Please ensure that CERE's dependencies are installed on your system:
-
LLVM 7.0 +, https://releases.llvm.org/download.html#7.0.0
-
gcc
-
GNU binutils (objcopy, readelf)
-
GNU autotools (automake, autoconf >= 2.69)
-
Python 3, and the following modules: matplotlib, jinja2, networkx, numpy, pydot, pygraphviz, sklearn
-
Graphviz, http://www.graphviz.org
-
gperftools, http://code.google.com/p/gperftools/
-
numactl and dc (needed when running the testsuite)
Then run the following command inside CERE directory:
$ ./autogen.sh
$ ./configure --without-flang
$ make
$ make install
The --with-flang
flag can be used to enable compilation of Fortran programs.
If needed, LLVM path, and gcc path can be configured with the following options:
$ ./configure --with-llvm=<path to llvm install directory> \
--without-flang \
CC=<path to gcc>
Once installation is over, we recommend that you run the test suite to ensure CERE works as expected on your system:
$ make check
For example on an Ubuntu x86_64 14.04 release, you should use the following install procedure:
$ wget http://llvm.org/releases/7.0.0/clang+llvm-7.0.0-x86_64-linux-gnu-ubuntu-14.04.tar.xz
tar -xf clang+llvm-7.0.0-x86_64-linux-gnu-ubuntu-14.04.tar.xz
export PATH=$PWD/clang+llvm-7.0.0-x86_64-linux-gnu-ubuntu-14.04/bin/:$PATH
export LIBRARY_PATH=$PWD/clang+llvm-7.0.0-x86_64-linux-gnu-ubuntu-14.04/lib/:$LIBRARY_PATH
export LD_LIBRARY_PATH=/usr/local/lib:$LD_LIBRARY_PATH
$ sudo apt-get install autoconf automake build-essential libtool ruby-ronn \
python python-matplotlib python-jinja2 python-pygraphviz python-pydot python-sklearn python-numpy \
python-networkx python-pip google-perftools libgoogle-perftools-dev numactl dc libiomp-dev
$ sudo ln -s /usr/lib/libiomp5.so /usr/lib/libomp.so
$ sudo pip install pulp
$ cd cere/
$ ./autogen.sh
$ ./configure --without-flang
$ make && make install
$ make check