Finding software and documentation

These web pages are intended as a starting guide for using palmetto. Each person will have additional interests or specific needs beyond the basics (e.g. special compiler options, certain application software) and to that end we provide some direction on how to find the software available to you and the documentation on your specific or advanced needs.

We can divide material into several areas:

Man pages
A great deal of documentation is available online via "man pages". Man pages are, as the name says, accessed through the Unix man command. Using the ls command as an example, you can get information about listing your files with

man ls

If you aren't sure of the item name you seek, you can use the keyword search such as

man -k fortran

to find information on fortran compilers, run time libraries, and so forth.

The Unix man command is actually a documentation formatting tool. The source documentation is in various locations across the file system. A Linux "environment variable" called $MANPATH is a default directory path that man searches to find documentation. Documentation for Linux commands and other CentOS applications and tools are in the default $MANPATH. (Type in echo $MANPATH to see the path.) Other paths are automatically loaded when you access certain compilers or software via the module package. (Reference to use of module will be made throughout these web pages as we document the use of particular software items.) In a few cases you may need to add a section parameter instead of modifying the path, as in

man 7 pbs_resources_linux.

Linux commands

The most commonly used Linux commands include

• ls - list files
• cat - list file contents
• more - page file contents
• rm - remove/delete files
• mkdir - make a new directory
• rmdir - remove an existing directory
• find - find files in a directory hierarchy
• locate - find files by name
• which - locate a file in user's path
• whereis - locate program by name

See their associated man pages for a complete description and information on tailoring their use.

Text editors

You will need a text editor to build PBS job submission scripts, modify your program code and input data, and generally view and change any text files you may use. Several well known text editors are available on user.

• vi - the "visual" editor. vi is a screen-oriented text editor. Two versions of interest are vim and gvim. vim is a "Much improved" version of vi. gvim is a GUI version of vim. For more information see man 1p vi and man vim.
• emacs - GNU Project Emacs. emacs, whose name comes from its early purpose as "Editor MACroS for the TECO editor", is a free, open source, and extensible editor. A favorite of programmers and system administrators, emacs is not a WYSIWYG (What You See Is What You Get) editor. See man emacs for more information.
• nano - "Nano's ANOther" editor. nano is a simple command line editor. It is an enhanced version of the pico editor. For more information, see man nano.

A Google search shows that many online tutorials are available for these editors as well.

Compilers

A number of Fortran, C, and C++ compilers are available for your use. Sources include the open source GNU compiler set as well as the high performance Intel compilers.

C Compilers

Fortran Compilers

Several versions of the GNU and Intel compilers are available to you. To see your choices, enter

module avail

--------------------------- /opt/Modules/versions ---------------------------
3.2.6

---------------------- /opt/Modules/3.2.6/modulefiles -----------------------
dot                gromacs/3.3.3      module-cvs         netcdf/4.0
fasta/35.4.2       gsl/1.10           module-info        null
fftw/2.1.5         hpl/1.0a           modules            openmpi/1.2.5
fftw/3.1.2         intel/10.0         mpiblast/1.4.0     openmpi/1.2.5-eth
fftw/3.2alpha3     intel/10.1         mpich/1.2.7        openmpi/1.2.7
fftw-mpi/2.1.5     intel/9.1          mpich/1.2.7-eth    pari/2.3.3
fluent/6.3.26      iprscan/4.3.1      mpich/1.2.7..7     paup/4b10
gaussian/g03       ls-dyna/971.R3.2.1 mpich2/1.0.6       perfsuite/0.6.2a6
gcc/3.4            matlab             mpich2/1.0.7       petsc/2.3.3-p8
gcc/4.1            mcnp5/1.20         mpich2/1.0.7-eth   pyMPI/2.5b0
gcc/4.3.2          mira/2.8.3         mpich2/1.0.7..2    pyMPI/2.5b0-eth
goto/1.25          mkl/10.0           namd/2.6           use.own
gromacs/3.3.1      mkl/8.1            ncbi/2.2.18

We need to let the operating system know the software that we want to run, which versions of that software, the location of the software, and so forth. Similar to the $MANPATH environment variable, the operating system searches for application and libraries that we use from a list in the $PATH and $LIBRARY environment variables. The defaults for these variables contain enough information to allow us to use the usual bash commands. The module command will supply the information about additional software we want to use.

So, to access the Intel version 10.1 compilers and their man pages for example, simply add the line

module add intel/10.1

to your .bashrc file (in your root directory.) Adding it to your .bashrc file ensures that access to the Intel compiler is given to the jobs you have running across palmetto nodes. Likewise, when using the AMD processors, be sure to include version 4.3.2 of the GNU Compiler Collection, as in

module add gcc/4.3.2

Application software and libraries

CITI provides various third party application software and libraries for your use. By entering

module avail

you can get an idea of what is available along with the current version numbers.

• Software
  

  R	R-mpi	abaqus	dot
  fasta	fftw	fftw-mpi	fluent
  gaussian	gcc	gems	goto
  gromacs	gsl	hpl	intel
  iprscan	libctl	ls-dyna	matlab
  mcnp5	meep	mira	mkl
  module-cvs	module-info	modules	mpb
  mpiblast	mpich	mpich2	namd
  ncbi	netcdf	null	openmpi
  pari	paup	perfsuite	petsc
  pgi	pgi32	pgi64	pvfs
  pyMPI	totalview	use.own

• Application descriptions
  • MATLAB® and companion toolboxes provide engineers, scientists, mathematicians, and educators with an environment for technical computing applications. See the MATLAB documentation pages for more information.
  • GROMACS is a versatile package to perform molecular dynamics, i.e. simulate the Newtonian equations of motion for systems with hundreds to millions of particles. Online man pages are available for each routine. A reference manual and tutorials can be found at GROMACs: Fast, Free, Flexible MD - Tutorial
  • PETSc, pronounced PET-see (the S is silent), is a suite of data structures and routines for the scalable (parallel) solution of scientific applications modeled by partial differential equations. It employs the MPI standard for all message-passing communication. See the PETSc Home Page for more information.
  • CHARM is a well know chemistry package from Princeton University. CHARM is available to those who have licensed it. Please contact us if you need more information about CHARM.
• Mathematical library descriptions

  The Intel Math Kernel Library (MKL) offers highly optimized, extensively threaded math routines for scientific, engineering, and financial applications that require maximum performance. MKL includes

  • Basic Linear Algebra Subprograms (BLAS)
  • Linear Algebra Package (LAPACK)
  • Scalable LAPACK (ScaLAPACK)
  • Vector Statistical Library
  • Vector Math Library
  • Fast Fourier Transforms

  See the Intel Math Kernel Library - Documentation web pages for documentation links.

• MPI library descriptions

  The Message Passing Interface (MPI) is a standardized set of procedures for communicating information and data between computers. It is commonly used to parallelize programs across a computing cluster. Two implementations of MPI are available to you: MPICH and OpenMPI.

  MPICH is available in two versions and two distributions. The versions are MPICH and MPICH2. The distributions are GNU and Intel. The compiler, your application, or the third party application you use will dictate which version of MPICH to use. Once you know that, you will add it via the module command

  module add <distribution> <version>

  (Remember to add this to the .bashrc file in your home directory.)

  OpenMPI, as the name implies, is an open source collaboration from across the High Performance Computing community. You will add it via the module command

  module add openmpi

  Documentation on MPI, MPICH, and OpenMPI can be found at

  Version	Web site links
  MPI	Message Passing Interface
  MPICH1	MPICH Home Page
  MPICH2	MPI: High-performance and Widely Portable MPI
  OpenMPI	Open MPI: Open Source High Performance Computing