Skip to content

Latest commit

 

History

History
230 lines (172 loc) · 7.71 KB

SYSTEM_INSTALL.md

File metadata and controls

230 lines (172 loc) · 7.71 KB
Raw

Installing GooFit on different systems

The following commands show you how to get a minimal install of GooFit on a vanilla system; you will probably want to add ROOT for your system, and possibly CUDA if you have a graphics card. If you do not have ROOT, some functionality, such as the Minuit1 version of the fitter, will not be available, and most of the examples will not be included in the build. Since ROOT is such a common addition, some of the system also have a note on installing ROOT.

CentOS 7: (click to expand)

For simplicity, this uses EPEL to get access to python-pip, and uses the pip version of CMake. Feel free to download CMake directly from Kitware instead. If you want to use Docker, you can start with docker run --rm -it centos. You can also use this recipe with an nvidia-docker CentOS image.

yum install epel-release -y
yum install python-pip git gcc-c++ make -y
pip install cmake plumbum
git clone --recursive https://github.com/GooFit/GooFit.git
cd GooFit
mkdir build
cd build
cmake ..
make
make test

If you'd like to add ROOT, add the following lines before running CMake:

mkdir root-6 && curl https://root.cern.ch/download/root_v6.14.02.Linux-centos7-x86_64-gcc4.8.tar.gz | tar --strip-components=1 -xz -C root-6
source root-6/bin/thisroot.sh

Alpine Linux 3.8: (click to expand)

A truly minimal system, Alpine gives you a working Docker system under 3 MB. Since it is unlikely that you'll be running Alpine outside of Docker, the Docker command is included.

docker run --rm -it alpine
apk add make cmake g++ git libexecinfo-dev
git clone --recursive https://github.com/GooFit/GooFit.git
cd GooFit
mkdir build
cd build
cmake ..
make
ctest

In the spirit of minimality, this is less instructive and contains more magic, but also would also work:

docker run --rm -it alpine
apk add build-base cmake git
git clone https://github.com/GooFit/GooFit.git
cd GooFit
make auto

If you'd like to use the Python version:

docker run --rm -it alpine
apk add python3-dev build-base cmake ninja git libexecinfo-dev
pip3 install scikit-build
pip3 -v install git+https://github.com/GooFit/GooFit.git

If you want Python2 instead, either add the py2-pip package or the line python2 -m ensurepip. Python 3 comes with Pip in Alpine. You can also pip install -v goofit if you want the latest released PyPI version.

Clang-style checking on Alpine (click to expand)

If you'd like to use LLVM's clang-format to check the style, probably the easiest way to do that is with Docker and Alpine. The following lines will run the style check for you:

docker run --rm -it alpine
apk add clang git
git clone https://github.com/GooFit/GooFit.git
cd GooFit
./scripts/check_style.sh

Ubuntu 18.04 LTS (click to expand)

Ubiquitous Ubuntu works also. Ubuntu was used for the NVidia Docker solution due to better support from NVidia. The following example uses ninja-build instead of make, but make works if you prefer it. You should also be able to use this recipe with docker run -it ubuntu or an nvidia-docker Ubuntu image if you don't have Ubuntu installed.

apt update && apt install -y git cmake ninja-build g++
git clone --recursive https://github.com/GooFit/GooFit.git
cd GooFit
mkdir build
cd build
cmake ..
cmake --build .
ctest

If you want Python bindings, add python-dev or python3-dev to the list.

If you'd like to add ROOT, add the following lines before running CMake:

mkdir root-6 && curl https://root.cern.ch/download/root_v6.14.02.Linux-ubuntu18-x86_64-gcc7.3.tar.gz | tar --strip-components=1 -xz -C root-6
source root-6/bin/thisroot.sh

OpenSUSE (click to expand)

If you use make, adding -jN where N is the number of cores will make builds much faster on multicore systems! ninja does this automatically. If you'd like to use Docker to provide an OpenSUSE environment, add docker run -it opensuse sh to the beginning of these commands.

zypper install -y git cmake gcc-c++
git clone --recursive https://github.com/GooFit/GooFit.git
cd GooFit
make

If you'd like to add ROOT:

As always, https://root.cern.ch/build-prerequisites#opensuse is a good resource, though glu-devel seems to now be required as well. I'm using ninja to build in parallel automatically (and it seems to be faster as well).

zypper install -y git bash cmake gcc-c++ gcc binutils xorg-x11-libX11-devel xorg-x11-libXpm-devel xorg-x11-devel xorg-x11-proto-devel xorg-x11-libXext-devel glu-devel
zypper install -y ninja tar
git clone --depth=1 --branch=v6-14-02 http://github.com/root-project/root.git root_git
mkdir root_build
cd root_build
cmake ../root_git -GNinja -DCMAKE_INSTALL_PREFIX=/opt/root-6-14-02
cmake --build . --target install

Then, you can activated that copy of root with:

source /opt/root-6-14-02/bin/thisroot.sh

You might want to add useful extra ROOT library: -Droofit=ON -Dmathmore=ON -Dminuit2=ON

Docker (click to expand)

If you are interested in actually running on Docker, you can use the official GooFit Docker images:

docker run --rm -it goofit/goofit-omp
nvidia-docker run --rm -it goofit/goofit-cuda

The CUDA version will need to build on your computer; the OMP version is prebuilt. You can also start with the ROOT Docker instance:

docker run --rm -it rootproject/root-ubuntu16 bash
cd
git clone --recursive https://github.com/GooFit/GooFit.git
cd GooFit
make

SLC 6 with CVMFS (LxPlus 6): (click to expand)

# If you have not run this already (automatic on LxPlus):
source /cvmfs/lhcb.cern.ch/group_login.sh

# Set up LCG releases
. /cvmfs/sft.cern.ch/lcg/views/LCG_94python3/x86_64-slc6-gcc62-opt/setup.sh

# Download and build as usual (ssh download doesn't seem to
# work properly due to bug on LxPlus)
git clone --recursive https://github.com/GooFit/GooFit.git
cd GooFit
mkdir build
cd build
cmake ..
make -j4

NVidia (click to expand)

docker run --rm -v $PWD:/goofit -it nvidia/cuda:9.2-devel-ubuntu18.04 bash
apt update && apt install -y cmake
mkdir build
cd build
cmake ../goofit -DGOOFIT_ARCH=3.5
make -j4

Or, to select a newer version of CMake:

docker run --rm -v $PWD:/goofit -it nvidia/cuda:9.2-devel-ubuntu18.04 bash
apt update && apt install -y wget
wget -qO- "https://cmake.org/files/v3.16/cmake-3.16.3-Linux-x86_64.tar.gz" | tar --strip-components=1 -xz -C /usr/local
cmake -S goofit -B build -DGOOFIT_ARCH=3.5
cmake --build build -j4

Note about installing CMake: (click to expand)

While other install methods for CMake, like pip, are easier, this way should always work. On Linux, you can manually get a version of CMake using:

mkdir cmake && wget -qO- "https://cmake.org/files/v3.16/cmake-3.16.3-Linux-x86_64.tar.gz" | tar --strip-components=1 -xz -C cmake
export PATH=`pwd`/cmake/bin:$PATH

The second line will need to be rerun whenever use a new shell. Feel free to make your updated CMake default; CMake is insanely backward compatible and will even "dumb itself down" when it sees a lower version in the minimum_required line in CMakeLists.txt.

If you are a fan of using ~/.local and already have ~/.local/bin in your path, you can instead use:

wget -qO- "https://cmake.org/files/v3.16/cmake-3.16.3-Linux-x86_64.tar.gz" | tar --strip-components=1 -xz -C ~/.local