CMAKE Parallel Build issue - cmake

I have external Project which is make based, I am able to compile link successfully. ( I have modified make infra of third party project to break large make into multiple smaller makes so that parallelly different small make can run, but it is orthogonal to this question)
Third party cmake is as below
External_Project_Add( ...
....
#NO BUILD_COMMAND)
But Compilation is too slow, and I can see that max number of compilation doesn't exceed 8 threads and CPU utilization is ~20%. (we are using distcc as well).
If I change as below
External_Project_Add( ...
....
BUILD_COMMAND ${CMAKE_COMMAND} --build <dir> --parallel --)
Compilation is super-fast by atleast 3 time. However occasionally system run out of PIDs.
Is there way to control max number of PIDs/process available to CMAKE ?
I had thought about using -j option is BUILD_COMMAND but it is artificially going to reduce the compilation compute available to my system.
I have tried changing $(MAKE) in my third party make to have -jn, but then it shows jobserver domain warning. It is not good idea to change third party make as well. So I dropped this idea.
Any help is appreciated !

$ cmake --build . --help
Unknown argument --help
Usage: cmake --build <dir> [options] [-- [native-options]]
Options:
-j [<jobs>] --parallel [<jobs>] = Build in parallel using
the given number of jobs. If <jobs> is omitted
the native build tool's default number is used.
The CMAKE_BUILD_PARALLEL_LEVEL environment variable
specifies a default parallel level when this option
is not given.
cmake passes the -j and -j <jobs> request to the build tool. cmake --build will run the appropriate build tool depending on the generator used and will add the correct options to run a parallel build. So just use -j <jobs> to limit the max number of requested jobs.

Related

How to build a CMake project in parallel on all available cores?

This related question shows how to build a CMake project using a specified numbers of cores. For example if I wanted to use 10 cores, I could invoke CMake like this:
cmake --build . -j 10
My question is: how can I build using all my available cores. I effectively want CMake to autodetect my core count and use all of them.
As of CMake 3.22, there is no standard way to do this. However, there are a few practical approaches.
If you use the Ninja or Ninja Multi-Config generators, on any platform, simply running the build with cmake --build /path/to/build-dir will use all cores.
If you are on a UNIX-like command line, you can run cmake --build . -j $(nproc)
If you are on Windows/cmd, you can run cmake --build . -j %NUMBER_OF_PROCESSORS%

How to restrict cmake commands based on which target is built

I have a cmake project which produces several executables. I want to package each executable in seperate Docker containers, so inside the Dockefile, I only built the target that I need:
RUN mkdir build \
&& cd build \
&& cmake /app/project -DCMAKE_BUILD_TYPE=Release
&& make -j 2 myExecutable \
&& make install/fast
This works as expected, but I run into an issue with the conan cmake integration. The installation is done when cmake is called, not during the actual build - this means that no matter which target I want to actually build, all the conan installation calls present in my cmake files are called - so way more packages are installed than necessary.
# for every target
# include conan dependencies (each target has its own conanfile.txt)
conan_cmake_run(CONANFILE conanfile.txt
BASIC_SETUP CMAKE_TARGETS
BUILD_TYPE "${CMAKE_BUILD_TYPE}"
BUILD outdated
${update_conan}
)
conan_target_link_libraries(${PROJECT_NAME})
Is there a way to make the cmake calls dependend on which target I actually want to build?
Unfortunately not, the macro conan_cmake_run has no distinction about which target is involved or even it was executed before. You could use CMake options to run or not conan_cmake_run.
Also, you could comment/vote your request thorough the issue https://github.com/conan-io/cmake-conan/issues/105
Regards!

From CMake setup 'make' to use '-j' option by default

I want my CMake project to be built by make -j N, whenever I call make from the terminal. I don't want to set -j option manually every time.
For that, I set CMAKE_MAKE_PROGRAM variable to the specific command line. I use the ProcessorCount() function, which gives the number of procesors to perform build in parallel.
When I do make, I do not see any speed up. However if I do make -j N, then it is built definitely faster.
Would you please help me on this issue? (I am developing this on Linux.)
Here is the snippet of the code that I use in CMakeList.txt:
include(ProcessorCount)
ProcessorCount(N)
message("number of processors: " ${N})
if(NOT N EQUAL 0)
set(CTEST_BUILD_FLAGS -j${N})
set(ctest_test_args ${ctest_test_args} PARALLEL_LEVEL ${N})
set(CMAKE_MAKE_PROGRAM "${CMAKE_MAKE_PROGRAM} -j ${N}")
endif()
message("cmake make program" ${CMAKE_MAKE_PROGRAM})
Thank you very much.
In case you want to speed up the build you can run multiple make processes in parallel but not cmake.
To perform every build with predefined number of parallel processes you can define this in MAKEFLAGS.
Set MAKEFLAGS in your environment script, e.g. ~/.bashrc as you want:
export MAKEFLAGS=-j8
On Linux the following sets MAKEFLAGS to the number of CPUs - 1: (Keep one CPU free for other tasks while build) and is useful in environments with dynamic ressources, e.g. VMware:
export MAKEFLAGS=-j$(($(grep -c "^processor" /proc/cpuinfo) - 1))
New from cmake v3.12 on:
The command line has a new option --parallel <JOBS>.
Example:
cmake --build build_arm --parallel 4 --target all
Example with number of CPUs- 1 using nproc:
cmake --build build_arm --parallel $(($(nproc) - 1)) --target all
Via setting the CMAKE_MAKE_PROGRAM variable you want to affect the build process. But:
This variable affects only the build via cmake --build, not on native tool (make) call:
The CMAKE_MAKE_PROGRAM variable is set for use by project code. The value is also used by the cmake(1) --build and ctest(1) --build-and-test tools to launch the native build process.
This variable should be a CACHEd one. It is used in such way by make-like generators:
These generators store CMAKE_MAKE_PROGRAM in the CMake cache so that it may be edited by the user.
That is, you need to set this variable with
set(CMAKE_MAKE_PROGRAM <program> CACHE PATH "Path to build tool" FORCE)
This variable should refer to the executable itself, not to a program with arguments:
The value may be the full path to an executable or just the tool name if it is expected to be in the PATH.
That is, value "make -j 2" cannot be used for that variable (splitting arguments as list
set(CMAKE_MAKE_PROGRAM make -j 2 CACHE PATH "Path to build tool" FORCE)
wouldn't help either).
In summary, you may redefine the behavior of cmake --build calls with setting the CMAKE_MAKE_PROGRAM variable to the script, which calls make with parallel options. But you may not affect the behavior of direct make calls.
You may set the env variable MAKEFLAGS using this command
export MAKEFLAGS=-j$(nproc)
My solution is to have a small script which will run make include all sorts of other features, not just the number of CPUs.
I call my script mk and I do a chmod 755 mk so I can run it with ./mk in the root of my project. I also have a few flags to be able to run various things with a simple command line. For example, while working on the code and I get many errors, I like to pipe the output to less. I can do that with ./mk -l without having to retype all the heavy duty Unix stuff...
As you can see, I have the -j4 in a couple of places where it makes sense. For the -l option, I don't want it because in this case it would eventually cause multiple errors to be printed at the same time (I tried that before!)
#!/bin/sh -e
#
# Execute make
case "$1" in
"-l")
make -C ../BUILD/Debug 2>&1 | less -R
;;
"-r")
make -j4 -C ../BUILD/Release
;;
"-d")
rm -rf ../BUILD/Debug/doc/lpp-doc-?.*.tar.gz \
../BUILD/Debug/doc/lpp-doc-?.*
make -C ../BUILD/Debug
;;
"-t")
make -C ../BUILD/Debug
../BUILD/Debug/src/lpp tests/suite/syntax-print.logo
g++ -std=c++14 -I rt l.cpp rt/*.cpp
;;
*)
make -j4 -C ../BUILD/Debug
;;
esac
# From the https://github.com/m2osw/lpp project
With CMake, it wouldn't work unless, as Tsyvarev mentioned, you create your own script. But I personally don't think it's sensible to call make from your make script. Plus it could break a build process which would not expect that strange script. Finally, my script, as I mentioned, allows me to vary the options depending on the situation.
I usually use alias in linux to set cm equal to cmake .. && make -j12. Or write a shell to specify make and clean progress ...
alias cm='cmake .. && make -j12'
Then use cm to make in a single command.

Set the number of threads in a CMake build

cmake --build . --config Release
Is it possible to set the number of cores to be used by the build process?
I'm looking for something similar to GNU make's -j option.
According to the Release Notes, with CMake 3.12 it can be done cross-platform:
The cmake(1) Build Tool Mode (cmake –build) gained --parallel [<jobs>] and -j [<jobs>] options to specify a parallel build level. They map to corresponding options of the native build tool.
You can pass arbitrary arguments to the native build tool with --. Everything after -- will be passed to the build tool. To pass -j 3 in your example, just use
cmake --build . --config Release -- -j 3
Documentation: https://cmake.org/cmake/help/v3.5/manual/cmake.1.html
You could also use Ninja as a build tool, it uses automatically an appropriate number of threads. Or you can modify the make command by defining CMAKE_MAKE_PROGRAM="make -j 3. But this is a rather hacky workaround.

What is CMake equivalent of 'configure --prefix=DIR && make all install '?

I do cmake . && make all install. This works, but installs to /usr/local.
I need to install to a different prefix (for example, to /usr).
What is the cmake and make command line to install to /usr instead of /usr/local?
You can pass in any CMake variable on the command line, or edit cached variables using ccmake/cmake-gui. On the command line,
cmake -DCMAKE_INSTALL_PREFIX:PATH=/usr . && make all install
Would configure the project, build all targets and install to the /usr prefix. The type (PATH) is not strictly necessary, but would cause the Qt based cmake-gui to present the directory chooser dialog.
Some minor additions as comments make it clear that providing a simple equivalence is not enough for some. Best practice would be to use an external build directory, i.e. not the source directly. Also to use more generic CMake syntax abstracting the generator.
mkdir build && cd build && cmake -DCMAKE_INSTALL_PREFIX:PATH=/usr .. && cmake --build . --target install --config Release
You can see it gets quite a bit longer, and isn't directly equivalent anymore, but is closer to best practices in a fairly concise form... The --config is only used by multi-configuration generators (i.e. MSVC), ignored by others.
The ":PATH" part in the accepted answer can be omitted. This syntax may be more memorable:
cmake -DCMAKE_INSTALL_PREFIX=/usr . && make all install
...as used in the answers here.
Note that in both CMake and Autotools you don't always have to set the installation path at configure time. You can use DESTDIR at install time (see also here) instead as in:
make DESTDIR=<installhere> install
See also this question which explains the subtle difference between DESTDIR and PREFIX.
This is intended for staged installs and to allow for storing programs in a different location from where they are run e.g. /etc/alternatives via symbolic links.
However, if your package is relocatable and doesn't need any hard-coded (prefix) paths set via the configure stage you may be able to skip it.
So instead of:
cmake -DCMAKE_INSTALL_PREFIX=/usr . && make all install
you would run:
cmake . && make DESTDIR=/usr all install
Note that, as user7498341 points out, this is not appropriate for cases where you really should be using PREFIX.
The way I build CMake projects cross platform is the following:
/project-root> mkdir build
/project-root> cd build
/project-root/build> cmake -G "<generator>" -DCMAKE_INSTALL_PREFIX=stage ..
/project-root/build> cmake --build . --target=install --config=Release
The first two lines create the out-of-source build directory
The third line generates the build system specifying where to put the installation result (which I always place in ./project-root/build/stage - the path is always considered relative to the current directory if it is not absolute)
The fourth line builds the project configured in . with the buildsystem configured in the line before. It will execute the install target which also builds all necessary dependent targets if they need to be built and then copies the files into the CMAKE_INSTALL_PREFIX (which in this case is ./project-root/build/stage. For multi-configuration builds, like in Visual Studio, you can also specify the configuration with the optional --config <config> flag.
The good part when using the cmake --build command is that it works for all generators (i.e. makefiles and Visual Studio) without needing different commands.
Afterwards I use the installed files to create packages or include them in other projects...
Starting with CMake 3.15, the correct way of achieving this would be using:
cmake --install <dir> --prefix "/usr"
Official Documentation
Starting with CMake 3.21 you can use the --install-prefix option instead of manually setting CMAKE_INSTALL_PREFIX.
The modern equivalent of configure --prefix=DIR && make all install would now be:
cmake -B build --install-prefix=DIR
cmake --build build
cmake --install build
Regarding Bruce Adams answer:
Your answer creates dangerous confusion. DESTDIR is intended for
installs out of the root tree. It allows one to see what would be
installed in the root tree if one did not specify DESTDIR.
PREFIX is the base directory upon which the real installation is
based.
For example, PREFIX=/usr/local indicates that the final destination
of a package is /usr/local. Using DESTDIR=$HOME will install the files
as if $HOME was the root (/). If, say DESTDIR, was /tmp/destdir, one
could see what 'make install' would affect. In that spirit, DESTDIR
should never affect the built objects.
A makefile segment to explain it:
install:
cp program $DESTDIR$PREFIX/bin/program
Programs must assume that PREFIX is the base directory of the final
(i.e. production) directory. The possibility of symlinking a program
installed in DESTDIR=/something only means that the program does not
access files based upon PREFIX as it would simply not work. cat(1)
is a program that (in its simplest form) can run from anywhere.
Here is an example that won't:
prog.pseudo.in:
open("#prefix#/share/prog.db")
...
prog:
sed -e "s/#prefix#/$PREFIX/" prog.pseudo.in > prog.pseudo
compile prog.pseudo
install:
cp prog $DESTDIR$PREFIX/bin/prog
cp prog.db $DESTDIR$PREFIX/share/prog.db
If you tried to run prog from elsewhere than $PREFIX/bin/prog,
prog.db would never be found as it is not in its expected location.
Finally, /etc/alternatives really does not work this way. There are
symlinks to programs installed in the root tree (e.g. vi -> /usr/bin/nvi,
vi -> /usr/bin/vim, etc.).
It is considered bad practice to invoke the actual build system (e.g. via the make command) if using CMake. It is highly recommended to do it like this:
Configure + Generation stages:
cmake -S foo -B _builds/foo/debug -G "Unix Makefiles" -D CMAKE_BUILD_TYPE:STRING=Debug -D CMAKE_DEBUG_POSTFIX:STRING=d -D CMAKE_INSTALL_PREFIX:PATH=/usr
Build and Install stages:
cmake --build _builds/foo/debug --config Debug --target install
When following this approach, the generator can be easily switched (e.g. -G Ninja for Ninja) without having to remember any generator-specific commands.
Note that the CMAKE_BUILD_TYPE variable is only used by single-config generators and the --config argument of the build command is only used by multi-config generators.
Lots of answer, but I figured I'd do a summary to properly group them and explain the differences.
First of all, you can define that prefix one of two ways: during configuration time, or when installing, and that's really up to your needs.
During configuration time
Two options:
cmake -S $src_dir -B $build_dir -D CMAKE_INSTALL_PREFIX=$install_dir
cmake -S $src_dir -B $build_dir --install-prefix=$install_dir # Since CMake 3.21
During install time
Advantage: no need to reconfigure if you want to change it.
Two options:
cmake DESTDIR=$install_dir --build $build_dir --target=install # Makefile only
cmake --install $build_dir --prefix=$install_dir