Monstergauss

Monstergauss was started in the 1970's, derived from early versions of parts of the Gaussian 80/82 series of programs. The aims were to have a program that was easily portable to a variety of computer systems, and which incorporated features not found in the Gaussian program suite.

It features analytic gradient optimization of geometries, plus various other options such as Restricted Hartree-Fock open shell SCF, pair-wise MC-SCF, Boys localization, one electron properties, quantitative PMO analysis, a direct CI, the Morokuma Energy Decomposition, and a basis set superposition error correction by the counterpoise method.

Monstergauss has been designed to be compatible, with minimal changes, with IBM and CDC mainframes, GOULD minicomputers, the DEC VAX minicomputers, and Unix-based systems (e.g. Sun and Apollo workstations).

Note: almost all the files here are compressed, and must be downloaded and uncompressed ('zcat FILENAME | tar xpf -') to be viewed. Most of the '.doc' files are in LaTeX format, but the basic text should be easy to read.

• Source code and test cases:
  • src.tar.Z - source code for Monstergauss and the support programs (size: 1081773 bytes).
  • test.tar.Z - 30 test cases (sample input and expected output) (size: 685677 bytes).
    The entire set of tests should be run and checked against the expected output before trusting any results produced by the program!
  • If you encounter problems with COMMON BLOCK sizes when compiling 'gl0903.f', make the following "fix":

    diff gl0903.f.old gl0903.f
    86c86
    <       INTEGER*2 MD1,MD2
    ---
    >       INTEGER*2 MD1,MD2,I2B,I2C
    93c93
    <       COMMON/C903B/MD1(MA1,MV1)
    ---
    >       COMMON/C903B/MD1(MA1,MV1),I2B
    95c95
    <       COMMON/C903C/MD2(MA1,MV1)
    ---
    >       COMMON/C903C/MD2(MA1,MV1),I2C
    370c370
    <       INTEGER*2 MD1,MD2
    ---
    >       INTEGER*2 MD1,MD2,I2B,I2C
    376c376
    <       COMMON/C903B/MD1(MA1,MV1)
    ---
    >       COMMON/C903B/MD1(MA1,MV1),I2B
    378c378
    <       COMMON/C903C/MD2(MA1,MV1)
    ---
    >       COMMON/C903C/MD2(MA1,MV1),I2C
    

    The problem with the COMMON blocks in 'gl0903.f' is that the variables in those COMMONs are INTEGER*2, and due to the values of MA and MV, it turns out the total size of the arrays MD1 and MD2 may be an odd number of INTEGER*2 values. Some 'f77' compilers round up the size so the COMMON block ends on a full word (INTEGER*4) boundary, which is not being done when the compiler sees and checks the size of the second use of the same COMMON block. Basically this is a compiler bug/error (it should round up before checking the size), but there is almost no chance it will be fixed.
    The real fix is to ensure that the product of MA1 and MV1 is an even number, so the COMMON blocks end on full word boundary. This means that some combinations of the 'MA' and 'MV' parameters (in particular the ones used in the example program conversion scripts) can't be used.
  • A problem with the CI was reported many years ago, so it is possible the fix above will not make everything work properly. I could never figure out where the problem was, but it is almost certainly some variables that need to have 'SAVE' specified or that need to be set to zero at compile time. Probably the easiest way to find it would be to compile all the 'gl09*.f' files with automatic SAVE / ZERO specified at compile time and make sure the program works properly. Then go back and recompile one file at a time without the automatic SAVE / ZERO, and see when it stops working; then try SAVE and ZERO separately to figure out which is the problem. Finding the variables that need SAVE / ZERO after that will be pretty hard.
  • Alan Liska from the J. Heyrovsky Institute of Physical Chemistry ('alan.liska at jh-inst.cas.cz') has ported Monstergauss to Windows and Android. You can get the files at:
    • Mobile Chemistry Portal
    • Monstergauss
    • Mobile Chemistry GitHub Repository
    • Standalone App On Google Play

• Documentation files:
  • gauss.doc.Z - Monstergauss installation notes (LaTeX version) (size: 59093 bytes).
  • gauss.out.Z - Monstergauss installation notes (plain text) (size: 61035 bytes).
  • gauss.l.Z - Unix 'man' page for 'mgauss'.
  • manual.out.Z - Monstergauss Users Manual (printable, includes FORTRAN carriage control) (size: 99651 bytes).

• Source code conversion programs:
  • convrt.doc.Z - source code conversion request form (LaTeX version).
    This is no longer needed as anyone with access to a Unix system can convert the source code themselves.
  • convrt.out.Z - source code conversion request form (plain text).
    This is no longer needed as anyone with access to a Unix system can convert the source code themselves.
  • convrt80.job - C shell script to convert Monstergauss to handle 80 basis functions on an Gould system.
  • convrt150.job - C shell script to convert Monstergauss to handle 150 basis functions on an IBM system.
  • convrt200.job - C shell script to convert Monstergauss to handle 200 basis functions on an IBM system.
  • convrt250.job - C shell script to convert Monstergauss to handle 250 basis functions on an IBM system.
  • convrtunx.job - C shell script to convert Monstergauss to handle more/fewer basis functions on a Unix system (this conversion is done "in-place", where the original source code is replaced by the new version).
  • convrtvax.job - C shell script to convert Monstergauss to handle more/fewer basis functions on a VAX system (this conversion is done "in-place", where the original source code is replaced by the new version).
  • maildir - Shell script to e-mail all files in the current directory.
  • manual.job - C shell script to prepare the Monstergauss Users Manual, using the source files (which contain all the documentation as comments).
  • tartape.job - Korn shell script to copy the source code to a tape drive or disk file (this is a good file to look at to see what all the other files contain and what they do).

• Basis Set Database:
  • basis1.data - H/He basis sets (size: 8385 bytes).
  • basis2a.data - Li/C basis sets (size: 54259 bytes).
  • basis2b.data - N/Ne basis sets (size: 70249 bytes).
  • basis3.data - Na/Ar basis sets (size: 92707 bytes).
  • basis4a.data - K/Mn basis sets (size: 49571 bytes).
  • basis4b.data - Fe/Kr basis sets (size: 54419 bytes).
  • basis5.data - Rb/Xe basis sets (size: 26363 bytes).
  • basis6.data - Cs/Lw basis sets (size: 22215 bytes).

  NOTE: These files are named '*.data' so they can be downloaded more easily, but once downloaded, they should be renamed to '*.dat' so they can be found automatically when needed.

• Other notes (not needed to install/run the program):
  • overlay.gould.doc.Z - conversion notes for FORTRAN 77 to Gould FORTX32, including a description of the overlay structure of the program (LaTeX version) (size: 20842 bytes).
  • rhfpaper.doc.Z - a paper on the 'Calculation of State Parameters for Open Shell Restricted Hartree-Fock Self-Consistent Field Theory' (LaTeX version) (size: 17433 bytes).

See also ChemCompute: computational chemistry software for undergraduate teaching and research. Draw a molecule and calculate.