Newer Software on Cluster

It is highly recommended to used the LCG softwares for better combination of different software sets.

Software from LHC Computing Grid(LCG)

CERN packs a lots of HEP softwares such as ROOT and Geant4 into releases and distributes them via CVMFS. Under /cvmfs/sft.cern.ch/lcg/views/ you can find directories start with LCG_ followed by the major release version. For most releases there is also a Python3 version. For example, LCG_98 is the major release 98 with Python2, and LCG_98python3 is the Python3 version. Let’s say we want use release 98 with Python3 and gcc9:

testuser@bl-0:~$ source /cvmfs/sft.cern.ch/lcg/views/LCG_98python3/x86_64-centos7-gcc9-opt/setup.sh
testuser@bl-0:~$ python --version
Python 3.7.6
testuser@bl-0:~$ gcc --version | head -n1
gcc (GCC) 9.2.0
testuser@bl-0:~$ root --version | head -n1
ROOT Version: 6.22/00

For users requires old ROOT 5 installation, the best choice is LCG_71 or LCG_72a:

[chenxun@bl-0 views]$ source /cvmfs/sft.cern.ch/lcg/views/LCG_72a/x86_64-slc6-gcc49-opt/setup.sh
[chenxun@bl-0 views]$ which root
/cvmfs/sft.cern.ch/lcg/views/LCG_72a/x86_64-slc6-gcc49-opt/bin/root
[chenxun@bl-0 views]$ root -b
  *******************************************
  *                                         *
  *        W E L C O M E  to  R O O T       *
  *                                         *
  *   Version   5.34/25   12 January 2015   *
  *                                         *
  *  You are welcome to visit our Web site  *
  *          http://root.cern.ch            *
  *                                         *
  *******************************************

  ROOT 5.34/25 (v5-34-25@v5-34-25, Jan 12 2015, 15:29:06 on linuxx8664gcc)

http://lcginfo.cern.ch/ lists all the packages in LCG release.

Python

Newer version of Python and related packages such as scipy and jupyter notebook also available through anaconda.

The advantage of anaconda over LCG is most of its packages is on local disk, LCG releases on the other hand need fetch from Internet therefor take a few minutes to start when load for the first time. To use anaconda, please see Using Python on Cluster.

GNU Compiler Collection(gcc)

CentOS provides recent versions of GNU Compiler Collection, GNU Debugger, and other development, debugging, and performance monitoring tools via Developer Toolset. To use gcc8:

$ source scl_source enable devtoolset-8

Similar for gcc9:

$ source scl_source enable devtoolset-9