pdf article.
Climate modeling: We have explored the application of automatic differentiation (AD) technology in the study of convective storms and, in particular, in a one-dimensional tornado model; the work was reported in the Monthly Weather Review. Currently, we are using AD-based schemes to tune sea ice model parameters.
Computational fluid dynamics:
The PETSc scalable nonlinear solvers provided the compute engine for one of 1999's Gordon Bell prizes for "Achieving High Sustained Performance in an Unstructured Mesh CFD Application.'' This code sustained over 227 Gflops on the ASCI Red machine and over 70% efficiency in going from 128 to 3072 processors. The code was developed jointly with researchers from NASA Langley, ICASE, and Old Dominion University.
Electric power markets: We are studying the ability of the largest producer in an electricity market to manipulate both the electricity and emission allowances markets to its advantage.
Analysis of the computed solution for the Pennsylvania - New Jersey - Maryland electricity market shows that the leader can gain substantial profits by withholding allowances and driving up NOx allowance costs for rival producers.
Fusion: We have been working with researchers at the Center for Extended MHD Modeling at PPPL on the assessment of high-order methods for future fusion simulation codes.
We have shown that high-order and
adaptive discretization technologies are decidedly superior to standard finite element approaches.
Nanophotonics:
In collaboration with researchers in Argonne's Chemistry Division, we are studying spectral postprocessing techniques as a means of increasing the performance
of
nanophotonics codes used for determining the assembly and material properties of nanoscale architectures.
Quantum chemistry:
We have had considerable success in the development of the parallel optimization algorithms in the Toolkit for Advanced Optimization (TAO).
In particular, TAO has enabled chemists to perform quantum chemistry simulations, which employ components based on (1) the NWChem and MPQC quantum chemistry codes for high-performance energy, gradient, and Hessian computations, (2) components based on Global Arrays and PETSc for parallel linear algebra operations, and (3) the TAO optimization component. During the past year, we performed parallel numerical experiments in molecular geometry design.
