1
mailto(bugs@ccache.samba.org)
2
manpage(ccache)(1)(April 2002)()()
3
manpagename(ccache)(a fast compiler cache)
8
ccache <compiler> [COMPILER OPTIONS]
10
<compiler> [COMPILER OPTIONS]
14
ccache is a compiler cache. It speeds up re-compilation of C/C++ code
15
by caching previous compiles and detecting when the same compile is
18
manpagesection(OPTIONS SUMMARY)
20
Here is a summary of the options to ccache.
23
-s show statistics summary
25
-c run a cache cleanup
26
-C clear the cache completely
27
-F <maxfiles> set maximum files in cache
28
-M <maxsize> set maximum size of cache (use G, M or K)
30
-V print version number
35
These options only apply when you invoke ccache as "ccache". When
36
invoked as a compiler none of these options apply. In that case your
37
normal compiler options apply and you should refer to your compilers
41
dit(bf(-h)) Print a options summary page
43
dit(bf(-s)) Print the current statistics summary for the cache. The
44
statistics are stored spread across the subdirectories of the
45
cache. Using "ccache -s" adds up the statistics across all
46
subdirectories and prints the totals.
48
dit(bf(-z)) Zero the cache statistics.
50
dit(bf(-V)) Print the ccache version number
52
dit(bf(-c)) Clean the cache and re-calculate the cache file count and
53
size totals. Normally the -c option should not be necessary as ccache
54
keeps the cache below the specified limits at runtime and keeps
55
statistics up to date on each compile. This option is mostly useful
56
if you manually modify the cache contents or believe that the cache
57
size statistics may be inaccurate.
59
dit(bf(-C)) Clear the entire cache, removing all cached files.
61
dit(bf(-F maxfiles)) This sets the maximum number of files allowed in
62
the cache. The value is stored inside the cache directory and applies
63
to all future compiles. Due to the way the value is stored the actual
64
value used is always rounded down to the nearest multiple of 16.
66
dit(bf(-M maxsize)) This sets the maximum cache size. You can specify
67
a value in gigabytes, megabytes or kilobytes by appending a G, M or K
68
to the value. The default is gigabytes. The actual value stored is
69
rounded down to the nearest multiple of 16 kilobytes.
73
manpagesection(INSTALLATION)
75
There are two ways to use ccache. You can either prefix your compile
76
commands with "ccache" or you can create a symbolic link between
77
ccache and the names of your compilers. The first method is most
78
convenient if you just want to try out ccache or wish to use it for
79
some specific projects. The second method is most useful for when you
80
wish to use ccache for all your compiles.
82
To install for usage by the first method just copy ccache to somewhere
85
To install for the second method do something like this:
87
cp ccache /usr/local/bin/
88
ln -s /usr/local/bin/ccache /usr/local/bin/gcc
89
ln -s /usr/local/bin/ccache /usr/local/bin/g++
90
ln -s /usr/local/bin/ccache /usr/local/bin/cc
92
This will work as long as /usr/local/bin comes before the path to gcc
93
(which is usually in /usr/bin). After installing you may wish to run
94
"which gcc" to make sure that the correct link is being used.
96
Note! Do not use a hard link, use a symbolic link. A hardlink will
97
cause "interesting" problems.
99
manpagesection(EXTRA OPTIONS)
101
When run as a compiler front end ccache usually just takes the same
102
command line options as the compiler you are using. The only exception
103
to this is the option '--ccache-skip'. That option can be used to tell
104
ccache that the next option is definitely not a input filename, and
105
should be passed along to the compiler as-is.
107
The reason this can be important is that ccache does need to parse the
108
command line and determine what is an input filename and what is a
109
compiler option, as it needs the input filename to determine the name
110
of the resulting object file (among other things). The heuristic
111
ccache uses in this parse is that any string on the command line that
112
exists as a file is treated as an input file name (usually a C
113
file). By using --ccache-skip you can force an option to not be
114
treated as an input file name and instead be passed along to the
115
compiler as a command line option.
117
manpagesection(ENVIRONMENT VARIABLES)
119
ccache uses a number of environment variables to control operation. In
120
most cases you won't need any of these as the defaults will be fine.
124
dit(bf(CCACHE_DIR)) the CCACHE_DIR environment variable specifies
125
where ccache will keep its cached compiler output. The default is
128
dit(bf(CCACHE_TEMPDIR)) the CCACHE_TEMPDIR environment variable specifies
129
where ccache will put temporary files. The default is the same as
130
CCACHE_DIR. Note that the CCACHE_TEMPDIR path must be on the same
131
filesystem as the CCACHE_DIR path, so that renames of files between
132
the two directories can work.
134
dit(bf(CCACHE_LOGFILE)) If you set the CCACHE_LOGFILE environment
135
variable then ccache will write some log information on cache hits
136
and misses in that file. This is useful for tracking down problems.
138
dit(bf(CCACHE_PATH)) You can optionally set CCACHE_PATH to a colon
139
separated path where ccache will look for the real compilers. If you
140
don't do this then ccache will look for the first executable matching
141
the compiler name in the normal PATH that isn't a symbolic link to
144
dit(bf(CCACHE_CC)) You can optionally set CCACHE_CC to force the name
145
of the compiler to use. If you don't do this then ccache works it out
146
from the command line.
148
dit(bf(CCACHE_PREFIX)) This option adds a prefix to the command line
149
that ccache runs when invoking the compiler. Also see the section
150
below on using ccache with distcc.
152
dit(bf(CCACHE_DISABLE)) If you set the environment variable
153
CCACHE_DISABLE then ccache will just call the real compiler,
154
bypassing the cache completely.
156
dit(bf(CCACHE_READONLY)) the CCACHE_READONLY environment variable
157
tells ccache to attempt to use existing cached object files, but not
158
to try to add anything new to the cache. If you are using this because
159
your CCACHE_DIR is read-only, then you may find that you also need to
160
set CCACHE_TEMPDIR as otherwise ccache will fail to create the
163
dit(bf(CCACHE_CPP2)) If you set the environment variable CCACHE_CPP2
164
then ccache will not use the optimisation of avoiding the 2nd call to
165
the pre-processor by compiling the pre-processed output that was used
166
for finding the hash in the case of a cache miss. This is primarily a
167
debugging option, although it is possible that some unusual compilers
168
will have problems with the intermediate filename extensions used in
169
this optimisation, in which case this option could allow ccache to be
172
dit(bf(CCACHE_NOSTATS)) If you set the environment variable
173
CCACHE_NOSTATS then ccache will not update the statistics files on
176
dit(bf(CCACHE_NLEVELS)) The environment variable CCACHE_NLEVELS allows
177
you to choose the number of levels of hash in the cache directory. The
178
default is 2. The minimum is 1 and the maximum is 8.
180
dit(bf(CCACHE_HARDLINK)) If you set the environment variable
181
CCACHE_HARDLINK then ccache will attempt to use hard links from the
182
cache directory when creating the compiler output rather than using a
183
file copy. Using hard links is faster, but can confuse programs like
184
'make' that rely on modification times.
186
dit(bf(CCACHE_RECACHE)) This forces ccache to not use any cached
187
results, even if it finds them. New results are still cached, but
188
existing cache entries are ignored.
190
dit(bf(CCACHE_UMASK)) This sets the umask for ccache and all child
191
processes (such as the compiler). This is mostly useful when you wish
192
to share your cache with other users. Note that this also affects the
193
file permissions set on the object files created from your
196
dit(bf(CCACHE_HASHDIR)) This tells ccache to hash the current working
197
directory when calculating the hash that is used to distinguish two
198
compiles. This prevents a problem with the storage of the current
199
working directory in the debug info of a object file, which can lead
200
ccache to give a cached object file that has the working directory in
201
the debug info set incorrectly. This option is off by default as the
202
incorrect setting of this debug info rarely causes problems. If you
203
strike problems with gdb not using the correct directory then enable
206
dit(bf(CCACHE_UNIFY)) If you set the environment variable CCACHE_UNIFY
207
then ccache will use the C/C++ unifier when hashing the pre-processor
208
output if -g is not used in the compile. The unifier is slower than a
209
normal hash, so setting this environment variable loses a little bit
210
of speed, but it means that ccache can take advantage of not
211
recompiling when the changes to the source code consist of
212
reformatting only. Note that using CCACHE_UNIFY changes the hash, so
213
cached compiles with CCACHE_UNIFY set cannot be used when
214
CCACHE_UNIFY is not set and vice versa. The reason the unifier is off
215
by default is that it can give incorrect line number information in
216
compiler warning messages.
218
dit(bf(CCACHE_EXTENSION)) Normally ccache tries to automatically
219
determine the extension to use for intermediate C pre-processor files
220
based on the type of file being compiled. Unfortunately this sometimes
221
doesn't work, for example when using the aCC compiler on HP-UX. On
222
systems like this you can use the CCACHE_EXTENSION option to override
223
the default. On HP-UX set this environment variable to "i" if you use
228
manpagesection(CACHE SIZE MANAGEMENT)
230
By default ccache has a one gigabyte limit on the cache size and no
231
maximum number of files. You can set a different limit using the
232
"ccache -M" and "ccache -F" options, which set the size and number of
235
When these limits are reached ccache will reduce the cache to 20%
236
below the numbers you specified in order to avoid doing the cache
237
clean operation too often.
239
manpagesection(HOW IT WORKS)
241
The basic idea is to detect when you are compiling exactly the same
242
code a 2nd time and use the previously compiled output. You detect
243
that it is the same code by forming a hash of:
246
it() the pre-processor output from running the compiler with -E
247
it() the command line options
248
it() the real compilers size and modification time
249
it() any stderr output generated by the compiler
252
These are hashed using md4 (a strong hash) and a cache file is formed
253
based on that hash result. When the same compilation is done a second
254
time ccache is able to supply the correct compiler output (including
255
all warnings etc) from the cache.
257
ccache has been carefully written to always produce exactly the same
258
compiler output that you would get without the cache. If you ever
259
discover a case where ccache changes the output of your compiler then
262
manpagesection(USING CCACHE WITH DISTCC)
264
distcc is a very useful program for distributing compilation across a
265
range of compiler servers. It is often useful to combine distcc with
266
ccache, so that compiles that are done are sped up by distcc, but that
267
ccache avoids the compile completely where possible.
269
To use distcc with ccache I recommend using the CCACHE_PREFIX
270
option. You just need to set the environment variable CCACHE_PREFIX to
271
'distcc' and ccache will prefix the command line used with the
272
compiler with the command 'distcc'.
274
manpagesection(SHARING A CACHE)
276
A group of developers can increase the cache hit rate by sharing a
277
cache directory. The hard links however cause unwanted side effects,
278
as all links to a cached file share the file's modification timestamp.
279
This results in false dependencies to be triggered by timestamp-based
280
build systems whenever another user links to an existing
281
file. Typically, users will see that their libraries and binaries are
282
relinked without reason. To share a cache without side effects, the
283
following conditions need to be met:
286
it() Use the same bf(CCACHE_DIR) environment variable setting
287
it() Set the bf(CCACHE_NOLINK) environment variable
288
it() Make sure everyone sets the CCACHE_UMASK environment variable
289
to 002, this ensures that cached files are accessible to everyone in
291
it() Make sure that all users have write permission in the entire
292
cache directory (and that you trust all users of the shared cache).
293
it() Make sure that the setgid bit is set on all directories in the
294
cache. This tells the filesystem to inherit group ownership for new
295
directories. The command "chmod g+s `find $CCACHE_DIR -type d`" might
299
manpagesection(HISTORY)
301
ccache was inspired by the compilercache shell script script written
302
by Erik Thiele and I would like to thank him for an excellent piece of
304
url(http://www.erikyyy.de/compilercache/)(http://www.erikyyy.de/compilercache/)
305
for the Erik's scripts.
307
I wrote ccache because I wanted to get a bit more speed out of a
308
compiler cache and I wanted to remove some of the limitations of the
309
shell-script version.
311
manpagesection(DIFFERENCES FROM COMPILERCACHE)
313
The biggest differences between Erik's compilercache script and ccache
316
it() ccache is written in C, which makes it a bit faster (calling out to
317
external programs is mostly what slowed down the scripts).
318
it() ccache can automatically find the real compiler
319
it() ccache keeps statistics on hits/misses
320
it() ccache can do automatic cache management
321
it() ccache can cache compiler output that includes warnings. In many
322
cases this gives ccache a much higher cache hit rate.
323
it() ccache can handle a much wider ranger of compiler options
324
it() ccache avoids a double call to cpp on a cache miss
329
When the cache is stored on an NFS filesystem, the filesystem must be
330
exported with the bf(no_subtree_check) option to make renames between
331
directories reliable.
333
manpagesection(CREDITS)
335
Thanks to the following people for their contributions to ccache
337
it() Erik Thiele for the original compilercache script
338
it() Luciano Rocha for the idea of compiling the pre-processor output
339
to avoid a 2nd cpp pass
340
it() Paul Russell for many suggestions and the debian packaging
345
ccache was written by Andrew Tridgell
346
url(http://samba.org/~tridge/)(http://samba.org/~tridge/)
348
If you wish to report a problem or make a suggestion then please email
349
bugs@ccache.samba.org
351
ccache is released under the GNU General Public License version 2 or
352
later. Please see the file COPYING for license details.