[R] Enable OpenMP with AppleClang in XGBoost R package (#5240)

* [R] Enable OpenMP with AppleClang in XGBoost R package

* Dramatically simplify install doc

Makefile

@@ -261,6 +261,8 @@ Rpack: clean_all
 	sed -i -e 's/-pthread/$$\(SHLIB_PTHREAD_FLAGS\)/g' xgboost/src/Makevars.win
 	sed -i -e 's/@ENDIAN_FLAG@/-DDMLC_CMAKE_LITTLE_ENDIAN=1/g' xgboost/src/Makevars.win
 	sed -i -e 's/@BACKTRACE_LIB@//g' xgboost/src/Makevars.win
+	sed -i -e 's/@OPENMP_LIB@//g' xgboost/src/Makevars.win
+	rm -f xgboost/src/Makevars.win-e   # OSX sed create this extra file; remove it
 	bash R-package/remove_warning_suppression_pragma.sh
 	rm xgboost/remove_warning_suppression_pragma.sh
 

R-package/configure

@@ -585,6 +585,7 @@ ac_subst_vars='LTLIBOBJS
 LIBOBJS
 BACKTRACE_LIB
 ENDIAN_FLAG
+OPENMP_LIB
 OPENMP_CXXFLAGS
 OBJEXT
 EXEEXT
@@ -612,7 +613,6 @@ infodir
 docdir
 oldincludedir
 includedir
-runstatedir
 localstatedir
 sharedstatedir
 sysconfdir
@@ -682,7 +682,6 @@ datadir='${datarootdir}'
 sysconfdir='${prefix}/etc'
 sharedstatedir='${prefix}/com'
 localstatedir='${prefix}/var'
-runstatedir='${localstatedir}/run'
 includedir='${prefix}/include'
 oldincludedir='/usr/include'
 docdir='${datarootdir}/doc/${PACKAGE_TARNAME}'
@@ -935,15 +934,6 @@ do
   | -silent | --silent | --silen | --sile | --sil)
     silent=yes ;;
 
-  -runstatedir | --runstatedir | --runstatedi | --runstated \
-  | --runstate | --runstat | --runsta | --runst | --runs \
-  | --run | --ru | --r)
-    ac_prev=runstatedir ;;
-  -runstatedir=* | --runstatedir=* | --runstatedi=* | --runstated=* \
-  | --runstate=* | --runstat=* | --runsta=* | --runst=* | --runs=* \
-  | --run=* | --ru=* | --r=*)
-    runstatedir=$ac_optarg ;;
-
   -sbindir | --sbindir | --sbindi | --sbind | --sbin | --sbi | --sb)
     ac_prev=sbindir ;;
   -sbindir=* | --sbindir=* | --sbindi=* | --sbind=* | --sbin=* \
@@ -1081,7 +1071,7 @@ fi
 for ac_var in	exec_prefix prefix bindir sbindir libexecdir datarootdir \
 		datadir sysconfdir sharedstatedir localstatedir includedir \
 		oldincludedir docdir infodir htmldir dvidir pdfdir psdir \
-		libdir localedir mandir runstatedir
+		libdir localedir mandir
 do
   eval ac_val=\$$ac_var
   # Remove trailing slashes.
@@ -1234,7 +1224,6 @@ Fine tuning of the installation directories:
   --sysconfdir=DIR        read-only single-machine data [PREFIX/etc]
   --sharedstatedir=DIR    modifiable architecture-independent data [PREFIX/com]
   --localstatedir=DIR     modifiable single-machine data [PREFIX/var]
-  --runstatedir=DIR       modifiable per-process data [LOCALSTATEDIR/run]
   --libdir=DIR            object code libraries [EPREFIX/lib]
   --includedir=DIR        C header files [PREFIX/include]
   --oldincludedir=DIR     C header files for non-gcc [/usr/include]
@@ -1827,11 +1816,7 @@ ac_compiler_gnu=$ac_cv_c_compiler_gnu
 
 
 
-### Check whether backtrace() is part of libc or the external lib libexecinfo
+# Use this line to set CC variable to a C compiler
-{ $as_echo "$as_me:${as_lineno-$LINENO}: checking Backtrace lib" >&5
-$as_echo_n "checking Backtrace lib... " >&6; }
-{ $as_echo "$as_me:${as_lineno-$LINENO}: result: " >&5
-$as_echo "" >&6; }
 ac_ext=c
 ac_cpp='$CPP $CPPFLAGS'
 ac_compile='$CC -c $CFLAGS $CPPFLAGS conftest.$ac_ext >&5'
@@ -2622,6 +2607,12 @@ ac_link='$CC -o conftest$ac_exeext $CFLAGS $CPPFLAGS $LDFLAGS conftest.$ac_ext $
 ac_compiler_gnu=$ac_cv_c_compiler_gnu
 
 
+### Check whether backtrace() is part of libc or the external lib libexecinfo
+{ $as_echo "$as_me:${as_lineno-$LINENO}: checking Backtrace lib" >&5
+$as_echo_n "checking Backtrace lib... " >&6; }
+{ $as_echo "$as_me:${as_lineno-$LINENO}: result: " >&5
+$as_echo "" >&6; }
+
 { $as_echo "$as_me:${as_lineno-$LINENO}: checking for backtrace in -lexecinfo" >&5
 $as_echo_n "checking for backtrace in -lexecinfo... " >&6; }
 if ${ac_cv_lib_execinfo_backtrace+:} false; then :
@@ -2706,7 +2697,8 @@ fi
 
 if test `uname -s` = "Darwin"
 then
-  OPENMP_CXXFLAGS="\$(SHLIB_OPENMP_CXXFLAGS)"
+  OPENMP_CXXFLAGS='-Xclang -fopenmp'
+  OPENMP_LIB='/usr/local/lib/libomp.dylib'
   ac_pkg_openmp=no
   { $as_echo "$as_me:${as_lineno-$LINENO}: checking whether OpenMP will work in a package" >&5
 $as_echo_n "checking whether OpenMP will work in a package... " >&6; }
@@ -2716,22 +2708,29 @@ $as_echo_n "checking whether OpenMP will work in a package... " >&6; }
 int
 main ()
 {
- return omp_get_num_threads ();
+ return (omp_get_max_threads() <= 1);
   ;
   return 0;
 }
 _ACEOF
-  PKG_CFLAGS="${OPENMP_CFLAGS}" PKG_LIBS="${OPENMP_CFLAGS}" "$RBIN" CMD SHLIB conftest.c 1>&5 2>&5 && "$RBIN" --vanilla -q -e "dyn.load(paste('conftest',.Platform\$dynlib.ext,sep=''))" 1>&5 2>&5 && ac_pkg_openmp=yes
+  ${CC} -o conftest conftest.c /usr/local/lib/libomp.dylib -Xclang -fopenmp 2>/dev/null && ./conftest && ac_pkg_openmp=yes
   { $as_echo "$as_me:${as_lineno-$LINENO}: result: ${ac_pkg_openmp}" >&5
 $as_echo "${ac_pkg_openmp}" >&6; }
   if test "${ac_pkg_openmp}" = no; then
     OPENMP_CXXFLAGS=''
+    OPENMP_LIB=''
+    echo '*****************************************************************************************'
+    echo 'WARNING: OpenMP is unavailable on this Mac OSX system. Training speed may be suboptimal.'
+    echo '         To use all CPU cores for training jobs, you should install OpenMP by running\n'
+    echo '             brew install libomp'
+    echo '*****************************************************************************************'
   fi
 fi
 
 
 
 
+
 ac_config_files="$ac_config_files src/Makevars"
 
 cat >confcache <<\_ACEOF

R-package/configure.ac

@@ -4,6 +4,9 @@ AC_PREREQ(2.62)
 
 AC_INIT([xgboost],[0.6-3],[],[xgboost],[])
 
+# Use this line to set CC variable to a C compiler
+AC_PROG_CC
+
 ### Check whether backtrace() is part of libc or the external lib libexecinfo
 AC_MSG_CHECKING([Backtrace lib])
 AC_MSG_RESULT([])
@@ -25,19 +28,26 @@ fi
 
 if test `uname -s` = "Darwin"
 then
-  OPENMP_CXXFLAGS="\$(SHLIB_OPENMP_CXXFLAGS)"
+  OPENMP_CXXFLAGS='-Xclang -fopenmp'
+  OPENMP_LIB='/usr/local/lib/libomp.dylib'
   ac_pkg_openmp=no
   AC_MSG_CHECKING([whether OpenMP will work in a package])
-  AC_LANG_CONFTEST(
+  AC_LANG_CONFTEST([AC_LANG_PROGRAM([[#include <omp.h>]], [[ return (omp_get_max_threads() <= 1); ]])])
-  [AC_LANG_PROGRAM([[#include <omp.h>]], [[ return omp_get_num_threads (); ]])])
+  ${CC} -o conftest conftest.c /usr/local/lib/libomp.dylib -Xclang -fopenmp 2>/dev/null && ./conftest && ac_pkg_openmp=yes
-  PKG_CFLAGS="${OPENMP_CFLAGS}" PKG_LIBS="${OPENMP_CFLAGS}" "$RBIN" CMD SHLIB conftest.c 1>&AS_MESSAGE_LOG_FD 2>&AS_MESSAGE_LOG_FD && "$RBIN" --vanilla -q -e "dyn.load(paste('conftest',.Platform\$dynlib.ext,sep=''))" 1>&AS_MESSAGE_LOG_FD 2>&AS_MESSAGE_LOG_FD && ac_pkg_openmp=yes
   AC_MSG_RESULT([${ac_pkg_openmp}])
   if test "${ac_pkg_openmp}" = no; then
     OPENMP_CXXFLAGS=''
+    OPENMP_LIB=''
+    echo '*****************************************************************************************'
+    echo 'WARNING: OpenMP is unavailable on this Mac OSX system. Training speed may be suboptimal.'
+    echo '         To use all CPU cores for training jobs, you should install OpenMP by running\n'
+    echo '             brew install libomp'
+    echo '*****************************************************************************************'
   fi
 fi
 
 AC_SUBST(OPENMP_CXXFLAGS)
+AC_SUBST(OPENMP_LIB)
 AC_SUBST(ENDIAN_FLAG)
 AC_SUBST(BACKTRACE_LIB)
 AC_CONFIG_FILES([src/Makevars])

R-package/src/Makevars.in

@@ -18,7 +18,7 @@ $(foreach v, $(XGB_RFLAGS), $(warning $(v)))
 
 PKG_CPPFLAGS=  -I$(PKGROOT)/include -I$(PKGROOT)/dmlc-core/include -I$(PKGROOT)/rabit/include -I$(PKGROOT) $(XGB_RFLAGS)
 PKG_CXXFLAGS= @OPENMP_CXXFLAGS@ @ENDIAN_FLAG@ -pthread
-PKG_LIBS = @OPENMP_CXXFLAGS@ @ENDIAN_FLAG@ @BACKTRACE_LIB@ -pthread
+PKG_LIBS = @OPENMP_CXXFLAGS@ @OPENMP_LIB@ @ENDIAN_FLAG@ @BACKTRACE_LIB@ -pthread
 OBJECTS= ./xgboost_R.o ./xgboost_custom.o ./xgboost_assert.o ./init.o\
          $(PKGROOT)/amalgamation/xgboost-all0.o $(PKGROOT)/amalgamation/dmlc-minimum0.o\
          $(PKGROOT)/rabit/src/engine_empty.o $(PKGROOT)/rabit/src/c_api.o

doc/build.rst

@@ -316,13 +316,21 @@ R Package Installation
 Installing pre-packaged version
 -------------------------------
 
-You can install xgboost from CRAN just like any other R package:
+You can install XGBoost from CRAN just like any other R package:
 
 .. code-block:: R
 
    install.packages("xgboost")
 
-For OSX users, single-threaded version will be installed. So only one thread will be used for training. To enable use of multiple threads (and utilize capacity of multi-core CPUs), see the section :ref:`osx_multithread` to install XGBoost from source.
+.. note:: Using all CPU cores (threads) on Mac OSX
+
+   If you are using Mac OSX, you should first install OpenMP library (``libomp``) by running
+
+   .. code-block:: bash
+
+      brew install libomp
+
+   and then run ``install.packages("xgboost")``. Without OpenMP, XGBoost will only use a single CPU core, leading to suboptimal training speed.
 
 Installing the development version
 ----------------------------------
@@ -339,48 +347,14 @@ Thus, one has to run git to check out the code first:
   cd xgboost
   git submodule init
   git submodule update
-  cd R-package
-  R CMD INSTALL .
-
-If the last line fails because of the error ``R: command not found``, it means that R was not set up to run from command line.
-In this case, just start R as you would normally do and run the following:
-
-.. code-block:: R
-
-  setwd('wherever/you/cloned/it/xgboost/R-package/')
-  install.packages('.', repos = NULL, type="source")
-
-The package could also be built and installed with CMake (and Visual C++ 2015 on Windows) using instructions from :ref:`r_gpu_support`, but without GPU support (omit the ``-DUSE_CUDA=ON`` cmake parameter).
-
-If all fails, try `Building the shared library`_ to see whether a problem is specific to R package or not.
-
-.. _osx_multithread:
-
-Installing R package on Mac OSX with multi-threading
-----------------------------------------------------
-
-First, obtain ``gcc-9`` and ``OpenMP`` with Homebrew (https://brew.sh/) to enable multi-threading (i.e. using multiple CPU threads for training). The default Apple Clang compiler does not support OpenMP, so using the default compiler would have disabled multi-threading.
-
-.. code-block:: bash
-
-  brew install gcc@9 libomp
-
-Now, clone the repository:
-
-.. code-block:: bash
-
-  git clone --recursive https://github.com/dmlc/xgboost
-
-Create the ``build/`` directory and invoke CMake with option ``R_LIB=ON``. Make sure to add ``CC=gcc-9 CXX=g++-9`` so that Homebrew GCC is selected. After invoking CMake, you can install the R package by running ``make`` and ``make install``:
-
-.. code-block:: bash
-
   mkdir build
   cd build
-  CC=gcc-9 CXX=g++-9 cmake .. -DR_LIB=ON
+  cmake .. -DR_LIB=ON
   make -j4
   make install
 
+If all fails, try `Building the shared library`_ to see whether a problem is specific to R package or not.
+
 .. _r_gpu_support:
 
 Installing R package with GPU support