cmake

View page source

Using CMake to Configure CppAD

The CMake Program

The cmake program enables one to create a single set of scripts, called CMakeLists.txt , that can be used to test and install a program on Unix, Microsoft, or Apple operating systems; see cmake resources .

Build Directory

Create build , a subdirectory of the Distribution Directory , change into the build .

Eigen

The cmake command will automatically search for the eigen package. If it is found, the corresponding examples will be tested and the sparse2eigen utility will be installed; see eigen . The results of cppad’s search for Eigen are reported in the output of the cmake command.

CMake Command

Simple

The simplest version of the cmake command is

cmake ..

This assumes that you are starting in the build directory (so that .. refers to the distribution directory). The command also assumes that cmake is in your execution path with version greater than or equal 2.8. If not, you can put the path to the version of cmake in font the command. It is suggested that you should start with the simple version of the cmake command above and then add options one at a time.

Warning

Version 3.10.2 of cmake will generates the following warning if boost is not present your system:

CMake Warning at /usr/share/cmake-3.10/Modules/FindBoost.cmake:567 ( message ):
      Imported targets and dependency information not available for Boost version
      ( all versions older than 1.33 )
      …

This is not a problem because CppAD will automatically exclude the examples and tests that use boost. Newer version of cmake; e.g., 3.13.4, do not generate this warning (when boost is not present on your system).

true_or_false

The syntax -D name = true_or_false , mean that the variable name can be assigned the value true or false .

Options

The full version of the command, with all its optional arguments is:

cmake \
      -D CMAKE_VERBOSE_MAKEFILE = cmake_verbose_makefile \
      -D CMAKE_BUILD_TYPE = cmake_build_type \
      -G generator \
      \
      -D cppad_prefix = cppad_prefix \
      -D cppad_postfix = cppad_postfix \
      \
      -D cmake_install_includedirs = cmake_install_includedirs \
      -D cmake_install_libdirs = cmake_install_libdirs \
      \
      -D cmake_install_datadir = cmake_install_datadir \
      -D cmake_install_docdir = cmake_install_docdir \
      \
      -D include_doc = true_or_false \
      -D cmake_defined_ok = true_or_false \
      -D cppad_static_lib = true_or_false \
      -D cppad_debug_and_release = true_or_false \
      \
      -D include_adolc = true_or_false \
      -D include_ipopt = true_or_false \
      -D include_cppadcg = true_or_false \
      \
      -D colpack_prefix = colpack_prefix \
      -D fadbad_prefix = fadbad_prefix \
      -D sacado_prefix = sacado_prefix \
      \
      -D cppad_cxx_flags = cppad_cxx_flags \
      -D cppad_link_flags = cppad_link_flags \
      -D cppad_profile_flag = cppad_profile_flag \
      -D cppad_testvector = cppad_testvector \
      -D cppad_max_num_threads = cppad_max_num_threads \
      -D cppad_tape_id_type = cppad_tape_id_type \
      -D cppad_tape_addr_type = cppad_tape_addr_type \
      -D cppad_debug_which = cppad_debug_which \
      \
      ..

msys2

The msys2 system, with mingw-64 and g++ , the following seems to work: Use pacman to install make , gcc , and mingw-w64-x86_64-cmake .

Visual Studio

  1. Make sure you have a version of git for windows .

  2. Launch the Visual Studio developer command window

  3. At the command prompt enter

    where cmake.exe
    

    to see if cmake.exe comes with your version of Visual Studio. If not, download and install the win64-x64 version of the cmake program. Make sure you select the box that adds cmake to your execution path.

  4. Download the CppAD repository, and make it the current directory, using the following commands

       clone https://github.com/coin-or/CppAD.git cppad.git
       cd cppad.git
    
  5. Execute the following commands:

if exist build ( rmdir /S build )
mkdir build
cd build
cmake ^
   -G "NMake Makefiles"^
   -D cppad_cxx_flags="/MP /EHs /EHc /std:c++17 /Zc:__cplusplus"^
   -D cppad_static_lib=TRUE^
   ..
nmake check

Note that when using windows DLL’s, CppAD builds a static version of cppad_lib . There are problems using a DLL for cppad_lib because Windows makes separate copies of static class member functions, one for library and one for the rest of the program.

autotools

The autotools build with the Visual Studio compiler should work with the following configure and test commands

mkdir build
cd build
../configure CC=cl CXX=cl CXX_FLAGS="-DEBUG"
make test

An optimized versions was tested using

../configure CXX=cl CC=cl CXX_FLAGS="-DNDEBUG -O2 -EHsc"
make test

The commands above need to be run a Dos window, that has the vcvarsall.bat settings, extend path that includes the msys2 bin directory, and was running the bash shell inside the Doc window. It is highly recommended that you use the cmake commands above, an not ../configure , when building with Visual Studio.

make check

Important information about the CppAD configuration is output by this command. If you have the grep program, and store the output in cmake.log , you can get a list of all the test options with the command:

grep 'make check' cmake.log

cmake_verbose_makefile

This value should be either true or false . The default value, when it is not present, is false . If it is true , then the output of the make commands will include all of the files and flags used to run the compiler and linker. This can be useful for seeing how to compile and link your own applications.

cmake_build_type

This value should be one of the valid CMake build types; e.g., Debug , Release , RelWithDebInfo , MinSizeRel . The value of this option is not case sensitive; e.g., Debug and debug yield the same result. If cmake_build_type is specified, cppad_debug_which must be the empty string.

generator

The CMake program is capable of generating different kinds of files. Below is a table with a few of the possible files

generator

Description

"Unix Makefiles"

make files for unix operating system

"NMake Makefiles"

make files for Visual Studio

"MSYS Makefiles"

make files for msys2

Ninja

Ninja build system

Other generator choices are available; see the cmake generators documentation.

cppad_prefix

This is the top level absolute path below which all of the CppAD files are installed by the command

make install

For example, if cppad_prefix is /usr , cmake_install_includedirs is include , and cppad_postfix is not specified, the file cppad.hpp is installed in the location

/ usr/include/cppad/cppad.hpp

The default value for cppad_prefix is the value of CMAKE_INSTALL_PREFIX ; see the cmake documentation. (Before 2019-10-02 the default value was /usr ; see the heading 10-02 on 2019 .)

cppad_postfix

This is the bottom level relative path below which all of the CppAD files are installed. For example, if cppad_prefix is /usr , cmake_install_includedirs is include , and cppad_postfix is coin-or , the file cppad.hpp is installed in the location

/ usr/include/coin-or/cppad/cppad.hpp

The default value for cppad_postfix is empty; i.e, there is no bottom level relative directory for the installed files.

cmake_install_includedirs

This is one directory, or a list of directories separated by spaces or by semi-colons. This first entry in the list is the middle level relative path below which the CppAD include files are installed. The entire list is used for searching for include files. For example, if cppad_prefix is /usr , cmake_install_includedirs is include , and cppad_postfix is not specified, the file cppad.hpp is installed in the location

/ usr/include/cppad/cppad.hpp

The default value for this directory list is include .

cmake_install_libdirs

This is one directory, or a list of directories separated by spaces or by semi-colons. This first entry in the list is the middle level relative path below which the CppAD library files are installed. The entire list is used for searching for library files. The default value for this directory list is lib .

cppad_lib

As an example of where the CppAD library is installed, if cppad_prefix is /usr , cmake_install_libdirs is lib , cppad_postfix is not specified, the CppAD library is installed in

/ usr/lib/libcppad_lib. ext . major . release

Here major is the major library version number, release is the release number for this version, and ext is the extension for shared libraries on this system. If yyyymmdd is the CppAD Version number, the major CppAD library version number is

major = dd - 1 + 31 * ( mm - 1 + 12 * ( yyyy - 2019 ) )

If there is no Release for this version of CppAD, the CppAD library is installed in

/ usr/lib/libcppad_lib. ext . major

The Mac is special, ext is dylib and comes at the end of the file name for that system.

cmake_install_datadir

This is the middle level relative path below which the CppAD data files are installed. For example, if cppad_prefix is /usr , cmake_install_datadir is share , and cppad_postfix is not specified, the pkgconfig file cppad.pc is installed in the location

/ usr/share/pkgconfig/cppad.pc

The default value for cmake_install_datadir is share .

cmake_install_docdir

This is the middle level relative path below which the CppAD user documentation files are installed. For example, if cppad_prefix is /usr , cmake_install_docdir is share/doc , and cppad_postfix is not specified, the file user_guide.html is installed in the location

/ usr/share/doc/cppad/user_guide.html

If cmake_install_docdir is not specified, or include_doc is false, the documentation files are not installed.

cmake_defined_ok

The default value for this option is true. It must be true when running cmake with a non-empty CMakeCache.txt file. If it is false, some checks will be made to make sure certain variables are not defined twice.

include_doc

The default value for this option is false. If it is true, is in the command line, the xrst program must be installed. In this case, the user and developer documentation targets doc_user and doc_dev will be available. For example, if the generator is Ninja the commands ninja doc_user and ninja doc_dev will build the corresponding documentation in the build/html sub-directory of the Distribution Directory .

cppad_static_lib

The default value for this option is true on msys, cygwin and Windows and false otherwise. the cppad library will be a static (shared) library. If the target system uses windows dlls, this setting should be true and if it is false, a warning will be issued and it will be ignored.

cppad_debug_and_release

The default value for this option is true. If it is true, the debug and release versions of CppAD can be mixed in the same program. This must be true if cppad_debug_which is debug_even or debug_odd . If NDEBUG is defined and cppad_debug_and_release is false, thread_alloc does more error checking. For programs that do a lot of memory allocation, this can take a significant amount of time. This is meant for testing CppAD and as a last resort when debugging.

include_adolc

The default value for this option is false. If it is true, the adolc examples will be compiled and tested. In this case, adolc.pc must be in the PKG_CONFIG_PATH .

include_ipopt

The default value for this option is false. If it is true, the ipopt examples will be compiled and tested. In this case, ipopt.pc must be in the PKG_CONFIG_PATH . In addition, ipopt_solve and cppad_ipopt_nlp will be installed.

include_cppadcg

The default value for this option is false. If it is true, the cppadcg examples will be compiled and tested. Warning : Do not use this option when installing cppad because the cppadcg package depends on cppad and using this option makes cppad depend on cppadcg. This option, and the script get_cppadcg.sh are only intended for testing purposes.

package_prefix

Each of these packages do not have pkg-config files corresponding to optional CppAD examples. If a prefix listed below is on the command line, the corresponding examples will be compiled and run:

cppad_cxx_flags

This specifies the compiler flags that are used when compiling the CppAD examples, tests, and library. This flags are in addition to the flags automatically generated by cmake for debug and release build; i.e., CMAKE_CXX_FLAGS_DEBUG and CMAKE_CXX_FLAGS_RELEASE . The default value for these flags is the empty string "" . These flags must be valid for the C++ compiler on your system. For example, if you are using g++ you could specify

-D cppad_cxx_flags="-Wall -ansi -pedantic-errors -std=c++17 -Wshadow"

C++17

In order for the compiler to take advantage of features that are in C++17 ,but not in C++11, the cppad_cxx_flags must enable these features.

cppad_profile_flag

This specifies an addition compiler and link flag that is used for profiling the speed tests. A profile version of the speed test is only build when this argument is present. If this flag is specified, you can run the gnu profiler with the following command:

      cd build/speed/profile
      make check_speed_profile
      ./ speed_profile test seed option_list
      gprof -b speed_profile | sed -f gprof.sed

cppad_testvector

See Choosing the CppAD Test Vector Template Class.

cppad_max_num_threads

The value cppad_max_num_threads must be greater than or equal to four; i.e., max_num_threads >= 4 . The current default value for cppad_max_num_threads is 48, but it may change in future versions of CppAD. The value cppad_max_num_threads in turn specifies the default value for the preprocessor symbol CPPAD_MAX_NUM_THREADS .

cppad_tape_id_type

The type cppad_tape_id_type is used for identifying different tapes. The valid values for this type are unsigned char , unsigned short int , unsigned int , and size_t . The smaller the value of sizeof ( cppad_tape_id_type ) , the less memory is used. On the other hand, the value

std::numeric_limits< cppad_tape_id_type >:: max ()

must be larger than the maximum number of tapes used by one thread times CPPAD_MAX_NUM_THREADS .

cstdint

If all of the following cstdint types are defined, they can also be used as the value of cppad_tape_addr_type : uint8_t , uint16_t , uint32_t , uint64_t .

cppad_tape_addr_type

The type cppad_tape_addr_type is used for address in the AD recordings (tapes). The valid values for this argument are unsigned char , unsigned short int , unsigned int , size_t . The smaller the value of sizeof ( cppad_tape_addr_type ) , the less memory is used. On the other hand, the value

std::numeric_limits< cppad_tape_addr_type >:: max ()

must be larger than any of the following: size_op , size_op_arg , size_par , size_text , size_VecAD .

cstdint

If all of the following cstdint types are defined, they can also be used as the value of cppad_tape_addr_type : uint8_t , uint16_t , uint32_t , uint64_t .

cppad_debug_which

All of the CppAD examples and test can optionally be tested in debug or release mode (see exception below). This option controls which mode is chosen for the corresponding files. The value cppad_debug_which be one of the following:

cppad_debug_which

CMAKE_BUILD_TYPE

debug_all

Debug

debug_none

Release

debug_even

not specified

debug_odd

not specified

empty string

not changed

If CMAKE_BUILD_TYPE is specified on the command line, then cppad_debug_which must be the empty string (its default value).