Research

Here are some of the research projects that are conducted via the help of HPCC resources.


Frank Hagelberg
, Physics and Astronomy

My present research focuses on the field of nanoscience and nanotechnology as the initiative to assemble and manipulate matter at the molecular or even atomic level. Physics, chemistry, biology and materials engineering are jointly contributing to the effort of tailoring novel materials by arranging small constituents in a controlled manner. This is referred to as bottom-up as opposed to the traditional top-down approach which consists in slicing down larger units. Nanoscience research is largely guided by the promise of novel devices at the nanoscale that will be vastly more compact, fast, flexible and energy efficient than the tools used in our current technology. To realize this goal, it is necessary to understand the principles that govern the nanosphere, characterized by spatial dimensions in the order of 10-9 m. Importantly, in-depth knowledge of the physical and chemical properties of nanomaterials is required.

My computational research in this field deals with low-dimensional systems that provide the physical basis for novel nanotechnological devices. At this juncture, I focus on advanced materials involving self-assembled monolayers (SAMs) and carbon nanostructures, mostly metallofullerenes and single-walled carbon nanotubes (SWCNTs). My studies on these and related problems involve various computational methods based on quantum physics and chemistry, such as density functional theory (DFT) as well as methods of ab initio theory. The computations are performed in part on ETSU machines, namely the computer clusters Parallel Quantum Solutions (PQS), Blackpearl, and Knightrider, in part on external machines. For more information please click here.


Christopher Wallace
, Computer Science

My research area is discrete/combinatorical optimization usually with emphasis on heuristics used for integer programming. Currently my main focus is looking for Ramsey Graphs bounds on good (4,5) graphs.

The current projects I am working on are:

  1. Porting over a mixed integer program solver such as CPLEX to work with clusters.
  2. Expanding a fast complete search heuristic to general binary programs.
  3. The bound on good Ramsey good (4,5) graphs is 24 but not all (4,5) graphs are known. For example it is estimated to be 14,600,000,000,000,000,000 good (4,6) graphs with 19 nodes. More information on (4,5) graphs in this area could improved bounds on r(4,6) and r(5,5), but the size of the search space is enormous. Searching for (5,5) graphs could be compared to searching the entire universe for just one specific atom.

The Knightrider cluster greatly improves that search capability by allowing increasing the the search five hundred times faster than one machine alone; what would take two years on my office machine can be done it one day on the cluster. The cluster also allows one to save large data sets, for example the number of graphs found just on 13 nodes is over 8 terabytes of data.