Each successful candidate for the traditional DOE CSGF must have a specific science or engineering application for their research.
In contrast, the program’s math/computer science track is intended for candidates focusing on fundamental research into enabling technologies for high-performance computing (HPC) that are broadly relevant to science and engineering applications of interest to DOE.
Such areas include (but are not limited to):
- ODE, PDE, and integral discretization methods
- Linear and nonlinear solvers
- Multiscale, multi-physics coupling methods
- Verification, validation, and uncertainty quantification
- In situ data analysis
- High-dimensional data analysis
- Large-scale data visualization
- High-performance compilers
- Programming models and abstractions for heterogeneous computing
- Domain-specific languages
- Dynamic runtime environments
- Power management
- HPC development tools
- HPC performance analysis and tools
- Debugging at extreme scale
- Scalable I/O
- Scalable machine learning
- Interpretable machine learning
- Physics-constrained machine learning
- Robust machine learning
- Scientific data management and engineering
The interdisciplinary program of study for fellows in this track will still include science and engineering course requirements, ensuring that they are exposed to the computational needs of applications that will use these new enabling technologies.
Please contact us with questions about this track.