Actions

Difference between revisions of "Accessing software"

From ALICE Documentation

 
Line 1: Line 1:
[[Category: About ALICE]]
 
 
= Software  =
 
= Software  =
 
Before we start using individual software packages, we need to understand why multiple versions of the software are available on ALICE systems and why users need to have a way to control which version they are using. The three biggest factors are:
 
Before we start using individual software packages, we need to understand why multiple versions of the software are available on ALICE systems and why users need to have a way to control which version they are using. The three biggest factors are:
Line 14: Line 13:
 
{{:Environment modules}}
 
{{:Environment modules}}
 
[[Category:Advanced]]
 
[[Category:Advanced]]
 +
[[Category:About ALICE]]

Latest revision as of 12:04, 29 June 2020

Software

Before we start using individual software packages, we need to understand why multiple versions of the software are available on ALICE systems and why users need to have a way to control which version they are using. The three biggest factors are:

  • software incompatibilities;
  • versioning;
  • dependencies.

Software incompatibility is a major headache for programmers. Sometimes the presence (or absence) of a software package will break others that depend on it. Two of the most famous examples are Python 2 and 3 and C compiler versions. Python 3 famously provides a python command that conflicts with that provided by Python 2. Software compiled against a newer version of the C libraries and then used when they are not present will result in a nasty 'GLIBCXX_3.4.20' not found error, for instance.

Software versioning is another common issue. A team might depend on a certain package version for their research project - if the software version was to change (for instance, if a package was updated), it might affect their results. Having access to multiple software versions allow a set of researchers to prevent software versioning issues from affecting their results.

Dependencies are where a particular software package (or even a particular version) depends on having access to another software package (or even a particular version of another software package). For example, the VASP materials science software may depend on having a particular version of the FFTW (Fastest Fourier Transform in the West) software library available for it to work. Environment modules are the solution to these problems.

Environment modules

A module is a self-contained description of a software package - it contains the settings required to run a software package and, usually, encodes required dependencies on other software packages.

There are a number of different environment module implementations commonly used on HPC systems: the two most common are TCL modules and Lmod. Both of these use similar syntax and the concepts are the same so learning to use one will allow you to use whichever is installed on the system you are using. In both implementations the module command is used to interact with environment modules. An additional sub-command is usually added to the command to specify what you want to do. For a list of sub-commands you can use module -h or module help. As for all commands, you can access the full help on the man pages with man module.

List of currently loaded modules

On login, you usually start out with a default set of modules.

The module list command shows which modules you currently have loaded in your environment. This is what you will most likely see after logging into ALICE.

  [me@nodelogin02~]$ module list
  
  Currently Loaded Modules:
    1) shared   2) DefaultModules   3) gcc/8.2.0   4) slurm/19.05.1

You can see that by default the module for Slurm and the gcc compiler is loaded. The number behind the slash sign represent the version number.

List available modules

To see the available modules, use module -d avail. With the -d option, you will only get the default versions of the modules. For various software packages, there are also older/other versions available, that might be used if necessary. You can see all version by omitting the -d option.

 [me@nodelogin02~]$ module -d avail
  ------------------------------------------------ /cm/shared/easybuild/modules/all -------------------------------------------------
  AMUSE-Miniconda2/4.7.10                                         VTK/8.2.0-foss-2019b-Python-3.7.4
  AMUSE-VADER/12.0.0-foss-2018a-Python-2.7.14                     WebKitGTK+/2.24.1-GCC-8.2.0-2.31.1
  AMUSE/12.0.0-foss-2018a-Python-2.7.14                           X11/20190717-GCCcore-8.3.0
  ATK/2.32.0-GCCcore-8.2.0                                        XML-Parser/2.44_01-GCCcore-7.3.0-Perl-5.28.0
  Autoconf/2.69-GCCcore-8.3.0                                     XZ/5.2.4-GCCcore-8.3.0
  Automake/1.16.1-GCCcore-8.3.0                                   Yasm/1.3.0-GCCcore-8.3.0
  Autotools/20180311-GCCcore-8.3.0                                ZeroMQ/4.3.2-GCCcore-8.2.0
  Bazel/0.20.0-GCCcore-8.2.0                                      amuse-framework/12.0.0-foss-2018a-Python-2.7.14
  Bison/3.3.2                                                     at-spi2-atk/2.32.0-GCCcore-8.2.0
  Boost.Python/1.67.0-foss-2018b-Python-3.6.6                     at-spi2-core/2.32.0-GCCcore-8.2.0
  Boost/1.71.0-gompi-2019b                                        binutils/2.32
  etc etc etc .......

If you are searching for a specific software package or tool you can search for the full module name like this:

 [me@nodelogin02~]$ module avail python
  ---------------------------------------- /cm/shared/easybuild/modules/all -----------------------------------------
  AMUSE-GPU/12.0.0-foss-2018a-Python-2.7.14
  AMUSE-VADER/12.0.0-foss-2018a-Python-2.7.14
  AMUSE/12.0.0-foss-2018a-Python-2.7.14
  AMUSE/13.1.0-foss-2018a-Python-3.6.4                    (D)
  Biopython/1.71-foss-2018a-Python-2.7.14
  Biopython/1.73-foss-2019a
  Biopython/1.75-foss-2019b-Python-3.7.4                  (D)
  Boost.Python/1.67.0-foss-2018b-Python-3.6.6
  CGAL/4.11.1-foss-2018b-Python-3.6.6
  CONCOCT/1.1.0-foss-2019a-Python-2.7.15
  CheckM/1.0.18-foss-2019a-Python-2.7.15
  Cython/0.25.2-foss-2018a-Python-2.7.14
  Cython/0.25.2-foss-2018a-Python-3.6.4
  Cython/0.29.3-foss-2019a-Python-3.7.2                   (D)
  DAS_Tool/1.1.1-foss-2019a-R-3.6.0-Python-3.7.2
  Docutils/0.9.1-foss-2018a-Python-2.7.14
  Docutils/0.9.1-foss-2018a-Python-3.6.4                  (D)
  GObject-Introspection/1.60.1-GCCcore-8.2.0-Python-3.7.2
  IPython/5.7.0-foss-2018a-Python-2.7.14
  IPython/6.4.0-foss-2018a-Python-3.6.4
  IPython/7.7.0-foss-2019a-Python-3.7.2                   (D)
  Keras/2.2.4-foss-2019a-Python-3.7.2
  Mako/1.0.7-foss-2017b-Python-2.7.14
  Mako/1.0.7-foss-2018a-Python-2.7.14
  Mako/1.0.7-foss-2018b-Python-2.7.15
  Meson/0.50.0-GCCcore-8.2.0-Python-3.7.2
  Meson/0.51.2-GCCcore-8.3.0-Python-3.7.4                 (D)
  MultiQC/1.7-foss-2018b-Python-3.6.6
  NLTK/3.2.4-foss-2018a-Python-3.6.4
  NLTK/3.2.4-foss-2019a-Python-3.7.2                      (D)
  PyCairo/1.18.0-foss-2018b-Python-3.6.6
  PyTorch/1.3.1-fosscuda-2019b-Python-3.7.4
  PyYAML/3.13-foss-2018b-Python-3.6.6
  Python/2.7.14-foss-2017b
  Python/2.7.14-foss-2018a
  Python/2.7.14-GCCcore-6.4.0-bare
  Python/2.7.15-foss-2018b
  Python/2.7.15-GCCcore-7.3.0-bare
  Python/2.7.15-GCCcore-8.2.0
  Python/2.7.16-GCCcore-8.3.0
  Python/3.6.4-foss-2018a
  Python/3.6.6-foss-2018b
  Python/3.7.2-GCCcore-8.2.0
  Python/3.7.4-GCCcore-8.3.0                              (D)
  etc etc etc .......


In the above output, you can see that there are modules with the flag "(D)". This indicates that this is the default module for a software package for which modules for different versions exists.

If you want to get more information about a specific module, you can use the whatis sub-command:

  [me@nodelogin02~]$ module whatis Python/3.7.4-GCCcore-8.3.0
  Python/3.7.4-GCCcore-8.3.0                            : Description: Python is a programming language that lets you work more quickly and 
  integrate your systems more effectively.
  Python/3.7.4-GCCcore-8.3.0                            : Homepage: https://python.org/
  Python/3.7.4-GCCcore-8.3.0                            : URL: https://python.org/
  Python/3.7.4-GCCcore-8.3.0                            : Extensions: alabaster-0.7.12, asn1crypto-0.24.0, atomicwrites-1.3.0, attrs-19.1.0, Babel- 
  2.7.0, bcrypt-3.1.7, bitstring-3.1.6, blist-1.3.6, certifi-2019.9.11, cffi-1.12.3, chardet-3.0.4, Click-7.0, cryptography-2.7, Cython-0.29.13, 
  deap-1.3.0, decorator-4.4.0, docopt-0.6.2, docutils-0.15.2, ecdsa-0.13.2, future-0.17.1, idna-2.8, imagesize-1.1.0, importlib_metadata-0.22, 
  ipaddress-1.0.22, Jinja2-2.10.1, joblib-0.13.2, liac-arff-2.4.0, MarkupSafe-1.1.1, mock-3.0.5, more-itertools-7.2.0, netaddr-0.7.19, netifaces- 
  0.10.9, nose-1.3.7, packaging-19.1, paramiko-2.6.0, pathlib2-2.3.4, paycheck-1.0.2, pbr-5.4.3, pip-19.2.3, pluggy-0.13.0, psutil-5.6.3, py-1.8.0, 
  py_expression_eval-0.3.9, pyasn1-0.4.7, pycparser-2.19, pycrypto-2.6.1, Pygments-2.4.2, PyNaCl-1.3.0, pyparsing-2.4.2, pytest-5.1.2, python- 
  dateutil-2.8.0, pytz-2019.2, requests-2.22.0, scandir-1.10.0, setuptools-41.2.0, setuptools_scm-3.3.3, six-1.12.0, snowballstemmer-1.9.1, Sphinx- 
  2.2.0, sphinxcontrib-applehelp-1.0.1, sphinxcontrib-devhelp-1.0.1, sphinxcontrib-htmlhelp-1.0.2, sphinxcontrib-jsmath-1.0.1, sphinxcontrib-qthelp- 
  1.0.2, sphinxcontrib-serializinghtml-1.1.3, sphinxcontrib-websupport-1.1.2, tabulate-0.8.3, ujson-1.35, urllib3-1.25.3, virtualenv-16.7.5, 
  wcwidth-0.1.7, wheel-0.33.6, xlrd-1.2.0, zipp-0.6.0

Load modules

To load a software module, use module load. In the example below, we will use Python 3.

Initially, Python 3 is not loaded and therefore not available for use. We can test this by using the command which that looks for programs the same way that Bash does. We can use it to tell us where a particular piece of software is stored.

  [me@nodelogin01~]$ which python3
  /usr/bin/which: no python3 in     (/cm/shared/apps/slurm/18.08.4/sbin:/cm/shared/apps/slurm/18.08.4/bin:/cm/local/apps/gcc/8.2.0/bin:/usr/local/bin:/usr/bin:/usr/local/sbin:/usr/sbin:/home/me/.local/bin:/home/me/bin)


We can load the python3 command with module load:

  [me@nodelogin01 ~]$ module load Python/3.7.2-GCCcore-8.2.0
  [me@nodelogin01 ~]$ which python3
  /cm/shared/easybuild/software/Python/3.7.4-GCCcore-8.3.0/bin/python3 

So what just happened? To understand the output, first we need to understand the nature of the $PATH environment variable. $PATH is a special environment variable that controls where a Linux operating system (OS) looks for software. Specifically $PATH is a list of directories (separated by :) that the OS searches through for a command. As with all environment variables, we can print it using echo.

  [me@nodelogin01 ~]$ echo $PATH
  /cm/shared/easybuild/software/Python/3.7.2-GCCcore-8.2.0/bin:/cm/shared/easybuild/software/XZ/5.2.4-GCCcore-  8.2.0/bin:/cm/shared/easybuild/software/SQLite/3.27.2-GCCcore-8.2.0/bin:/cm/shared/easybuild/software/Tcl/8.6.9-GCCcore-8.2.0/bin:/cm/shared/easybuild/software/libreadline/8.0-GCCcore-8.2.0/bin:/cm/shared/easybuild/software/ncurses/6.1-GCCcore-8.2.0/bin:/cm/shared/easybuild/software/bzip2/1.0.6- GCCcore-8.2.0/bin:/cm/shared/easybuild/software/GCCcore/8.2.0/bin:/cm/shared/apps/slurm/19.05.1/sbin:/cm/shared/apps/slurm/19.05.1/bin:/cm/local/apps/gcc/8.2.0/bin:/usr/local/bin:/usr/bin:/usr/local/sbin:/usr/sbin:/home/me/.local/bin:/home/me/bin

You will notice a similarity to the output of the which command. In this case, there’s only one difference: the /cm/shared/easybuild/software/Python/3.7.2-GCCcore-8.2.0/bin directory at the beginning.

When we used module load Python/3.7.2-GCCcore-8.2.0, it added this directory to the beginning of our $PATH. Let us examine what is there:

   [me@nodelogin01 ~]$ ls /cm/shared/easybuild/software/Python/3.7.2-GCCcore-8.2.0/bin
   2to3		  futurize	 pip	      pytest		 python3-config  rst2odt_prepstyles.py	sphinx-apidoc
   2to3-3.7	  idle3		 pip3	      py.test		 pyvenv		 rst2odt.py		sphinx-autogen
   chardetect	  idle3.7	 pip3.7       python		 pyvenv-3.7	 rst2pseudoxml.py	sphinx-build
   cygdb	  netaddr	 pybabel      python3		 rst2html4.py	 rst2s5.py		sphinx-quickstart
   cython	  nosetests	 __pycache__  python3.7		 rst2html5.py	 rst2xetex.py		tabulate
   cythonize	  nosetests-3.7  pydoc3       python3.7-config	 rst2html.py	 rst2xml.py		virtualenv
   easy_install	  pasteurize	 pydoc3.7     python3.7m	 rst2latex.py	 rstpep2html.py		wheel
   easy_install-3.7  pbr	 pygmentize   python3.7m-config  rst2man.py	 runxlrd.py


Taking this to its conclusion, module load adds software to your $PATH. It “loads” software.

A special note on this, module load will also load required software dependencies. If you compare the output below with what you had when you first logged in to ALICE, you will notice several other modules have been load automatically, because the Python module depends on them.

   [me@nodelogin01 ~]$ module list
   Currently Loaded Modules:
     1) shared           5) GCCcore/8.2.0               9) libreadline/8.0-GCCcore-8.2.0  13) GMP/6.1.2-GCCcore-8.2.0
     2) DefaultModules   6) bzip2/1.0.6-GCCcore-8.2.0  10) Tcl/8.6.9-GCCcore-8.2.0        14) libffi/3.2.1-GCCcore-8.2.0
     3) gcc/8.2.0        7) zlib/1.2.11-GCCcore-8.2.0  11) SQLite/3.27.2-GCCcore-8.2.0    15) Python/3.7.2-GCCcore-8.2.0
     4) slurm/19.05.1    8) ncurses/6.1-GCCcore-8.2.0  12) XZ/5.2.4-GCCcore-8.2.0

Also a note of warning: When you load several modules, it is possible that their dependencies can cause conflicts and problems later on. It is best to always check what other modules have been automatically loaded.

Unload modules

The command module unload “un-loads” a module. For the above example:

 [me@nodelogin01 ~]$ module unload Python/3.7.2-GCCcore-8.2.0
 [me@nodelogin01 ~]$ module list
 Currently Loaded Modules:
   1) shared           5) GCCcore/8.2.0               9) libreadline/8.0-GCCcore-8.2.0  13) GMP/6.1.2-GCCcore-8.2.0
   2) DefaultModules   6) bzip2/1.0.6-GCCcore-8.2.0  10) Tcl/8.6.9-GCCcore-8.2.0        14) libffi/3.2.1-GCCcore-8.2.0
   3) gcc/8.2.0        7) zlib/1.2.11-GCCcore-8.2.0  11) SQLite/3.27.2-GCCcore-8.2.0
   4) slurm/18.08.4    8) ncurses/6.1-GCCcore-8.2.0  12) XZ/5.2.4-GCCcore-8.2.0

Important: Currently, unloading a module does not unload its dependencies (as you can see from the above output).


If you want to remove all the modules that are currently loaded, you can use the command module purge:

 [me@nodelogin01 ~]$ module purge
 [me@nodelogin01 ~]$ module list
 No modules loaded


Note that this command will also unload the modules loaded by default on login including Slurm. You can either manually load the modules back or source your bashrc with source ~/.bashrc.