Research
Computational Research at ScaleConducts multidisciplinary research in large‑scale scientific simulation and high‑performance computing, advancing methods, software, and applications across engineering and physics.
Advancing State-Of-The-Art, Large-Scale Simulations
The computational simulation of complex physical processes plays a significant role in many industries and in our national security. Simulations help designers and analysts assess the behavior of engineered and natural systems under a variety of conditions. They also play a vital role in the design of experiments and in the interpretation of those results.
The Center for Large-scale Scientific Simulations (CLASS) has built multidisciplinary teams of faculty members and graduate students from several departments within Texas A&M University’s Colleges of Engineering and Arts and Sciences to work collaboratively with key researchers at national laboratories on education and research relevant to computational multiphysics simulations.
The OpenSn open-sourse Sn Code
CLASS develops and maintains the OpenSn open-source Sn code (https://github.com/open-sn). OpenSn is a 3-D unstructured polyhedral-mesh Sn code that serves as a research vehicle for Sn methods development as well as a tool for reactor analysis and neutral- and charged-particle transport calculations. CLASS researchers are internationally recognized as experts in the development of massively-parallel Sn solution algorithms.
Collaborative Research and Development supporting LLNL Missions
This is a computational research project with components in strength of materials, microstructure evolution/phase-field modeling, hydrodynamics, and machine learning for radiation transport and associated applications. Participants represent three departments in the Texas A&M College of Engineering and one department in the College of Arts and Sciences with close collaboration between Texas A&M faculty, students, and LLNL scientists. The latter serve on student research committees and students do at least one summer internship at LLNL.
Collaborative Research and Development supporting Stockpile Stewardship
This computational research project includes components consisting of a new Sn-like method for computing the uncollided flux from an isotropic point source of radiation, new “non-invasive” discontinuous finite-element discretization techniques for the angular derivative term in the 2-D cylindrical-geometry Sn equations, and massively-parallel GPU-based solution algorithms for the Sn equations on 3-D unstructured meshes. All participating faculty and students are from the Nuclear Engineering Department. There is close collaboration between the faculty, students, and LLNL staff. The latter serve on student research committees and students do at least one summer internship at LLNL.
Other Research Projects with National Labs
Sandia National Laboratories
Development of Lebedev-Based Galerkin Quadratures for Coupled Electron-Photon Sn Calculations. All participants are faculty and students in the Nuclear Engineering Department.
Los Alamos National Laboratory
Development of Transport Methods with Applications. All participants are faculty and students in the Nuclear Engineering Department.
Research Resources
Software
C++ Libraries: BOOST
Scientific Visualization: VISIT
Parallel Graph Partitioning and Fill-reducing Matrix Ordering: parMETIS
Finite-Element Libraries: deal.II, libMesh
Linear Solvers: PETSc/, TRILINOS, HYPRE
Radiation Transport: PDT, PARTISN, DENOVO, MCNP
Facilities
CLASS has access to various clusters at Texas A&M University, including the HYDRA IBM p5-575 Cluster, as well as various massively parallel computers at Lawrence Livermore National Laboratory.