View Position
| Project Title | Parallel Numerical Software Developer |
| Summary | Develop software to solve partial differential equations using high spatial resolution spectral methods. The student will develop and use spectral codes in two and three dimensions. Two dimensional codes will run on upto thousands of cores and three dimensional codes on upto hundreds of thousands of cores. |
| Job Description | The intern will help in the development and running of spectral simulation codes in ongoing projects to understand solutions to partial differential equations where high spatial resolution is required to understand how small spatial scales develop and evolve. Current codes have been demonstrated to work on upto 70,000 cores for 12,288^3 discretizations on NSF and DOE machines. A responsible and reliable individual is required to help optimize the codes and run scientific simulations. The student will join ongoing projects investigating nonlinear wave equations (in collaboration with research groups in Europe) and simulations of viscoleasticity. These equations model a variety of phenomena, including ocean waves, lasers and shape memory alloys. There is some scope for the intern to focus on simulating and understanding the behavior of a particular equation, or on software development or on visualization. The internship will serve as a good introduction to computational science which will give the student a broad understanding of the field and a detailed knowledge on one of the three aforementioned aspects. In both the above mentioned projects, the overall goal is to characterize the asymptotic behavior of highly oscillatory solutions that develop as the singular perturbation term is made very small. |
| Conditions/Qualifications | Preferably a student at Michigan. Coursework in differential equations, multivariable calculus and good programming skills. Interest in and knowledge of computer architecture a plus. |
| Start Date | 05/25/2011 |
| End Date | 04/30/2012 |
| Location | Ann Arbor, Michigan. Department of Mathematics |
| Interns | Paul Rigge
|