My work in the compiler enables the the easy programmability of sophisticated visualization algorithms. The compiler represents the high level of math by the design and implementation of our intermediate representation EIN
. To test the implementation of our compiler I developed DATm
our automated testing tool. Alongside a bright advisee and other collaborators we are extending the language for finite element community.
Diderot users should be able to write intuitive code, compile programs with complicated tensor math, believe in the correctness of the compiler, and define fields by other sources of data.
The work is organized by research Topic.
The EIN IR + Language design and Implementation details
Representing a high level of tensor math with the EIN IR, compiler design to evaluate and apply computations to PDE solutions, and new language features.
New language features include user-defined functions and writing directly in EIN notation
Compiler Optimization details
Loop optimization and compilation techniques to reduce the size of the IR
Automated Testing and Correctness details
DATm: Diderot's automated testing tool and proofs organized around the rewriting system.
Scientific Visualization and Scientific Computing Applications details
The application of Diderot to the scientific computing domain to create sophisticated and informative visualizations.
charisee.chiw at gmail.com