1.1. Copyright and License

This document, Scientific Computing with free software on GNU/Linux HOWTO, is copyrighted (c) 2002 by Manoj Warrier. Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.1 or any later version published by the Free Software Foundation; with no Invariant Sections, with no Front-Cover Texts, and with no Back-Cover Texts. A copy of the license is available here

1.2. Disclaimer

No liability for the contents of this document is accepted. Use of the concepts, examples, links and information is entirely at your own risk. There may be errors and inaccuracies, that could damage your system, waste your time, etc... Proceed with caution, and although this is unlikely, the author takes no responsibility whatsoever.

All copyrights are held by their respective owners, unless specifically noted otherwise. Use of a term in this document should not be regarded as affecting the validity of any trademark or service mark. Naming of particular products, software or brands should not be seen as endorsements

I have not used many of the software applications to which links are provided. There are simply too many applications that do the same thing, that one cannot be expected to have used all of them. In a book on Scientific Computing using GNU/Linux, one would mention ones favorite tool to carry out a task and describe it in detail. However this is a howto providing links to various available free tools for scientific computing and may contain links to some software that promises much but delivers little and vice versa.

1.3. Motivation

This howto mainly consists of the links provided at http://Scilinux.sourceforge.net which has to be disbanded due to a name conflict. The best alternative seems to be to make it a Linux document and host it at the LDP site. Another reason is that there seems to be many free software applications doing the same things. We hope to provide links to the available software thereby making it easy for the scientific community to make a choice without spending much time.

1.4. Credits / Contributors

In this document, I have the pleasure of acknowledging:

• Linus Trovalds, Richard M. Stallman and their merry men for Linux, GNU and also for indirectly broadening various perspectives which were not really obvious.

• A host of colleagues and friends from the Institute for Plasma Research, India for discussions at various times.

• Marcel Bose, Ivan Lamouret, K. Scott Hunziker, Livine Christin, W. Herbert, Simon Pinches and many others for suggesting various links mentioned in this document.

• Vasudha my wife for letting me do what I wish and egging me on with comments like "let us hope that you will finish at least this project"

Shishir and Ashoka are co-authors of this document because such a collection of links was Shishir's idea and Ashoka is always contributing by providing links, suggestions and a second point of view. They will be helping me maintain this HOWTO too.

1.5. Feedback

Feedback is most certainly welcome for this document. Send your additions, comments and criticisms to the following email address : <m_war at users.sourceforge.net>.

1.6. Translations

No translations yet.

3.1. Programming Languages

Links are provided to various compilers used in Scientific Computing like FORTRAN, C, C++, Java and more recently Python.

• GNU Compiler Collection : GNU's project to produce a world class optimizing compiler. It works on multiple architectures and diverse environments. Currently GCC contains front ends for C, C++, Objective C, GNU Fortran-95, Java, and Ada, as well as libraries for these languages (libstdc++, libgcj,..).

  For manuals on using the various GCC compilers check out The GCC online documentation

• g77 : The GCC front end for FORTRAN 77. It is a very good FORTRAN77 compiler. It however does not have the -r8 option which compiles a program as double precision. This could be a good compiler design philosophy but in many cases gives problems when porting a code from SUN / DEC / HP workstations onto Linux systems. The g77 manual is available at The Gcc Online documentation site.

• gfortran. I was happy to receive this link by mail. It was 3 years since I had migrated to using the GNU C compiler for scientific computing because there was no "truly free" FORTRAN-95 compiler available then. I thank Paul Thomas for this link.

• g95. gfortran above and g95 are reportedly offshoots from the same CVS tree. Has an impressive list of programs that compiles and runs using this compiler.

• fort77 and f2c: fort77 is a perl program which invokes the f2c command (a Fortran to C translator) transparently, so it can be used just like a real Fortran compiler. Fort77 can be used to compile Fortran, C and assembler code and can link the code with f2c libraries. If you install fort77, you'll also need to install the f2c package. This does not have the "-r8" problem. You can download fort77 and f2c from the above link.

• lush: An object-oriented programming language, which combines the flexibility of an interpretive language, with the efficiency of a compiled language. It has full interfaces to numerical libraries (GSL, LAPACK, BLAS), graphics libraries (OpenGL), which allows creation of graphics and 3D animations and many other features that sound too good to be true. I have not yet tried this out, but it sounds very promising.

• Scientific Python: You may want to explore Python for your scientific computing needs. Python is an interpreted, interactive, object-oriented programming language. It has a number of extensions for numerics, plotting, data storage and combined with Tk lets you develop very good GUIs for your codes. The most exciting aspect is that it simplifies programming because it has modules for almost anything (vectors, tensors, transformations, derivatives, linear algebra, Fourier transforms, statistics, etc ...) are available. You can also wrap C and Fortran libraries from Python. Finally if you want to write a numerical scheme of your own you may find that it is simpler in Python. There are also interfaces to netCDF (portable binary files), MPI and BSPlib (parallel programming).

  You can further explore Python for Scientific computing here:

  • Scientific-Python: A collection of modules for scientific computing on Python. All the necessary modules can be downloaded as either a tar file or an RPM file from here. The maintainer Konrad HINSEN also has a nice tutorial on Scientific Computing in Python.

  • SciPy An open source library of scientific tools for Python. It includes modules for graphics and plotting, optimization, integration, special functions, signal and image processing, genetic algorithms, ODE solvers, etc.

3.2. Debugging Tools

In this section links are given to mainly debugging tools for GCC and FORTRAN. I understand that python has a debugging module built in though I have not used it. The purpose of a debugger is to allow you to see what is going on inside a program while it executes or what the program was doing when/if it crashed.

• Ftnchek: A FORTRAN checker designed to detect errors in a Fortran program that a compiler usually does not. Therefore it is best to run ftnchek on your FORTRAN programs after it has compiled without errors. Its purpose is to assist the user in finding semantic errors. Semantic errors are legal in the Fortran language but are wasteful or may cause incorrect operation. An on-line manual is available. This project is looking for volunteers to bringing ftnchek up to the Fortran 90 standard.

• gdb : All programs written in the languages supported by GCC can be debugged using gdb, an excellent interactive, command line debugger. You can compile your programs using a -g option which then compiles your code with debugging information inserted into the executable. It can start your programs, stop your programs on specified conditions and at specified locations, examine what happened when your program stops. In a large code with multiple cascading calls to various functions it can back trace the function calls. You can also Download the document Debugging with GDB and a quick reference card.

• xxgdb: It is a front end to the gdb debugger. Useful for beginners to gdb as it lists out the whole gdb commands as buttons with a area for viewing source on which one can include break points, etc by a click of the mouse, and another area for viewing the debugging results.

• DDD: The GNU Data Display Debugger, GNU DDD, is a graphical front-end for command-line debuggers such as GDB, DBX, WDB, Ladebug, JDB, XDB, the Perl debugger, or the Python debugger. Besides ``usual'' front-end features such as viewing source texts it also has a good interactive graphical data display, where data structures are displayed as graphs. Follow this link for a DDD manual in postscript / HTML / PDF format.

3.3. Version Control Tools

It will be worth your while investing some time in learning to use one of the version control tools below (cvs is what I use ..) if you are into any serious code development.

• Concurrent Versions System : CVS is one of the most popular version control systems running on the Linux operating system. Popular Linux projects like Apache, EGCS, GIMP, and others are using CVS to coordinate their efforts ... This is how the URL linked above describes their effort.

  A tutorial on CVS is available at Gentoo Linux Documentation and a free CVS book is available here

• Project Revision Control System : PRCS, the Project Revision Control System, is the front end to a set of tools that (like CVS) provide a way to deal with sets of files and directories as an entity, preserving coherent versions of the entire set. PRCS was designed primarily by Paul N. Hilfinger, with input and modifications by Luigi Semenzato and Josh MacDonald. PRCS is written and maintained by Josh MacDonald. Its purpose is similar to that of SCCS, RCS, and CVS, but (according to its authors, at least), it is much simpler than any of those systems. This page is where information on the latest developments in PRCS can be found.

• Gbuild : gbuild is a script written in the Bourne shell language to simplify package maintenance by allowing you to automate code update from CVS, compilation, building tar/rpms/srpms of your package. some external scripts which certain functions of gbuild depend on are written in Perl. gbuild is released under the GPL.

3.4. Integrated Development Environments

Integrated development environments (IDEs) can be very useful for building code and ideally come with all the above tools (i.e a compiler, a debugger and a version control tool). In addition to that IDEs also usually provide a makefile generator, documenting help, online help manuals, etc.

• Kdeveloper : A easy to use C/C++ IDE (Integrated Development Environment) for Linux. It supports KDE/Qt, GNOME, plain C and C++ projects. This site has a lot of documentation ..... a highly browsable site for software developers. Specifically, KDevelop manages or provides:

  All development tools needed for C++ programming like Compiler, Linker, automake and autoconf; KAppWizard, which generates complete, ready-to-go sample applications; Class generator, for creating new classes and integrating them into the current project; File management for sources, headers, documentation etc. to be included in the project; The creation of User-Handbooks written with SGML and the automatic generation of HTML-output with the KDE look and feel; Automatic HTML-based API-documentation for your project's classes with cross-references to the used libraries; Internationalization support for your application, allowing translators to easily add their target language to a project; KDevelop also includes WYSIWYG (What you see is what you get)-creation of user interfaces with a built-in dialog editor; Debugging your application by integrating KDbg; Editing of project-specific pixmaps with KIconEdit; The inclusion of any other program you need for development by adding it to the "Tools"-menu according to your individual needs.

• VDKbuilder: VDKbuilder is a tool that helps programmers in constructing GUI interfaces, editing, compiling, linking, and debugging within an integrated environment. Using VDKBuilder dramatically reduces developing time since all code related to GUI construction and signal processing is automatically generated, maintained and updated. It is distributed under the GNU Public License. Visit the site for downloading the software.

5.1. Repositories

• Netlib: An amazing amount of free source code for Numerical Methods. Netlib is THE source code repository which contains an innumerable amount of source code for Numerical Methods. It also has an active discussion forum wherein you can submit your queries and stay posted for help. Netlib also has a Parallel Tools Library and a search by subject.

• GAMS: Guide to Available Mathematical Software GAMS has a very useful search using which one can search for keywords (example: ``diffusion'' to search for a diffusion equation solver). However the browse by package at GAMS reveals that a lot of the software they provide is a link to the netlib repository.

• Object Oriented Numerics A site devoted to object oriented numerics. It has a Mailing list, Extensive Links to freely available libraries (OO of course) and freely available tools for object oriented scientific computing.

• GNU Scientific Library The GNU Scientific Library (GSL) is a collection of numerical routines written from scratch in C. It provides an Applications Programming Interface (API) for C programmers and also allows wrappers to be written for very high level languages. It covers a wide range of numerical computing topics, has a good manual, is widely portable and is distributed under the GNU General Public License.

• GiNaC GiNaC is designed to allow the creation of software which need symbolic manipulations embedded in them. It extends C++ by a set of algebraic capabilities and is recursively named for GiNaC is not a Computer Algebra system. It is distributed under the terms and conditions of the GNU general public license (GPL).

5.2. Other topic specific numerical libraries

• FFTW FFTW is a collection of fast C routines for computing the Discrete Fourier Transform in one or more dimensions. It includes complex, real, and parallel transforms, and can handle arbitrary array sizes efficiently. This package includes both the double- and single-precision FFTW uniprocessors and the threads libraries.

• LAPACK LAPACK (Linear Algebra PACKage) is a standard library for numerical linear algebra. LAPACK provides routines for solving systems of simultaneous linear equations, least-squares solutions of linear systems of equations, eigenvalue problems, and singular value problems. LAPACK is coded in Fortran77 and is built with egcs. It is well documented and widely used (and therefore widely tested).

• SuperLU SuperLU is a general purpose library which performs an LU decomposition for the direct solution of large, sparse, non-symmetric systems of linear equations on high performance machines. Its written in C and is callable from either C or Fortran.

• ARPACK ARPACK is a set of Fortran77 subroutines designed to solve large scale eigenvalue problems. A Users Guide is available. The above link also gives information about a parallel version of ARPACK - PARPACK and a object oriented version ARPACK++.

• Computational Fluid Dynamics codes This link contains a comprehensive listing of public domain, shareware and freeware Computational Fluid Dynamics codes links with a description of each CFD code.

7.1. The GNU/Linux Workstation

As with most documentation related to GNU/Linux, the Linux Documentation project's home page is a priceless source. You might first want to read The Linux Installation HOWTO. For those who want to install Linux along with Windows might want to browse through The Linux + Windows HOWTO. When installing Linux make sure that you choose to install all documentation. After installing Linux, a good, comprehensive document to getting started with using Linux is The Rute Users Tutorial and Exposition which is a beginners guide to Linux and UNIX like systems. I'd like to give a less intimidating (size-wise) link to a small beginners guide, but U will find this useful after taking the plunge. You might also want to go through The Linux System Administrator's Guide and to check out The Linux Administration Made Easy (LAME) guide It attempts to describe day-to-day administration and maintenance issues commonly faced by Linux system administrators.

7.2. Parallel Processing and Symmetric Multiprocessing: Supercomputing

It is possible to get large volume number crunching without spending millions of rupees on a supercomputer. You only need to link together (by some high speed network) the requisite number of CPUs, with GNU/LINUX as the underlying OS. Add some freely available message passing software and a effective parallel processing number crunching machine is made. Such clusters are called "Beowulf clusters". The other advantages of such a cluster other than building costs is, up-gradation costs are minimal. The two best resources for Linux cluster builders are

• The Beowulf Project home page and

• The Extreme Linux Project

8.1. Word Processing Tools

• Latex: LaTeX is a high-quality typesetting system, with features designed for the production of technical and scientific documentation. LaTeX is the de facto standard for the communication and publication of scientific documents. David R.Wilkin's primer "Getting Started with LaTeX" is a good tutorial to getting started with LaTeX. For those who have to live with a WYSIWYG documenting tool, check LyX. This is a front-end for latex. It isn't as powerful as latex proper, but helps with a good WYSIWIG.

• Lout: A document formatting system similar to latex. Good features, documentation and history. Light weight and outputs postscript. Thanks to Emiliano Gavilan for this link.

• Abiword: As the AbiWord home page says, "AbiWord is suitable for typing papers, letters, reports, memos, and so forth". It has won many awards and seems to be the best open source WYSIWYG word processor. Check out the above link to know more about it and download it.

• kword: As the kword home page says, "KWord is a FrameMaker-like word-processing and desktop publishing application. KWord is capable of creating demanding and professional looking documents. Whether you are a corporate or home user, production artist or student, KWord will prove a valuable and easy to use tool for all your word processing and layout needs". Check out the above link to know more about it and download it. (you might want to know more about the whole koffice suite).

8.2. Poster presentation tools

• KPresenter: KPresenter is the presentation tool of the KOffice suite of office utilities. It allows screen presentations with all the trappings one is used to seeing in costly presentation tools. It also allows honest, real scientific presentations where one does not have to impress the audience with non subject specific stuff :-). The best thing about it is the possibility of saving the presentation as a html file. It makes portable network graphics files with each presentation slide. With a smattering of knowledge of html files one could put in a animated image as a image link thereby allowing one to show movies too when necessary.

• Xfig : Though the man page claims that it is a facility for the interactive generation of figures ...., It in fact much more than that. Other than generating figures for elucidating what you want to say in a poster, it helps you import and export figures in a variety of formats, write text in various fonts and sizes, generate Greek symbols and color text, Save as latex picture file or any other format supported by your word processor for inclusion in your publications, generate GIFs of each page of the poster to put on your web site, and finally it generates *.fig files which are small in size. The only thing on my wish list for xfig is the capability to edit the imported pictures which are not in *.fig format. Therefore for a computer screen projected poster presentation you need a frames capable browser with contents in one frame and the xfig generated posters (exported as *.png or *.jpg from Xfig) on the other.