Computational GS500 methods (full term; 4 credits) GS503 Introduction to computer programming (half term; 2 credits) Computational methods have become increasingly important for quantitative modeling and data analysis in the Geological Sciences. This set of courses provides a survey of the most important computational techniques, ranging from an introduction to computer programming to advanced parallel computing on linux clusters. The course is introductory and hands-on, with the grade determined by weekly computational problem sets. The courses 500+503 meet together for the first half of the semester. GS500 continues in the second semester with a specific focus on parallel computing. Instructors: Peter van Keken, 4502 CC Little Building, [e-mail]; Erik Kneller, 4514 CC Little Building, [e-mail]; Mike Messina, 25xx CC Little Building, [e-mail]; Meeting times: lecture TTh 10-11:30a, lab M7-9p (lab hours are still flexible). Course material: notes, programming examples and assignments will be distributed through the course website. For an example of see www.geo.lsa.umich.edu/~keken/535 and www.geo.lsa.umich.edu/~keken/534 for materials provided for two ad-hoc seminars that were held in Fall 2003 and 2004. Recommended reading: Press et al., Numerical recipes, Cambridge Sobell, A practical guide to linux, Addison-Wesley Prata, C++ Primer Plus, 4th edition, SAMS Gropp et al., Using MPI, 2nd edition, MIT press. Nyhoff and Leestma, Fortran 90 for engineers and scientists, Prentice Hall. Course topics: GS500/503 (2 credits; first half term) Introduction to Unix Comparison to MacOS/Windows; file management, editors (vi/emacs) Computer programming History and philosophy; best practices; error checking & benchmarking Introduction to classical and modern languages (e.g., Fortran/C/C++/java/python) Comparison with high-end computational applications (matlab/mathematica) Basic numerical techniques Numerical integration/differentiation/extrapolation Data analysis Regression, multivariate analysis, visualization GS500 only (for 2 additional credits during second half of term) Shell programming bash/cshrc, batch jobs Introduction to solution of partial differential equations Poisson/Laplace equation, relevance for geophysical and geochemical modeling Introduction to parallel computing Distributed vs shared memory programming, MPI Use of parallel clusters Secure shell management, batch jobs, resource management Large data set Parallel issues IO, data mining Visualization GMT, Paraview, volview
The course schedule is somewhat flexible and can be adjusted based on students interest. There is plenty of opportunity for students to work on individual projects. We encourage team work. Wireless access will be available during the lectures and we encourage you to bring in your own laptop at all times.
Tentative lecture schedule Tu 9/5 Introduction + overview, history of computing Th 9/7 Nuts and bolts: computer hardware + generic OS issues (MM) Tu 9/12 Introduction to linux; comparison with Windows/Mac; simple programming Th 9/14 Introduction to programming: elements of Fortran/C, compilation, IO (EAK) Tu 9/19 Introduction to programming (in 5501) (EAK) Th 9/21 Basic numerical techniques I: interpolation, extrapolation, differentiation, integration Tu 9/26 Basic numerical techniques II. Th 9/28 Poisson equation: background & relevance; solution techniques, mathematica etc. Tu 10/3 Poisson equation: programming issues/numerical recipes Th 10/5 Data regression: background, approaches; maple, matlab, excel Tu 10/10 Data regression: programming issues/numerical recipes Th 10/12 Student presentations: comparison between approaches Tu 10/17 Fall Break Th 10/19 Automating linux: shell programming in bash Tu 10/24 (GSA) Introduction to parallel programming: shared vs distributed memory models Th 10/26 Introduction to MPI Tu 10/31 Parallel applications: embarrassingly parallel Th 11/2 Parallel ...