Siesta FAQ

This is a short FAQ about the computational Quantum Chemistry code "Siesta" on the HPCVL clusters. HPCVL requires all users of this software to sign a statement. The software is available on all nodes of the cluster, but can only be accessed by persons who belong to a specific Unix group. See details below. Note that this FAQ is incomplete and only discusses how to sign up for Siesta usage.

  1. What is "Siesta"?
  2. Where is Siesta located and how do I access it?
  3. How do I use Siesta?
  4. How do I set up and submit Siesta batch jobs?
  5. Where can I get further help?

  1. What is "Siesta"?

    SIESTA (Spanish Initiative for Electronic Simulations with Thousands of Atoms) is both a method and its computer program implementation, to perform electronic structure calculations and ab initio molecular dynamics simulations of molecules and solids. It is mainly used for linear-scaling ab-initio calculations for large and complex systems. At present, both serial and MPI-parallel versions of the program are installed, but the parallel version encounters problems which we are working to resolve.

    Back to Top...

  2. Where is Siesta located and how do I access it?

     The present version of Siesta is 2.0.2. The programs in the Siesta package reside in /opt/siesta/2.0.2. There are two subdirectories called "32" and "64" which contain the 32-bit and 64-bit version of the program, respectively. If you encounter memory limitations with the 32-bit version, please try the 64-bit alternative.

    To use Siesta on the HPCVL clusters, it is required that you read our licensing agreement and sign a statement. You will then be made a member of a Unix group siesta, which enables you to run the software.

    Back to Top...

  3. How do I run Siesta interactively? The program is run by typing 

    /opt/siesta/2.0.2/32/serial-siesta < input.fdf
    for the 32-bit version, or 
    /opt/siesta/2.0.2/64/serial-siesta < input.fdf
    for the 64-bit version. Here, the input.fdf file is meant to be the input file for a Siesta run. To find out how to make an input file, please check the Siesta Manual.

    [...This FAQ is in progress. Please check later...]

  4. How do I set up and run a parallel Siesta batch job?

    While both a serial version serial-siesta and an MPI-parallel version mpi-siesta are installed, only the former works at this point. Please contact us for suggestions on how get multiple Siesta runs started within a single program submission.

    The MPI version of Siesta is run through the mpirun command: 

    mpirun -np 8 /opt/siesta/2.0.2/32/mpi-siesta < input.fdf
    for the 32-bit version, or 
    mpirun -np 8 /opt/siesta/2.0.2/64/mpi-siesta < input.fdf
    for 64 bit. This example uses input.fdf as an input file and runs using 8 processes.

    [...This FAQ is in progress. Please check later...]


    Back to Top...

  5. Where can I get further help?

     A User's Guides for Siesta is available here. The software provider of Siesta operates a very informative website with lots of information, including examples, manuals, FAQ's, etc. If you have problems with the Grid Engine, read our Grid Engine FAQ on that subject, and maybe consult the manual for that software which is accessible as a PDF file. HPCVL also provides user support in the case of technical problems. Contact us here.

 
 
   
© HPCVL 2009