using this option has the same effect as setting the
environment variable CCACHE_DIR temporarily.
--evict-namespace NAMESPACE
Remove files created in the given namespace from the cache.
--evict-older-than AGE
Remove files used less recently than AGE from the cache. AGE should
be an unsigned integer with a d (days) or s (seconds) suffix. If
combined with --evict-namespace, only remove files within that
namespace.
-h, --help
Print a summary of command line options.
-F NUM, --max-files NUM
Set the maximum number of files allowed in the cache to NUM. Use 0
for no limit. The value is stored in a configuration file in the
cache directory and applies to all future compilations.
-M SIZE, --max-size SIZE
Set the maximum size of the files stored in the cache. SIZE should
be a number followed by an optional suffix: kB, MB, GB, TB
(decimal), KiB, MiB, GiB or TiB (binary). The default suffix is
GiB. Use 0 for no limit. The value is stored in a configuration
file in the cache directory and applies to all future compilations.
-X LEVEL, --recompress LEVEL
Recompress the cache to level LEVEL using the Zstandard algorithm.
The level can be an integer, with the same semantics as the
compression_level configuration option, or the special value
uncompressed for no compression. See [Cache compression] for more
information. This can potentially take a long time since all files
in the cache need to be visited. Only files that are currently
compressed with a different level than LEVEL will be recompressed.
--recompress-threads THREADS
Use up to THREADS threads when recompressing the cache. The default
is to use one thread per CPU.
-o KEY=VALUE, --set-config KEY=VALUE
Set configuration option KEY to VALUE in the configuration file.
See CONFIGURATION for more information.
-x, --show-compression
Print cache compression statistics. See CACHE COMPRESSION for more
information. This can potentially take a long time since all files
in the cache need to be visited.
-p, --show-config
Print current configuration options and from where they originate
(environment variable, configuration file or compile-time default)
in human-readable format.
--show-log-stats
Print statistics counters from the stats log in human-readable
format. See stats_log. Use -v/--verbose once or twice for more
details.
-s, --show-stats
Print a summary of configuration and statistics counters in
human-readable format. Use -v/--verbose once or twice for more
details.
-v, --verbose
Increase verbosity. The option can be given multiple times.
-V, --version
Print version and copyright information.
-z, --zero-stats
Zero the cache statistics (but not the configuration options).
Options for remote file-based storage
--trim-dir PATH
Remove not recently used files from directory PATH until it is at
most the size specified by --trim-max-size.
Warning
Don't use this option to trim the local cache. To trim the
local cache
directory to a certain size, use CCACHE_MAXSIZE=SIZE ccache -c.
--trim-max-size SIZE
Specify the maximum size for --trim-dir. SIZE should be a number
followed by an optional suffix: kB, MB, GB, TB (decimal), KiB, MiB,
GiB or TiB (binary). The default suffix is GiB. Use 0 for no limit.
--trim-method METHOD
Specify the method to trim a directory with --trim-dir. Possible
values are:
atime
LRU (least recently used) using the file access timestamp. This
is the default.
mtime
LRU (least recently used) using the file modification
timestamp.
--trim-recompress LEVEL
Recompress to level LEVEL using the Zstandard algorithm when using
--trim-dir. The level can be an integer, with the same semantics as
the compression_level configuration option, or the special value
uncompressed for no compression. See [Cache compression] for more
information. This can potentially take a long time since all files
in the cache need to be visited. Only files that are currently
compressed with a different level than LEVEL will be recompressed.
--trim-recompress-threads THREADS
Recompress using up to THREADS threads with --trim-recompress. The
default is to use one thread per CPU.
Options for scripting or debugging
--checksum-file PATH
Print the checksum (128 bit XXH3) of the file at PATH (- for
standard input).
--extract-result PATH
Extract data stored in the result file at PATH (- for standard
input). The data will be written to ccache-result.* files in to
the current working directory. This option is only useful when
debugging ccache and its behavior.
--format FORMAT
Specify format for --print-log-stats and --print-stats. Possible
values are:
tab
Tab separated. This is the default.
json
JSON formatted.
-k KEY, --get-config KEY
Print the value of configuration option KEY. See CONFIGURATION for
more information.
--hash-file PATH
Print the hash (160 bit BLAKE3) of the file at PATH (- for standard
input). This is only useful when debugging ccache and its behavior.
--inspect PATH
Print the content of a result or manifest file at PATH (- for
standard input) to standard output in human-readable format. File
content embedded in a result file will however not be printed; use
--extract-result to extract the file content. This option is only
useful when debugging ccache and its behavior.
--print-log-stats
Print statistics counters from the stats log in machine-parsable
(tab-separated or JSON) format. See stats_log and --format.
--print-stats
Print statistics counter IDs and corresponding values in
machine-parsable (tab-separated or JSON) format. See --format.
--print-version
Print version and don't do anything else.
Extra options
When run as a compiler, ccache usually just takes the same command line
options as the compiler you are using. The only exception to this is
the option --ccache-skip. That option can be used to tell ccache to
avoid interpreting the next option in any way and to pass it along to
the compiler as-is.
Note
--ccache-skip currently only tells ccache not to interpret the next
option as a special compiler option -- the option will still be
included in the direct mode hash.
The reason this can be important is that ccache does need to parse the
command line and determine what is an input filename and what is a
compiler option, as it needs the input filename to determine the name
of the resulting object file (among other things). The heuristic ccache
uses when parsing the command line is that any argument that exists as
a file is treated as an input file name. By using --ccache-skip you can
force an option to not be treated as an input file name and instead be
passed along to the compiler as a command line option.
Another case where --ccache-skip can be useful is if ccache interprets
an option specially but shouldn't, since the option has another meaning
for your compiler than what ccache thinks.
See also ignore_options.
CONFIGURATION
Ccache's default behavior can be overridden by options in configuration
files, which in turn can be overridden by environment variables with
names starting with CCACHE_. Ccache normally reads configuration from
two files: first a system-level configuration file and secondly a
cache-specific configuration file. The priorities of configuration
options are as follows (where 1 is highest):
1. Command line settings in KEY=VALUE form. Example:
ccache debug=true compiler_check="%compiler% --version" gcc -c example.c
2. Environment variables.
3. The cache-specific configuration file (see below).
4. The system (read-only) configuration file <sysconfdir>/ccache.conf
(typically /etc/ccache.conf or /usr/local/etc/ccache.conf).
5. Compile-time defaults.
As a special case, if the environment variable CCACHE_CONFIGPATH is set
it specifies the configuration file, and the system configuration file
won't be read.
Location of the configuration file
The location of the cache-specific configuration file is determined
like this on non-Windows systems:
1. If CCACHE_CONFIGPATH is set, use that path.
2. Otherwise, if the environment variable CCACHE_DIR is set then use
$CCACHE_DIR/ccache.conf.
3. Otherwise, if cache_dir is set in the system configuration file
then use <cache_dir>/ccache.conf.
4. Otherwise, if there is a legacy $HOME/.ccache directory then use
$HOME/.ccache/ccache.conf.
5. Otherwise, if XDG_CONFIG_HOME is set then use
$XDG_CONFIG_HOME/ccache/ccache.conf.
6. Otherwise, use $HOME/Library/Preferences/ccache/ccache.conf (macOS)
or $HOME/.config/ccache/ccache.conf (other systems).
On Windows, this is the method used to find the configuration file:
1. If CCACHE_CONFIGPATH is set, use that path.
2. Otherwise, if the environment variable CCACHE_DIR is set then use
%CCACHE_DIR%/ccache.conf.
3. Otherwise, if cache_dir is set in the system configuration file
then use <cache_dir>\ccache.conf. The system-wide configuration on
Windows is %ALLUSERSPROFILE%\ccache\ccache.conf by default. The
ALLUSERSPROFILE environment variable is usually C:\ProgramData.
4. Otherwise, if there is a legacy %USERPROFILE%\.ccache directory
then use %USERPROFILE%\.ccache\ccache.conf.
5. Otherwise, use %LOCALAPPDATA%\ccache\ccache.conf if it exists.
6. Otherwise, use %APPDATA%\ccache\ccache.conf.
See also the cache_dir configuration option for how the cache directory
location is determined.
Configuration value syntax
All configuration values support expansion of environment variables.
The syntax is similar to POSIX shell syntax: $VAR or ${VAR}. Both
variants will expand to the value of the environment variable VAR.
Two consecutive dollar signs ($$) will expand to a single dollar sign
($).
Configuration file syntax
Configuration files are in a simple "key = value" format, one option
per line. Lines starting with a hash sign are comments. Blank lines are
ignored, as is whitespace surrounding keys and values. Example:
# Set maximum cache size to 10 GB:
max_size = 10G
Boolean values
Some configuration options are boolean values (i.e. truth values). In a
configuration file, such values must be set to the string true or
false. For the corresponding environment variables, the semantics are
a bit different:
o A set environment variable means "true" (even if set to the empty
string).
o The following case-insensitive negative values are considered an
error (instead of surprising the user): 0, false, disable and no.
o An unset environment variable means "false".
Each boolean environment variable also has a negated form starting with
CCACHE_NO. For example, CCACHE_COMPRESS can be set to force compression
and CCACHE_NOCOMPRESS can be set to force no compression.
Configuration options
Below is a list of available configuration options. The corresponding
environment variable name is indicated in parentheses after each
configuration option key.
absolute_paths_in_stderr (CCACHE_ABSSTDERR)
This option specifies whether ccache should rewrite relative paths
in the compiler's textual output (standard error and standard
output) to absolute paths. This can be useful if you use base_dir
with a build system (e.g. CMake with the "Unix Makefiles"
generator) that executes the compiler in a different working
directory, which makes relative paths in compiler errors or
warnings incorrect. The default is false.
base_dir (CCACHE_BASEDIR)
This option should be an absolute path to a directory. If set,
ccache will rewrite absolute paths into paths relative to the
current working directory, but only absolute paths that begin with
base_dir. Cache results can then be shared for compilations in
different directories even if the project uses absolute paths in
the compiler command line. See also the discussion under COMPILING
IN DIFFERENT DIRECTORIES. If set to the empty string (which is the
default), no rewriting is done.
A typical path to use as base_dir is your home directory or another
directory that is a parent of your project directories. Don't use /
as the base directory since that will make ccache also rewrite
paths to system header files, which typically is counterproductive.
For example, say that Alice's current working directory is
/home/alice/project1/build and that she compiles like this:
ccache gcc -I/usr/include/example -I/home/alice/project2/include -c /home/alice/project1/src/example.c
Here is what ccache will actually execute for different base_dir
values:
# Current working directory: /home/alice/project1/build
# With base_dir = /:
gcc -I../../../../usr/include/example -I../../project2/include -c ../src/example.c
# With base_dir = /home or /home/alice:
gcc -I/usr/include/example -I../../project2/include -c ../src/example.c
# With base_dir = /home/alice/project1 or /home/alice/project1/src:
gcc -I/usr/include/example -I/home/alice/project2/include -c ../src/example.c
If Bob has put project1 and project2 in /home/bob/stuff and both
users have set base_dir to /home or /home/$USER, then Bob will get
a cache hit (if they share ccache directory) since the actual
command line will be identical to Alice's command line:
# Current working directory: /home/bob/stuff/project1/build
# With base_dir = /home or /home/bob:
gcc -I/usr/include/example -I../../project2/include -c ../src/example.c
Without base_dir there will be a cache miss since the absolute
paths will differ. With base_dir set to / there will be a cache
miss since the relative path to /usr/include/example will be
different. With base_dir set to /home/bob/stuff/project1 there will
a cache miss since the path to project2 will be a different
absolute path.
Warning
Rewriting absolute paths to relative is kind of a brittle hack.
It
works OK in many cases, but there might be cases where things
break. One known issue is that absolute paths are not
reproduced in dependency files, which can mess up dependency
detection in tools like Make and Ninja. If possible, use
relative paths in the first place instead of using base_dir.
cache_dir (CCACHE_DIR)
This option specifies where ccache will keep its cached compiler
outputs.
On non-Windows systems, the default is $HOME/.ccache if such a
directory exists, otherwise $XDG_CACHE_HOME/ccache if
XDG_CACHE_HOME is set, otherwise $HOME/Library/Caches/ccache
(macOS) or $HOME/.cache/ccache (other systems).
On Windows, the default is %USERPROFILE%\.ccache if such a
directory exists, otherwise %LOCALAPPDATA%\ccache.
Warning
Previous ccache versions defaulted to storing the cache in
%APPDATA%\ccache on Windows. This can result in large network
file transfers of the cache in domain environments and similar
problems. Please check this directory for cache directories and
either delete them or the whole directory, or move them to the
%LOCALAPPDATA%\ccache directory.
See also Location of the configuration file.
compiler (CCACHE_COMPILER or (deprecated) CCACHE_CC)
This option can be used to force the name of the compiler to use.
If set to the empty string (which is the default), ccache works it
out from the command line.
compiler_check (CCACHE_COMPILERCHECK)
By default, ccache includes the modification time ("mtime") and
size of the compiler in the hash to ensure that results retrieved
from the cache are accurate. If compiler plugins are used, these
plugins will also be added to the hash. This option can be used to
select another strategy. Possible values are:
content
Hash the content of the compiler binary. This makes ccache very
slightly slower compared to mtime, but makes it cope better
with compiler upgrades during a build bootstrapping process.
mtime
Hash the compiler's mtime and size, which is fast. This is the
default.
none
Don't hash anything. This may be good for situations where you
can safely use the cached results even though the compiler's
mtime or size has changed (e.g. if the compiler is built as
part of your build system and the compiler's source has not
changed, or if the compiler only has changes that don't affect
code generation). You should only use none if you know what you
are doing.
string:value
Hash value. This can for instance be a compiler revision number
or another string that the build system generates to identify
the compiler.
a command string
Hash the standard output and standard error output of the
specified command. The string will be split on whitespace to
find out the command and arguments to run. No other
interpretation of the command string will be done, except that
the special word %compiler% will be replaced with the path to
the compiler. Several commands can be specified with semicolon
as separator. Examples:
%compiler% -v
%compiler% -dumpmachine; %compiler% -dumpversion
You should make sure that the specified command is as fast as
possible since it will be run once for each ccache invocation.
Identifying the compiler using a command is useful if you want
to avoid cache misses when the compiler has been rebuilt but
not changed.
Another case is when the compiler (as seen by ccache) actually
isn't the real compiler but another compiler wrapper -- in that
case, the default mtime method will hash the mtime and size of
the other compiler wrapper, which means that ccache won't be
able to detect a compiler upgrade. Using a suitable command to
identify the compiler is thus safer, but it's also slower, so
you should consider continue using the mtime method in
combination with the prefix_command option if possible. See
USING CCACHE WITH OTHER COMPILER WRAPPERS.
compiler_type (CCACHE_COMPILERTYPE)
Ccache normally guesses the compiler type based on the compiler
name. The compiler_type option lets you force a compiler type. This
can be useful if the compiler has a non-standard name but is
actually one of the known compiler types. Possible values are:
auto
Guess one of the types below based on the compiler name
(following symlinks). This is the default.
clang
Clang-based compiler.
clang-cl
clang-cl.
gcc
GCC-based compiler.
icl
Intel compiler on Windows.
icx
Intel LLVM-based compiler.
icx-cl
Intel LLVM-based MSVC-compatible compiler.
msvc
Microsoft Visual C++ (MSVC).
nvcc
NVCC (CUDA) compiler.
other
Any compiler other than the known types.
compression (CCACHE_COMPRESS or CCACHE_NOCOMPRESS, see Boolean values
above)
If true, ccache will compress data it puts in the cache. However,
this option has no effect on how files are retrieved from the
cache; compressed and uncompressed results will still be usable
regardless of this option. The default is true.
Compression is done using the Zstandard algorithm. The algorithm is
fast enough that there should be little reason to turn off
compression to gain performance. One exception is if the cache is
located on a compressed file system, in which case the compression
performed by ccache of course is redundant.
Compression will be disabled if file cloning (the file_clone
option) or hard linking (the hard_link option) is enabled.
compression_level (CCACHE_COMPRESSLEVEL)
This option determines the level at which ccache will compress
object files using the real-time compression algorithm Zstandard.
It only has effect if compression is enabled (which it is by
default). Zstandard is extremely fast for decompression and very
fast for compression for lower compression levels. The default is
0.
Semantics of compression_level:
> 0
A positive value corresponds to normal Zstandard compression
levels. Lower levels (e.g. 1) mean faster compression but worse
compression ratio. Higher levels (e.g. 19) mean slower
compression but better compression ratio. The maximum possible
value depends on the libzstd version, but at least up to 19 is
available for all versions. Decompression speed is essentially
the same for all levels. As a rule of thumb, use level 5 or
lower since higher levels may slow down compilations
noticeably. Higher levels are however useful when recompressing
the cache with command line option -X/--recompress.
< 0
A negative value corresponds to Zstandard's "ultra-fast"
compression levels, which are even faster than level 1 but with
less good compression ratios. For instance, level -3
corresponds to --fast=3 for the zstd command line tool. In
practice, there is little use for levels lower than -5 or so.
0 (default)
The value 0 means that ccache will choose a suitable level,
currently 1.
See the Zstandard documentation <http://zstd.net> for more
information.
cpp_extension (CCACHE_EXTENSION)
This option can be used to force a certain extension for the
intermediate preprocessed file. The default is to automatically
determine the extension to use for intermediate preprocessor files
based on the type of file being compiled, but that sometimes
doesn't work. For example, when using the "aCC" compiler on HP-UX,
set the cpp extension to i.
debug (CCACHE_DEBUG or CCACHE_NODEBUG, see Boolean values above)
If true, enable the debug mode. The debug mode creates per-object
debug files that are helpful when debugging unexpected cache
misses. Note however that ccache performance will be reduced
slightly. See CACHE DEBUGGING for more information. The default is
false.
debug_dir (CCACHE_DEBUGDIR)
Specifies where to write per-object debug files if the debug mode
is enabled. If set to the empty string, the files will be written
next to the object file. If set to a directory, the debug files
will be written with full absolute paths in that directory,
creating it if needed. The default is the empty string.
For example, if debug_dir is set to /example, the current working
directory is /home/user and the object file is build/output.o then
the debug log will be written to
/example/home/user/build/output.o.ccache-log. See also CACHE
DEBUGGING.
debug_level (CCACHE_DEBUGLEVEL)
Specifies the amount of information that is written when the debug
mode is enabled. See CACHE DEBUGGING for more information. The
default is 2.
depend_mode (CCACHE_DEPEND or CCACHE_NODEPEND, see Boolean values
above)
If true, the depend mode will be used. The default is false. See
The depend mode.
direct_mode (CCACHE_DIRECT or CCACHE_NODIRECT, see Boolean values
above)
If true, the direct mode will be used. The default is true. See The
direct mode.
disable (CCACHE_DISABLE or CCACHE_NODISABLE, see Boolean values above)
When true, ccache will just call the real compiler, bypassing the
cache completely. The default is false.
It is also possible to disable ccache for a specific source code
file by adding the string ccache:disable in a comment in the first
4096 bytes of the file.
extra_files_to_hash (CCACHE_EXTRAFILES)
This option is a list of paths to files that ccache will include in
the the hash sum that identifies the build. The list separator is
semicolon on Windows systems and colon on other systems.
file_clone (CCACHE_FILECLONE or CCACHE_NOFILECLONE, see Boolean values
above)
If true, ccache will attempt to use file cloning (also known as
"copy on write", "CoW" or "reflinks") to store and fetch cached
compiler results. file_clone has priority over hard_link. The
default is false.
Files stored by cloning cannot be compressed, so the cache size
will likely be significantly larger if this option is enabled.
However, performance may be improved depending on the use case.
Unlike the hard_link option, file_clone is completely safe to use,
but not all file systems support the feature. For such file
systems, ccache will fall back to use plain copying (or hard links
if hard_link is enabled).
hard_link (CCACHE_HARDLINK or CCACHE_NOHARDLINK, see Boolean values
above)
If true, ccache will attempt to use hard links to store and fetch
cached object files. The default is false.
Files stored via hard links cannot be compressed, so the cache size
will likely be significantly larger if this option is enabled.
However, performance may be improved depending on the use case.
Warning
Do not enable this option unless you are aware of these
caveats:
o If the resulting file is modified, the file in the cache will
also be modified since they share content, which corrupts the
cache entry. As of version 4.0, ccache makes stored and fetched
object files read-only as a safety measure. Furthermore, a
simple integrity check is made for cached object files by
verifying that their sizes are correct. This means that
mistakes like strip file.o or echo >file.o will be detected
even if the object file is made writable, but a modification
that doesn't change the file size will not.
o Programs that don't expect that files from two different
identical compilations are hard links to each other can fail.
o Programs that rely on modification times (like make) can be
confused if several users (or one user with several build
trees) use the same cache directory. The reason for this is
that the object files share i-nodes and therefore modification
times. If file.o is in build tree A (hard-linked from the
cache) and file.o then is produced by ccache in build tree B by
hard-linking from the cache, the modification timestamp will be
updated for file.o in build tree A as well. This can retrigger
relinking in build tree A even though nothing really has
changed.
hash_dir (CCACHE_HASHDIR or CCACHE_NOHASHDIR, see Boolean values above)
If true (which is the default), ccache will include the current
working directory (CWD) in the hash that is used to distinguish two
compilations when generating debug info (compiler option -g with
variations). Exception: The CWD will not be included in the hash
if the compiler options -fdebug-prefix-map or
-fdebug-compilation-dir are used appropriately. See also the
discussion under COMPILING IN DIFFERENT DIRECTORIES.
The reason for including the CWD in the hash by default is to
prevent a problem with the storage of the current working directory
in the debug info of an object file, which can lead ccache to
return a cached object file that has the working directory in the
debug info set incorrectly.
You can disable this option to get cache hits when compiling the
same source code in different directories if you don't mind that
CWD in the debug info might be incorrect.
ignore_headers_in_manifest (CCACHE_IGNOREHEADERS)
This option is a list of paths to files (or directories with
headers) that ccache will not include in the manifest list that
makes up the direct mode. Note that this can cause stale cache hits
if those headers do indeed change. The list separator is semicolon
on Windows systems and colon on other systems.
ignore_options (CCACHE_IGNOREOPTIONS)
This option is a space-delimited list of compiler options that
ccache will ignore. Entries in the list can optionally end with an
asterisk (*) to matching any option suffix. For example,
-fmessage-length=* will match both -fmessage-length=20 and
-fmessage-length=70. A matching compiler option will neither be
interpreted specially nor be part of the input hash. Ignoring a
compiler option from the hash can be useful when you know it
doesn't affect the result (and ccache doesn't know that), or when
it does and you don't care. See also Extra options.
inode_cache (CCACHE_INODECACHE or CCACHE_NOINODECACHE, see Boolean
values above)
If true, ccache will cache source file hashes based on device,
inode and timestamps. This reduces the time spent on hashing
include files since the result can be reused between compilations.
The default is true. The feature requires temporary_dir to be
located on a local filesystem of a supported type.
Note
Support for the inode cache feature on Windows is experimental.
On Windows
the default is false.
keep_comments_cpp (CCACHE_COMMENTS or CCACHE_NOCOMMENTS, see Boolean
values above)
If true, ccache will not discard the comments before hashing
preprocessor output. The default is false. This can be used to
check documentation with -Wdocumentation.
log_file (CCACHE_LOGFILE)
If set to a file path, ccache will write information on what it is
doing to the specified file. This is useful for tracking down
problems.
If set to syslog, ccache will log using syslog() instead of to a
file. If you use rsyslogd, you can add something like this to
/etc/rsyslog.conf or a file in /etc/rsyslog.d:
# log ccache to file
:programname, isequal, "ccache" /var/log/ccache
# remove from syslog
& ~
max_files (CCACHE_MAXFILES)
This option specifies the maximum number of files to keep in the
cache. Use 0 for no limit (which is the default). See also CACHE
SIZE MANAGEMENT.
max_size (CCACHE_MAXSIZE)
This option specifies the maximum size of the cache. Use 0 for no
limit. The default value is 5G. Available suffixes: k, M, G, T
(decimal) and Ki, Mi, Gi, Ti (binary). The default suffix is G. See
also CACHE SIZE MANAGEMENT.
msvc_dep_prefix (CCACHE_MSVC_DEP_PREFIX)
This option specifies the prefix of included files output for MSVC
compiler. The default prefix is "Note: including file:". If you
use a localized compiler, this should be set accordingly.
namespace (CCACHE_NAMESPACE)
If set, the namespace string will be added to the hashed data for
each compilation. This will make the associated cache entries
logically separate from cache entries with other namespaces, but
they will still share the same storage space. Cache entries can
also be selectively removed from the local cache with the command
line option --evict-namespace, potentially in combination with
--evict-older-than.
For instance, if you use the same local cache for several disparate
projects, you can use a unique namespace string for each one. This
allows you to remove cache entries that belong to a certain project
if you stop working with that project.
path (CCACHE_PATH)
If set, ccache will search directories in this list when looking
for the real compiler. The list separator is semicolon on Windows
systems and colon on other systems. If not set, ccache will look
for the first executable matching the compiler name in the normal
PATH that isn't a symbolic link to ccache itself.
pch_external_checksum (CCACHE_PCH_EXTSUM or CCACHE_NOPCH_EXTSUM, see
Boolean values above)
When this option is set, and ccache finds a precompiled header
file, ccache will look for a file with the extension ".sum" added
(e.g. "pre.h.gch.sum"), and if found, it will hash this file
instead of the precompiled header itself to work around the
performance penalty of hashing very large files.
prefix_command (CCACHE_PREFIX)
This option adds a list of prefixes (separated by space) to the
command line that ccache uses when invoking the compiler. See also
USING CCACHE WITH OTHER COMPILER WRAPPERS.
prefix_command_cpp (CCACHE_PREFIX_CPP)
This option adds a list of prefixes (separated by space) to the
command line that ccache uses when invoking the preprocessor.
read_only (CCACHE_READONLY or CCACHE_NOREADONLY, see Boolean values
above)
If true, ccache will attempt to use existing cached results, but it
will not add new results to any cache backend. Statistics counters
will still be updated, though, unless the stats option is set to
false.
If you are using this because your ccache directory is read-only,
you need to set temporary_dir since ccache will fail to create
temporary files otherwise. You may also want to set stats to false
make ccache not even try to update stats files.
read_only_direct (CCACHE_READONLY_DIRECT or CCACHE_NOREADONLY_DIRECT,
see Boolean values above)
Just like read_only except that ccache will only try to retrieve
results from the cache using the direct mode, not the preprocessor
mode. See documentation for read_only regarding using a read-only
ccache directory.
recache (CCACHE_RECACHE or CCACHE_NORECACHE, see Boolean values above)
If true, ccache will not use any previously stored result. New
results will still be cached, possibly overwriting any pre-existing
results.
remote_only (CCACHE_REMOTE_ONLY or CCACHE_NOREMOTE_ONLY, see Boolean
values above)
If true, ccache will only use remote storage. The default is false.
Note that cache statistics counters will still be kept in the local
cache directory unless stats is false. See also Storage
interaction.
remote_storage (CCACHE_REMOTE_STORAGE)
This option specifies one or several storage backends (separated by
space) to query after checking the local cache (unless remote_only
is true). See REMOTE STORAGE BACKENDS for documentation of syntax
and available backends.
Examples:
o file:/shared/nfs/directory
o file:///shared/nfs/one|read-only file:///shared/nfs/two
o file:///Z:/example/windows/folder
o http://example.com/cache
o redis://example.com
Note
In previous ccache versions this option was called
secondary_storage
(CCACHE_SECONDARY_STORAGE), which can still be used as an
alias.
reshare (CCACHE_RESHARE or CCACHE_NORESHARE, see Boolean values above)
If true, ccache will write results to remote storage even for local
storage cache hits. The default is false.
response_file_format (CCACHE_RESPONSE_FILE_FORMAT)
Ccache normally guesses the response file format based on the
compiler type. The response_file_format option lets you force the
response file quoting behavior. This can be useful if the compiler
supports both POSIX and Windows response file quoting. Possible
values are:
auto
Guess one of the formats below based on the compiler type. This
is the default.
posix
POSIX quoting behavior.
windows
Windows quoting behavior.
run_second_cpp (CCACHE_CPP2 or CCACHE_NOCPP2, see Boolean values above)
If true, ccache will first run the preprocessor to preprocess the
source code (see The preprocessor mode) and then on a cache miss
run the compiler on the source code to get hold of the object file.
This is the default.
If false, ccache will first run preprocessor to preprocess the
source code and then on a cache miss run the compiler on the
preprocessed source code instead of the original source code. This
makes cache misses slightly faster since the source code only has
to be preprocessed once. The downside is that some compilers won't
produce the same result (for instance diagnostics warnings) when
compiling preprocessed source code.
A solution to the above mentioned downside is to set run_second_cpp
to false and pass -fdirectives-only (for GCC) or -frewrite-includes
(for Clang) to the compiler. This will cause the compiler to leave
the macros and other preprocessor information, and only process the
#include directives. When run in this way, the preprocessor
arguments will be passed to the compiler since it still has to do
some preprocessing (like macros).
This option is ignored with MSVC, as there is no way to make it
compile without preprocessing first.
sloppiness (CCACHE_SLOPPINESS)
By default, ccache tries to give as few false cache hits as
possible. However, in certain situations it's possible that you
know things that ccache can't take for granted. This option makes
it possible to tell ccache to relax some checks in order to
increase the hit rate. The value should be a comma-separated string
with one or several of the following values:
clang_index_store
Ignore the Clang compiler option -index-store-path and its
argument when computing the manifest hash. This is useful if
you use Xcode, which uses an index store path derived from the
local project path. Note that the index store won't be updated
correctly on cache hits if you enable this sloppiness.
file_stat_matches
Ccache normally examines a file's contents to determine whether
it matches the cached version. With this sloppiness set, ccache
will consider a file as matching its cached version if the
mtimes and ctimes match.
file_stat_matches_ctime
Ignore ctimes when file_stat_matches is enabled. This can be
useful when backdating files' mtimes in a controlled way.
gcno_cwd
By default, ccache will include the current working directory
in the hash when producing a .gcno file (when compiling with
-ftest-coverage). This is because GCC 9+ includes the current
working directory in the .gcno file. The gcno_cwd sloppiness
makes ccache not hash the current working directory so that you
can get cache hits when compiling in different directories,
with the tradeoff of potentially getting an incorrect directory
in the .gcno file. gcno_cwd also disables hashing of the
current working directory if -fprofile-abs-path is used.
Note
gcno_cwd sloppiness will not have any effect when compiling
with
--coverage since that implies -fprofile-arcs which always
forces CWD to be included in the input hash.
incbin
By default, ccache will ignore all files containing an .incbin
directive. While this is the correct behaviour as ccache does
not detect incbin changes, this restriction can make some
projects difficult to cache. This sloppiness will pretend the
.incbin directive doesn't exist and simply allow caching.
include_file_ctime
By default, ccache will disable caching if a source code file
has a status change time (ctime) after the start of the ccache
invocation. This sloppiness disables that check. See also
[Handling of newly created source files].
include_file_mtime
By default, ccache will disable caching if a source code file
has a modification time (mtime) after the start of the ccache
invocation. This sloppiness disables that check. See also
[Handling of newly created source files].
ivfsoverlay
Ignore the Clang compiler option -ivfsoverlay and its argument.
This is useful if you use Xcode, which uses a virtual file
system (VFS) for things like combining Objective-C and Swift
code.
locale
Ccache includes the environment variables LANG, LC_ALL,
LC_CTYPE and LC_MESSAGES in the hash by default since they may
affect localization of compiler warning messages. Set this
sloppiness to tell ccache not to do that.
modules
By default, ccache will not cache compilations if -fmodules is
used since it cannot hash the state of compiler's internal
representation of relevant modules. This sloppiness allows
caching in such a case. See C++ MODULES for more information.
pch_defines
Be sloppy about #define directives when precompiling a header
file. See PRECOMPILED HEADERS for more information.
random_seed
Ignore the -frandom-seed option and its arguments when
computing the input hash. This is useful if your build system
generates different seeds between builds and you are OK with
reusing cached results.
system_headers
Only check non-system headers in direct mode. This can be
useful if e.g. your system headers tend to change but you know
that the changes don't matter. Notes:
o This sloppiness is only supported for GCC-like compilers,
not MSVC.
o System headers are still taken into account for
preprocessed lookup.
o You can get stale cache hits if the system headers do
change in incompatible ways.
o See also the ignore_headers_in_manifest setting.
time_macros
Ignore __DATE__, __TIME__ and __TIMESTAMP__ being present in
the source code.
See the discussion under TROUBLESHOOTING for more information.
stats (CCACHE_STATS or CCACHE_NOSTATS, see Boolean values above)
If true, ccache will update the statistics counters on each
compilation. The default is true. If false, [automatic cleanup]
will be disabled as well.
stats_log (CCACHE_STATSLOG)
If set to a file path, ccache will write statistics counter updates
to the specified file. This is useful for getting statistics for
individual builds. To show a summary of the current stats log, use
ccache --show-log-stats.
Note
Lines in the stats log starting with a hash sign (#) are
comments.
temporary_dir (CCACHE_TEMPDIR)
This option specifies where ccache will put temporary files. The
default is $XDG_RUNTIME_DIR/ccache-tmp (typically
/run/user/<UID>/ccache-tmp) if XDG_RUNTIME_DIR is set and the
directory exists, otherwise <cache_dir>/tmp.
Note
In previous versions of ccache, CCACHE_TEMPDIR had to be on the
same
filesystem as the CCACHE_DIR path, but this requirement has
been relaxed.
umask (CCACHE_UMASK)
This option (an octal integer) specifies the umask for files and
directories in the cache directory. This is mostly useful when you
wish to share your cache with other users.
Disabling ccache
To disable ccache completely for all invocations, set disable = true
(CCACHE_DISABLE=1). You can also disable ccache for a certain source
code file by adding the string ccache:disable in a comment in the first
4096 bytes of the file. In the latter case the Ccache disabled
statistics counter will be increased.
REMOTE STORAGE BACKENDS
The remote_storage option lets you configure ccache to use one or
several remote storage backends. By default, the local cache directory
located in cache_dir will be queried first and remote storage second,
but remote_only can be set to true to disable local storage. Note that
cache statistics counters will still be kept in the local cache
directory -- remote storage backends only store compilation results and
manifests.
A remote storage backend is specified with a URL, optionally followed
by a pipe (|) and a pipe-separated list of attributes. An attribute is
key=value or just key as a short form of key=true. Attribute values
must be percent-encoded
<https://en.wikipedia.org/wiki/Percent-encoding> if they contain
percent, pipe or space characters.
Attributes for all backends
These optional attributes are available for all remote storage
backends:
o read-only: If true, only read from this backend, don't write. The
default is false.
o shards: A comma-separated list of names for sharding (partitioning)
the cache entries using Rendezvous hashing
<https://en.wikipedia.org/wiki/Rendezvous_hashing>, typically to
spread the cache over a server cluster. When set, the storage URL
must contain an asterisk (*), which will be replaced by one of the
shard names to form a real URL. A shard name can optionally have an
appended weight within parentheses to indicate how much of the key
space should be associated with that shard. A shard with weight w
will contain w/S of the cache, where S is the sum of all shard
weights. A weight could for instance be set to represent the
available memory for a memory cache on a specific server. The
default weight is 1.
Examples:
o redis://cache-*.example.com|shards=a(3),b(1),c(1.5) will put
55% (3/5.5) of the cache on redis://cache-a.example.com, 18%
(1/5.5) on redis://cache-b.example.com and 27% (1.5/5.5) on
redis://cache-c.example.com.
o http://example.com/*|shards=alpha,beta will put 50% of the
cache on http://example.com/alpha and 50% on
http://example.com/beta.
Storage interaction
The table below describes the interaction between local and remote
storage on cache hits and misses if remote_only is false (which is the
default):
+--------------+----------------+-------------------+
|Local storage | Remote storage | What happens |
+--------------+----------------+-------------------+
|miss | miss | Compile, write to |
| | | local, write to |
| | | remote[1] |
+--------------+----------------+-------------------+
|miss | hit | Read from remote, |
| | | write to local |
+--------------+----------------+-------------------+
|hit | - | Read from local, |
| | | don't write to |
| | | remote[2] |
+--------------+----------------+-------------------+
[1] Unless remote storage has attribute read-only=true.
[2] Unless local storage is set to share its cache hits with the
reshare option.
If remote_only is true:
+--------------+----------------+---------------------+
|Local storage | Remote storage | What happens |
+--------------+----------------+---------------------+
|- | miss | Compile, write to |
| | | remote, don't write |
| | | to local |
+--------------+----------------+---------------------+
|- | hit | Read from remote, |
| | | don't write to |
| | | local |
+--------------+----------------+---------------------+
File storage backend
URL format: file:DIRECTORY or file://[HOST]DIRECTORY
This backend stores data as separate files in a directory structure
below DIRECTORY, similar (but not identical) to the local cache
storage. A typical use case for this backend would be sharing a cache
on an NFS directory. DIRECTORY must start with a slash. HOST can be
the empty string or localhost. On Windows, HOST can also be the name of
a server hosting a shared folder.
Important
ccache will not perform any cleanup of the storage -- that has to
be
done by other means, for instance by running ccache --trim-dir
periodically.
Examples:
o file:/shared/nfs/directory
o file:///shared/nfs/directory|umask=002|update-mtime=true
o file:///Z:/example/windows/folder
o file://example.com/shared/ccache%20folder
Optional attributes:
o layout: How to store file under the cache directory. Available
values:
o flat: Store all files directly under the cache directory.
o subdirs: Store files in 256 subdirectories of the cache
directory.
The default is subdirs.
o umask: This attribute (an octal integer) overrides the umask to use
for files and directories in the cache directory.
o update-mtime: If true, update the modification time (mtime) of
cache entries that are read. The default is false.
HTTP storage backend
URL format: http://HOST[:PORT][/PATH]
This backend stores data in an HTTP-compatible server. The required
HTTP methods are GET, PUT and DELETE.
Important
ccache will not perform any cleanup of the storage -- that has to
be
done by other means, for instance by running ccache --trim-dir
periodically.
Note
HTTPS is not supported.
Tip
See
How to set up HTTP storage
<https://ccache.dev/howto/http-storage.html> for hints on how to
set up an HTTP server for use with ccache.
Examples:
o http://localhost
o http://someusername:p4ssw0rd@example.com/cache/
o http://localhost:8080|layout=bazel|connect-timeout=50
Optional attributes:
o bearer-token: Bearer token used to authorize the HTTP requests.
o connect-timeout: Timeout (in ms) for network connection. The
default is 100.
o keep-alive: If true, keep the HTTP connection to the storage server
open to avoid reconnects. The default is true.
o layout: How to map key names to the path part of the URL. Available
values:
o bazel: Store values in a format compatible with the Bazel HTTP
caching protocol. More specifically, the entries will be stored
as 64 hex digits under the /ac/ part of the cache.
Note
You may have to disable verification of action cache values
in the server
for this to work since ccache entries are not valid action
result metadata values.
o flat: Append the key directly to the path part of the URL (with
a leading slash if needed).
o subdirs: Append the first two characters of the key to the URL
(with a leading slash if needed), followed by a slash and the
rest of the key. This divides the entries into 256 buckets.
The default is subdirs.
o header: Add the key=value pair to the HTTP headers of the request.
For example: header=Content-Type=application/octet-stream adds
"Content-Type: application/octet-stream" to the http headers of the
request.
o operation-timeout: Timeout (in ms) for HTTP requests. The default
is 10000.
Redis storage backend
URL formats:
redis://[[USERNAME:]PASSWORD@]HOST[:PORT][/DBNUMBER]
redis+unix:SOCKET_PATH[?db=DBNUMBER]
redis+unix://[[USERNAME:]PASSWORD@localhost]SOCKET_PATH[?db=DBNUMBER]
This backend stores data in a Redis <https://redis.io> (or
Redis-compatible) server. There are implementations for both
memory-based and disk-based storage. PORT defaults to 6379 and
DBNUMBER defaults to 0.
Note
ccache will not perform any cleanup of the Redis storage, but you
can
configure LRU eviction <https://redis.io/topics/lru-cache>.
Tip
See
How to set up Redis <https://ccache.dev/howto/redis-storage.html>
storage" for hints on setting up a Redis server for use with
ccache.
Tip
You can set up a cluster of Redis servers using the shards
attribute
described in REMOTE STORAGE BACKENDS.
Examples:
o redis://localhost
o redis://p4ssw0rd@cache.example.com:6379/0|connect-timeout=50
o redis+unix:/run/redis.sock
o redis+unix:///run/redis.sock
o redis+unix://p4ssw0rd@localhost/run/redis.sock?db=0
Optional attributes:
o connect-timeout: Timeout (in ms) for network connection. The
default is 100.
o operation-timeout: Timeout (in ms) for Redis commands. The default
is 10000.
CACHE SIZE MANAGEMENT
By default, ccache has a 5 GB limit on the total size of files in the
cache and no limit on the number of files. You can set different limits
using the command line options -M/--max-size and -F/--max-files. Use
the -s/--show-stats option to see the cache size and the currently
configured limits (in addition to other various statistics).
Cleanup can be triggered in two different ways: automatic and manual.
Automatic cleanup
After a new compilation result has been written to the local cache,
ccache will trigger an automatic cleanup if max_size or max_files is
exceeded. The cleanup removes cache entries in approximate LRU (least
recently used) order based on the modification time (mtime) of files in
the cache. For this reason, ccache updates mtime of the cache files
read on a cache hit to mark them as recently used.
For performance reasons only entries in a subset of the cache are
considered when automatic cleanup is triggered. This means that there
is no guarantee that the oldest entries of the whole cache are evicted
on each individual cleanup, but over time the cache as a whole will
have "approximate LRU eviction" behavior.
Manual cleanup
You can run ccache -c/--cleanup to force cleanup of the whole cache.
This will recalculate the cache size information and also make sure
that the cache size does not exceed max_size and max_files.
Note that there is no guarantee that only the oldest entries are
evicted, as discussed in Automatic cleanup above. To evict based on age
you can use ccache --evict-older-than AGE.
CACHE COMPRESSION
Ccache will by default compress all data it puts into the cache using
the compression algorithm Zstandard <http://zstd.net> (zstd) using
compression level 1. The algorithm is fast enough that there should be
little reason to turn off compression to gain performance. One
exception is if the cache is located on a compressed file system, in
which case the compression performed by ccache of course is redundant.
See the documentation for the configuration options compression and
compression_level for more information.
You can use the command line option -x/--show-compression to print
information related to compression. Example:
Total data: 14.8 GB (16.0 GB disk blocks)
Compressed data: 11.3 GB (30.6% of original size)
Original size: 36.9 GB
Compression ratio: 3.267 x (69.4% space savings)
Incompressible data: 3.5 GB
Notes:
o The "disk blocks" size is the cache size when taking disk block
size into account. This value should match the "Cache size" value
from "ccache --show-stats". The other size numbers refer to actual
content sizes.
o "Compressed data" refers to result and manifest files stored in the
cache.
o "Incompressible data" refers to files that are always stored
uncompressed (triggered by enabling file_clone or hard_link) or
unknown files (for instance files created by older ccache
versions).
o The compression ratio is affected by compression_level.
The cache data can also be recompressed to another compression level
(or made uncompressed) with the command line option -X/--recompress. If
you choose to disable compression by default or to use a low
compression level, you can (re)compress newly cached data with a higher
compression level after the build or at another time when there are
more CPU cycles available, for instance every night. Full recompression
potentially takes a lot of time, but only files that are currently
compressed with a different level than the target level will be
recompressed.
CACHE STATISTICS
ccache --show-stats shows a summary of statistics, including cache
size, cleanups (number of performed cleanups, either implicitly due to
a cache size limit being reached or due to explicit ccache -c calls),
overall hit rate, hit rate for direct/preprocessed modes and hit rate
for local and remote storage.
The summary also includes counters called "Errors" and "Uncacheable",
which are sums of more detailed counters. To see those detailed
counters, use the -v/--verbose flag. The verbose mode can show the
following counters:
+---------------------------+----------------------------+
|Counter | Description |
+---------------------------+----------------------------+
|Autoconf compile/link | Uncacheable compilation or |
| | linking by an Autoconf |
| | test. |
+---------------------------+----------------------------+
|Bad compiler arguments | Malformed compiler |
| | argument, e.g. missing a |
| | value for a compiler |
| | option that requires an |
| | argument or failure to |
| | read a file specified by a |
| | compiler option argument. |
+---------------------------+----------------------------+
|Called for linking | The compiler was called |
| | for linking, not |
| | compiling. Ccache only |
| | supports compilation of a |
| | single file, i.e. calling |
| | the compiler with the -c |
| | option to produce a single |
| | object file from a single |
| | source file. |
+---------------------------+----------------------------+
|Called for preprocessing | The compiler was called |
| | for preprocessing, not |
| | compiling. |
+---------------------------+----------------------------+
|Ccache disabled | Ccache was disabled by a |
| | ccache:disable string in |
| | the source code file. |
+---------------------------+----------------------------+
|Could not use modules | Preconditions for using |
| | C++ MODULES were not |
| | fulfilled. |
+---------------------------+----------------------------+
|Could not use precompiled | Preconditions for using |
|header | precompiled headers were |
| | not fulfilled. |
+---------------------------+----------------------------+
|Could not read or parse | An input file could not be |
|input file | read or parsed (see the |
| | debug log for details). |
+---------------------------+----------------------------+
|Could not write to output | The output path specified |
|file | with -o could not be |
| | written to. |
+---------------------------+----------------------------+
|Compilation failed | The compilation failed. No |
| | result stored in the |
| | cache. |
+---------------------------+----------------------------+
|Compiler check failed | A compiler check program |
| | specified by |
| | compiler_check |
| | (CCACHE_COMPILERCHECK) |
| | failed. |
+---------------------------+----------------------------+
|Compiler output file | One of the files expected |
|missing | to be produced by the |
| | compiler was missing after |
| | compilation. |
+---------------------------+----------------------------+
|Compiler produced empty | The compiler's output file |
|output | (typically an object file) |
| | was empty after |
| | compilation. |
+---------------------------+----------------------------+
|Could not find the | The compiler to execute |
|compiler | could not be found. |
+---------------------------+----------------------------+
|Error hashing extra file | Failure reading a file |
| | specified by |
| | extra_files_to_hash |
| | (CCACHE_EXTRAFILES). |
+---------------------------+----------------------------+
|Forced recache | CCACHE_RECACHE was used to |
| | overwrite an existing |
| | result. |
+---------------------------+----------------------------+
|Input file modified during | An input file was modified |
|compilation | during compilation. |
+---------------------------+----------------------------+
|Internal error | Unexpected failure, e.g. |
| | due to problems |
| | reading/writing the cache. |
+---------------------------+----------------------------+
|Missing cache file | A file was unexpectedly |
| | missing from the cache. |
| | This only happens in rare |
| | situations, e.g. if one |
| | ccache instance is about |
| | to get a file from the |
| | cache while another |
| | instance removed the file |
| | as part of cache cleanup. |
+---------------------------+----------------------------+
|Multiple source files | The compiler was called to |
| | compile multiple source |
| | files in one go. This is |
| | not supported by ccache. |
+---------------------------+----------------------------+
|No input file | No input file was |
| | specified to the compiler. |
+---------------------------+----------------------------+
|Output to stdout | The compiler was |
| | instructed to write its |
| | output to standard output |
| | using -o -. This is not |
| | supported by ccache. |
+---------------------------+----------------------------+
|Preprocessing failed | Preprocessing the source |
| | code using the compiler's |
| | -E option failed. |
+---------------------------+----------------------------+
|Unsupported code directive | Code like the assembler |
| | .incbin directive was |
| | found. This is not |
| | supported by ccache. |
+---------------------------+----------------------------+
|Unsupported compiler | A compiler option not |
|option | supported by ccache was |
| | found. |
+---------------------------+----------------------------+
|Unsupported environment | An environment variable |
|variable | not supported by ccache |
| | was set. |
+---------------------------+----------------------------+
|Unsupported source | A source language e.g. |
|language | specified with -x was |
| | unsupported by ccache. |
+---------------------------+----------------------------+
HOW CCACHE WORKS
The basic idea is to detect when you are compiling exactly the same
code a second time and reuse the previously produced output. The
detection is done by hashing different kinds of information that should
be unique for the compilation and then using the hash sum to identify
the cached output. Ccache uses BLAKE3, a very fast cryptographic hash
algorithm, for the hashing. On a cache hit, ccache is able to supply
all of the correct compiler outputs (including all warnings, dependency
file, etc) from the cache. Data stored in the cache is checksummed with
XXH3, an extremely fast non-cryptographic algorithm, to detect
corruption.
Ccache has two ways of gathering information used to look up results in
the cache:
o the preprocessor mode, where ccache runs the preprocessor on the
source code and hashes the result
o the direct mode, where ccache hashes the source code and include
files directly
The direct mode is generally faster since running the preprocessor has
some overhead.
If no previous result is detected (i.e., there is a cache miss) using
the direct mode, ccache will fall back to the preprocessor mode unless
the depend mode is enabled. In the depend mode, ccache never runs the
preprocessor, not even on cache misses. Read more in The depend mode
below.
Common hashed information
The following information is always included in the hash:
o the extension used by the compiler for a file with preprocessor
output (normally .i for C code and .ii for C++ code)
o the compiler's size and modification time (or other
compiler-specific information specified by compiler_check)
o the name of the compiler
o the current directory (if hash_dir is enabled)
o contents of files specified by extra_files_to_hash (if any)
The preprocessor mode
In the preprocessor mode, the hash is formed of the common information
and:
o the preprocessor output from running the compiler with -E
o the command line options except those that affect include files
(-I, -include, -D, etc; the theory is that these command line
options will change the preprocessor output if they have any effect
at all)
o any standard error output generated by the preprocessor
Based on the hash, the cached compilation result can be looked up
directly in the cache.
The direct mode
In the direct mode, the hash is formed of the common information and:
o the input source file
o the compiler options
Based on the hash, a data structure called "manifest" is looked up in
the cache. The manifest contains:
o references to cached compilation results (object file, dependency
file, etc) that were produced by previous compilations that matched
the hash
o paths to the include files that were read at the time the
compilation results were stored in the cache
o hash sums of the include files at the time the compilation results
were stored in the cache
The current contents of the include files are then hashed and compared
to the information in the manifest. If there is a match, ccache knows
the result of the compilation. If there is no match, ccache falls back
to running the preprocessor. The output from the preprocessor is parsed
to find the include files that were read. The paths and hash sums of
those include files are then stored in the manifest along with
information about the produced compilation result.
There is a catch with the direct mode: header files that were used by
the compiler are recorded, but header files that were not used, but
would have been used if they existed, are not. So, when ccache checks
if a result can be taken from the cache, it currently can't check if
the existence of a new header file should invalidate the result. In
practice, the direct mode is safe to use in the absolute majority of
cases.
The direct mode will be disabled if any of the following holds:
o direct_mode is false
o a modification time of one of the include files is too new (needed
to avoid a race condition)
o a compiler option not supported by the direct mode is used, for
example:
o a -Wp,* compiler option other than -Wp,-MD,<path>,
-Wp,-MMD,<path>, -Wp,-D<macro[=defn]> or -Wp,-U<macro>
o most uses of -Xpreprocessor
o the string __TIME__ is present in the source code
The depend mode
If the depend mode is enabled, ccache will not use the preprocessor at
all. The hash used to identify results in the cache will be based on
the direct mode hash described above plus information about include
files read from the dependency list generated by MSVC with
/showIncludes, or the dependency file generated by other compilers with
-MD or -MMD.
Advantages:
o The ccache overhead of a cache miss will be much smaller.
o Not running the preprocessor at all can be good if compilation is
performed remotely, for instance when using distcc or similar;
ccache then won't make potentially costly preprocessor calls on the
local machine.
Disadvantages:
o The cache hit rate will likely be lower since any change to
compiler options or source code will make the hash different.
Compare this with the default setup where ccache will fall back to
the preprocessor mode, which is tolerant to some types of changes
of compiler options and source code changes.
o If -MD is used, the manifest entries will include system header
files as well, thus slowing down cache hits slightly, just as using
-MD slows down make. This is also the case for MSVC with
/showIncludes.
o If -MMD is used, the manifest entries will not include system
header files, which means ccache will ignore changes in them.
The depend mode will be disabled if any of the following holds:
o depend_mode is false.
o run_second_cpp is false.
o The compiler is not generating dependencies using -MD or -MMD (for
MSVC, /showIncludes is added automatically if not specified by the
user).
HANDLING OF NEWLY CREATED SOURCE FILES
If modification time (mtime) or status change time (ctime) of the
source file or one of the include files is equal to (or newer than) the
time that ccache was invoked, ccache disables caching completely. This
is done as a safety measure to avoid a race condition (see below). In
practice, this is only a problem when using file systems with very low
timestamp granularity. You can set sloppiness to
include_file_ctime,include_file_mtime to opt out of the safety measure.
For reference, the race condition mentioned above consists of these
events:
1. A source code file is read by ccache and added to the input hash.
2. The source code file is modified.
3. The compiler is executed and reads the modified source code.
4. Ccache stores the compiler output in the cache associated with the
incorrect key (based on the unmodified source code).
CACHE DEBUGGING
To find out what information ccache actually is hashing, you can enable
the debug mode via the configuration option debug or by setting
CCACHE_DEBUG in the environment. This can be useful if you are
investigating why you don't get cache hits. Note that performance will
be reduced slightly.
When the debug mode is enabled, ccache will create up to five
additional files next to the object file:
+-------------------------------------------+-------------+---------------------+
|Filename | Debug level | Description |
+-------------------------------------------+-------------+---------------------+
|<objectfile>.<timestamp>.ccache-input-c | 2 | Binary input hashed |
| | | by both the direct |
| | | mode and the |
| | | preprocessor mode. |
+-------------------------------------------+-------------+---------------------+
|<objectfile>.<timestamp>.ccache-input-d | 2 | Binary input only |
| | | hashed by the |
| | | direct mode. |
+-------------------------------------------+-------------+---------------------+
|<objectfile>.<timestamp>.ccache-input-p | 2 | Binary input only |
| | | hashed by the |
| | | preprocessor mode. |
+-------------------------------------------+-------------+---------------------+
|<objectfile>.<timestamp>.ccache-input-text | 2 | Human-readable |
| | | combined diffable |
| | | text version of the |
| | | three files above. |
+-------------------------------------------+-------------+---------------------+
|<objectfile>.<timestamp>.ccache-log | 1 | Log for this object |
| | | file. |
+-------------------------------------------+-------------+---------------------+
The timestamp format is
<year><month><day>_<hour><minute><second>_<microsecond>.
If you only need the log file, set debug_level (environment variable
CCACHE_DEBUGLEVEL) to 1.
If debug_dir (environment variable CCACHE_DEBUGDIR) is set, the files
above will be written to that directory with full absolute paths
instead of next to the object file.
In the direct mode, ccache uses the 160 bit BLAKE3 hash of the
"ccache-input-c" + "ccache-input-d" data (where + means concatenation),
while the "ccache-input-c" + "ccache-input-p" data is used in the
preprocessor mode.
The "ccache-input-text" file is a combined text version of the three
binary input files. It has three sections ("COMMON", "DIRECT MODE" and
"PREPROCESSOR MODE"), which is turn contain annotations that say what
kind of data comes next.
To debug why you don't get an expected cache hit for an object file,
you can do something like this:
1. Enable debug (CCACHE_DEBUG).
2. Build.
3. Clean and build again.
4. Compare the <objectfile>.<timestamp>.ccache-input-text files for
the two builds. This together with the
<objectfile>.<timestamp>.ccache-log files should give you some
clues about what is happening.
COMPILING IN DIFFERENT DIRECTORIES
Some information included in the hash that identifies a unique
compilation can contain absolute paths:
o The preprocessed source code may contain absolute paths to include
files if the compiler option -g is used or if absolute paths are
given to -I and similar compiler options.
o Paths specified by compiler options (such as -I, -MF, etc) on the
command line may be absolute.
o The source code file path may be absolute, and that path may
substituted for __FILE__ macros in the source code or included in
warnings emitted to standard error by the preprocessor.
This means that if you compile the same code in different locations,
you can't share compilation results between the different build
directories since you get cache misses because of the absolute build
directory paths that are part of the hash.
Here's what can be done to enable cache hits between different build
directories:
o If you build with -g (or similar) to add debug information to the
object file, you must either:
o use the compiler options -fdebug-prefix-map=<old>=<new> or
-fdebug-compilation-dir for relocating debug info to a common
prefix (e.g. -fdebug-prefix-map=$PWD=. or
-fdebug-compilation-dir=.); or
o set hash_dir = false.
o If you use absolute paths anywhere on the command line (e.g. the
source code file path or an argument to compiler options like -I
and -MF), you must set base_dir to an absolute path to a "base
directory". Ccache will then rewrite absolute paths under that
directory to relative before computing the hash.
PRECOMPILED HEADERS
Ccache has support for precompiled headers with GCC and Clang. However,
you have to do some things to make it work properly:
o You must set sloppiness to pch_defines,time_macros. The reason is
that ccache can't tell whether __TIME__, __DATE__ or __TIMESTAMP__
is used when using a precompiled header. Further, it can't detect
changes in #defines in the source code because of how preprocessing
works in combination with precompiled headers.
o You may also want to include include_file_mtime,include_file_ctime
in sloppiness. See HANDLING OF NEWLY CREATED SOURCE FILES.
o You must either:
o use the compiler option -include to include the precompiled
header (i.e., don't use #include in the source code to include
the header; the filename itself must be sufficient to find the
header, i.e. -I paths are not searched); or
o (for the Clang compiler) use the compiler option -include-pch
to include the PCH file generated from the precompiled header;
or
o (for the GCC compiler) add the compiler option -fpch-preprocess
when compiling.
o If you use Clang, you must compile with -fno-pch-timestamp.
If you don't do this, either the non-precompiled version of the header
file will be used (if available) or ccache will fall back to running
the real compiler and increase the statistics counter "Preprocessing
failed" (if the non-precompiled header file is not available).
C++ MODULES
Ccache has support for Clang's -fmodules option. In practice ccache
only additionally hashes module.modulemap files; it does not know how
Clang handles its cached binary form of modules so those are ignored.
This should not matter in practice: as long as everything else
(including module.modulemap files) is the same the cached result should
work. Still, you must set sloppiness to modules to allow caching.
You must use both direct mode and depend mode. When using the
preprocessor mode Clang does not provide enough information to allow
hashing of module.modulemap files.
SHARING A LOCAL CACHE
A group of developers can increase the cache hit rate by sharing a
local cache directory. To share a local cache without unpleasant side
effects, the following conditions should to be met:
o Use the same cache directory.
o Make sure that the configuration option hard_link is false (which
is the default).
o Make sure that all users are in the same group.
o Set the configuration option umask to 002. This ensures that cached
files are accessible to everyone in the group.
o Make sure that all users have write permission in the entire cache
directory (and that you trust all users of the shared cache).
o Make sure that the setgid bit is set on all directories in the
cache. This tells the filesystem to inherit group ownership for new
directories. The following command might be useful for this:
find $CCACHE_DIR -type d | xargs chmod g+s
The reason to avoid the hard link mode is that the hard links cause
unwanted side effects, as all links to a cached file share the file's
modification timestamp. This results in false dependencies to be
triggered by timestamp-based build systems whenever another user links
to an existing file. Typically, users will see that their libraries
and binaries are relinked without reason.
You may also want to make sure that a base directory is set
appropriately, as discussed in a previous section.
SHARING A CACHE ON NFS
It is possible to put the cache directory on an NFS filesystem (or
similar filesystems), but keep in mind that:
o Having the cache on NFS may slow down compilation. Make sure to do
some benchmarking to see if it's worth it.
o Ccache hasn't been tested very thoroughly on NFS.
A tip is to set temporary_dir to a directory on the local host to avoid
NFS traffic for temporary files.
It is recommended to use the same operating system version when using a
shared cache. If operating system versions are different then system
include files will likely be different and there will be few or no
cache hits between the systems. One way of improving cache hit rate in
that case is to set sloppiness to system_headers to ignore system
headers.
An alternative to putting the main cache directory on NFS is to set up
a remote storage file cache.
USING CCACHE WITH OTHER COMPILER WRAPPERS
The recommended way of combining ccache with another compiler wrapper
(such as "distcc") is by letting ccache execute the compiler wrapper.
This is accomplished by defining prefix_command, for example by setting
the environment variable CCACHE_PREFIX to the name of the wrapper (e.g.
distcc). Ccache will then prefix the command line with the specified
command when running the compiler. To specify several prefix commands,
set prefix_command to a colon-separated list of commands.
Unless you set compiler_check to a suitable command (see the
description of that configuration option), it is not recommended to use
the form ccache anotherwrapper compiler args as the compilation
command. It's also not recommended to use the masquerading technique
for the other compiler wrapper. The reason is that by default, ccache
will in both cases hash the mtime and size of the other wrapper instead
of the real compiler, which means that:
o Compiler upgrades will not be detected properly.
o The cached results will not be shared between compilations with and
without the other wrapper.
Another minor thing is that if prefix_command is used, ccache will not
invoke the other wrapper when running the preprocessor, which increases
performance. You can use prefix_command_cpp if you also want to invoke
the other wrapper when doing preprocessing (normally by adding -E).
CAVEATS
o The direct mode fails to pick up new header files in some rare
scenarios. See The direct mode above.
TROUBLESHOOTING
General
A general tip for getting information about what ccache is doing is to
enable debug logging by setting the configuration option debug (or the
environment variable CCACHE_DEBUG); see CACHE DEBUGGING for more
information. Another way of keeping track of what is happening is to
check the output of ccache -s.
Performance
Ccache has been written to perform well out of the box, but sometimes
you may have to do some adjustments of how you use the compiler and
ccache in order to improve performance.
Since ccache works best when I/O is fast, put the cache directory on a
fast storage device if possible. Having lots of free memory so that
files in the cache directory stay in the disk cache is also preferable.
A good way of monitoring how well ccache works is to run ccache -s
before and after your build and then compare the statistics counters.
Here are some common problems and what may be done to increase the hit
rate:
o If the counter for preprocessed cache hits has been incremented
instead of the one for direct cache hits, ccache has fallen back to
preprocessor mode, which is generally slower. Some possible reasons
are:
o The source code has been modified in such a way that the
preprocessor output is not affected.
o Compiler arguments that are hashed in the direct mode but not
in the preprocessor mode have changed (-I, -include, -D, etc)
and they didn't affect the preprocessor output.
o The compiler option -Xpreprocessor or -Wp,* (except
-Wp,-MD,<path>, -Wp,-MMD,<path>, and -Wp,-D<define>) is used.
o This was the first compilation with a new value of the base
directory.
o A modification or status change time of one of the include
files is too new . See HANDLING OF NEWLY CREATED SOURCE FILES.
o The __TIME__ preprocessor macro is (potentially) being used.
Ccache turns off direct mode if __TIME__ is present in the
source code. This is done as a safety measure since the string
indicates that a __TIME__ macro may affect the output. (To be
sure, ccache would have to run the preprocessor, but the sole
point of the direct mode is to avoid that.) If you know that
__TIME__ isn't used in practise, or don't care if ccache
produces objects where __TIME__ is expanded to something in the
past, you can set sloppiness to time_macros.
o The __DATE__ preprocessor macro is (potentially) being used and
the date has changed. This is similar to how __TIME__ is
handled. If __DATE__ is present in the source code, ccache
hashes the current date in order to be able to produce the
correct object file if the __DATE__ macro affects the output.
If you know that __DATE__ isn't used in practise, or don't care
if ccache produces objects where __DATE__ is expanded to
something in the past, you can set sloppiness to time_macros.
o The __TIMESTAMP__ preprocessor macro is (potentially) being
used and the source file's modification time has changed. This
is similar to how __TIME__ is handled. If __TIMESTAMP__ is
present in the source code, ccache hashes the string
representation of the source file's modification time in order
to be able to produce the correct object file if the
__TIMESTAMP__ macro affects the output. If you know that
__TIMESTAMP__ isn't used in practise, or don't care if ccache
produces objects where __TIMESTAMP__ is expanded to something
in the past, you can set sloppiness to time_macros.
o The input file path has changed. Ccache includes the input file
path in the direct mode hash to be able to take relative
include files into account and to produce a correct object file
if the source code includes a __FILE__ macro.
o If a cache hit counter was not incremented even though the same
code has been compiled and cached before, ccache has either
detected that something has changed anyway or a cleanup has been
performed (either explicitly or implicitly when a cache limit has
been reached). Some perhaps unobvious things that may result in a
cache miss are usage of __TIME__, __DATE__ or __TIMESTAMP__ macros,
or use of automatically generated code that contains a timestamp,
build counter or other volatile information.
o If "Multiple source files" has been incremented, it's an indication
that the compiler has been invoked on several source code files at
once. Ccache doesn't support that. Compile the source code files
separately if possible.
o If "Unsupported compiler option" has been incremented, enable debug
logging and check which compiler option was rejected.
o If "Preprocessing failed" has been incremented, one possible reason
is that precompiled headers are being used. See PRECOMPILED HEADERS
for how to remedy this.
o If "Could not use precompiled header" has been incremented, see
PRECOMPILED HEADERS.
o If "Could not use modules" has been incremented, see C++ MODULES.
Corrupt object files
It should be noted that ccache is susceptible to general storage
problems. If a bad object file sneaks into the cache for some reason,
it will of course stay bad. Some possible reasons for erroneous object
files are bad hardware (disk drive, disk controller, memory, etc),
buggy drivers or file systems, a bad prefix_command or compiler
wrapper. If this happens, the easiest way of fixing it is this:
1. Build so that the bad object file ends up in the build tree.
2. Remove the bad object file from the build tree.
3. Rebuild with CCACHE_RECACHE set.
An alternative is to clear the whole cache with ccache -C if you don't
mind losing other cached results.
There are no reported issues about ccache producing broken object files
reproducibly. That doesn't mean it can't happen, so if you find a
repeatable case, please report it.
MORE INFORMATION
Credits, mailing list information, bug reporting instructions, source
code, etc, can be found on ccache's web site: <https://ccache.dev>.
AUTHOR
Ccache was originally written by Andrew Tridgell and is currently
developed and maintained by Joel Rosdahl. See AUTHORS.txt or
AUTHORS.html and <https://ccache.dev/credits.html> for a list of
contributors.
Ccache 4.11 2025-03-11 ccache(1)
ccache 4.11 - Generated Thu Mar 13 07:15:59 CDT 2025