This award supports research and education activities that will make path integral Monte Carlo codes for simulation of atomic and molecular systems, and materials at the quantum mechanical level accessible to students and non-experts. This will stimulate new researchers to contribute fresh ideas and help build a larger base of users. The PI will pursue a strategy based on three principles: (i) On-going development of open-source software following software engineering practices must emphasize features that have a timely impact on science, encourage intuitive understanding, and assist non-expert users. (ii) Education and outreach in the form of tutorials, documentation, and workshops must dramatically lower the barrier for the use of path integral Monte Carlo by non-specialists. (iii) Development and execution of a set of high-impact projects -- in warm, dense matter, cold atoms, nano-electronics, and quantum chemistry -- by a network of collaborators which will demonstrate the research capabilities of the open-source path integral Monte Carlo code.
The code under development already embodies many software design principles, including object oriented design, XML input, HDF5 output, GNU GPL open-source distribution, and code management in subversion. Under this project the PI will: (i) complete documentation, including tutorials with explanations of the scientific concepts; (ii) develop a set of benchmarks and unit tests, to document standard applications and verify proper execution of the code; (iii) coordinate a set of high-impact projects across a network of collaborators, (iv) develop a timeline and milestones for formal releases, with executables for UNIX, Macintosh, and Windows, including the pyQt4 graphical user interface; and (v) integrate these improvements into existing nanoHUB port and other software infrastructures. New computer code will be distributed under the General Public License, and analysis tools will use open-source python libraries.
This award supports research and education activities to develop well engineered computer codes for the high accuracy simulation of electrons in materials and tiny structures made of a small number of atoms, and systems of atoms at the level of quantum mechanics. Developed computer codes will be made available to the broader research community through existing software centers, such as nanoHUB in a well documented form. This will enable students and non-experts to access and use the codes to perform simulations using path-integral techniques and to generate fresh ideas to advance the computational method.
Path-integral methods have been applied so far to various problems where quantum mechanics is important, including electronic systems and devices on the nanoscale, cold atoms trapped by lasers, and some atoms and small molecules. The technique has high accuracy and has advantages over existing methods, but it has so far not been adequately developed. It also has great pedagogical value for teaching the principles of quantum mechanics.
Students and junior scientists will be trained in the application of advanced computational methods to problems in nanotechnology.
(This project ended early when the PI left his academic job to work in the private sector, where he develops high performance computing software for the oil and gas industry. This report reflects work done on a smaller budget and shorter time period than the original proposal.) Software Infrastructure The pi-qmc program is publicly available as an open source project on GitHub. As promised in the grant, a test suite was developed using both unit tests and system tests. The unit tests validate correct behavior for functions in the code, and were used to test and develop new features, including the calculation of radiative recombination rates. The system tests verify that small, standard calculations always produce the correct results as the source code is refactored and improved. These tests are integrated into a continuous integration server hosted at jenkins-ci.org, so that the tests are run automatically whenever anyone contributes new source code to the project. A User's Manual was also developed with examples pulled from the system test suite. The User's Manual is automatically gfenerated and posted online on readthedocs.org. In addition to supporting this software, the testing and documentation practices serve as a useful template for other scientific computing project. Community Tools The open source software project was filled out with a supporting discussion board and error reporting, following guidelines from the book Producing Open Source Sofware (http://producingoss.com). Tools have been deployed to nanoHUB, though the project ended before significant updates could be published. The PI presented this work at the first annual NanoHUB User's Conference. Scientific Applications Applications of pi-qmc to spin-ordering in quantum wires was completed by graduate student Jianheng Liu and published in his Ph. D. dissertation. Astrophysics applications to the equation of state of hydrogen and helium were studied by a graduate student supported on the grant. A hypothesis of oxygen sequestration in Jupiter was written up and submitted to the journal Icarus. The calculations were inconclusive at the early completion of the project, and the graduate student moved from astrophysics to particle physics research. Research on quantum dots was conducted with Heriot Watt University. One paper on core-shell dots was published and presented at an invited talk. Additional calculations on responses to electric and magnetic fields were developed and included in the online software.
