Numerical simulation of condensed matter or fluid systems, can be done at several length scales. We discuss three and provide examples of public domain codes for electronic and atomic scale systems as well as a brief discussion of fluid modeling. The two codes to be presented in detail are QUANTUMESPRESSO for electronic properties and LAMMPS for molecular dynamics. Both are installed on TAMNUN. Especially for the electronic scale there are many alternatives, but these have been selected due to their public domain nature and to the availability of good tutorial and example material as well as local expertise and serendipity in obtaining help. Not to mention the codes successful and staightforward implementation on distributed MPI machines such as TAMNUN. For more information on TAMNUN, you can look at my links from earlier in the course.

The following is a guide to general procedure for simulation with large public or private codes at any scale for TAMNUN users.

  • If possible look at the list of codes on TAMNUN (or the computer you will use) which on TAMNUN can be found on the web from within the firewall at:
  • Check out the literature for simulations of the same or similar materials. Just like you compare a simple simulation with an analytic solution in some limit it is good to compare your results from a published simulation of a similar case. See which codes they used.
  • If you do not find overlap at this stage, you can either seek advice, officially from the HPC consultant, unofficially I will help students who have taken Computational physics in the future.
  • Select one or more of the codes on TAMNUN, codes, or if nothing suitable exists and you find a public domain code send a polite request to with the link to the download page. If its a public domain code that requires registration register and send your user/password with the request, or download and place in a folder on tx/aluf/t2 with full address. The systems people can lift things if you give the full path and its more efficient than emailing enormous files. If the code costs or has a limited license, check that it is suitable before buying it. If someone else at the Technion has bought it, be in touch to get advice, but do not expect them to let you use code they bought, although I, for example, will let people try out VASP for which I paid several thousand dollars but insist they pay for their own copy if its useful.
  • Once you have the code look for a source of demos/samples to run. These can be used to check the installation if it is a new one, and to teach you if its an old one.
  • There are many Computational Physics projects that include instructions for codes - and some of you will be adding to this collection. Not all are perfectly clear, but the really bad ones (and there were not many) have not been listed.
  • Many codes do not include graphics. There are instructions in many ofthe projects mentioned in the links below to output electronic density or atomistic configurations in AViz, but other graphics is used for graphs of functions, e.g. gnuplot. A few code use commercial graphics provided by the students' research group, but I try to avoid this so everyone can use the webpages later on.

    Let us start with a quick revision of condensed matter theory, and links to some great earlier projects.

    1. Bravais Lattices Shahar Rosen
    2. Reciprocal lattice - in momentum or k-space -FT of the direct lattice.
    3. Brillouin Zone Arik Rond is the Wigner-Seitz cell of the reciprocal lattice which contains all the unique k-vectors representing the periodicity of the waves allowed in a periodic structure.
    4. Fermi Surfaces Shir Kolangi

      This code can be downloaded from the websites. There also earlier codes here linked in a Wikipedia article on Brillouin zones.

    5. K-points - These are special points that can be used in Brillouin Zone sampling.
    6. Band structure - In a solid there are regions, ``bands" where electrons or holes are allowed to enter and bands that are forbidden. To understand material properties we need to calculate these.
    We will move to electronic scale, continue to atomistic scale and then cover some other topics from previous or current projects on large scale.