Numerical Methods
Large-scale data analysis and simulation models are continuously being refined to improve predictive accuracy and precision. CI research into numerical methods for optimizing parallel codes, compilers, and other computational tools is enabling increased performance from existing approaches and creating new approaches to investigating numerical models of real-world problems.
Current Projects

ACTS: Adjoint Compiler
PIs: Paul Hovland
[more info]

OpenAD/F: Automatic Differentiation of Fortran Codes
PIs: Paul Hovland, Boyana Norris, Jean Utke
[more info]

