Python for Scientific Computing at SIAM CSE 2011: Slides¶
At the 2011 SIAM CSE meeting held in Reno on February 28-March 6, Randy LeVeque from U. Washington, Hans-Petter Langtangen from the Simula Research Laboratory in Norway and I co-organized a 2-part minisymposium entitled Python for Scientific Computing. This was done partly as a followup to the successful one we had in 2009, and we had again good attendance and lively discussions.
Our minisymposium was divided in two parts, and I am posting here links to the original program pages as well as all the slides I have so far from the speakers, along with any links they may have provided to their personal or project pages. A short narrative is my blog and I also put some pictures up.
As a sign of the healthy growth of the scientific python ecosystem, this year there were also other python-focused minisymposia at this conference, as well as several python talks in other MS; those slides are below.
MS: Python in Scientific Computing - Part I of II
- Capabilities and Recent Developments of NumPy for Scientific Computing
- Travis E. Oliphant, Enthought, Inc., USA
- Cython: Compiled Code meets Dynamic Python
- Lisandro Dalcin, Centro Int. de Métodos Computacionales en Ingeniería, Argentina; Robert Bradshaw, University of Washington, Seattle, USA
- Matplotlib - from Interactive Exploration to Publication Graphics
- John Hunter, Tradelink Securities, Inc., USA
- Why Modern, High-performance Networking Matters for Interactive Computing
- Fernando Perez, University of California, Berkeley, USA; Brian Granger, California Polytechnic State University, San Luis Obispo, USA; Evan Patterson, California Institute of Technology, USA
MS: Python in Scientific Computing - Part II of II
- Interactive Parallel Python with ZeroMQ
- Min Ragan-Kelley and Fernando Perez, University of California, Berkeley, USA; Brian Granger, California Polytechnic State University, San Luis Obispo, USA
- Real-time Classification of Astronomical Events with Python
- Joshua Bloom, Dan Starr, Joseph Richards, Nathaniel Butler, and Dovi Poznanski, University of California, Berkeley, USA
- SymPy: Symbolic Mathematics in Pure Python
- Mateusz Paprocki, University of Nevada, Reno, USA
- FEMhub Online Numerical Methods Laboratory
- Pavel Solin and Ondrej Certik, University of Nevada, Reno, USA; Mateusz Paprocki, University of Wroclaw, Poland; Aayush Poudel, University of Nevada, Reno, USA
Other Python talks¶
There were a number of other sessions dedicated to Python tools, as well as talks based on Python elsewhere. Here is a summary of the ones for which I have slides (a few are pending review by National Labs for public release):
MS: Python-based Software for Solving Partial Differential Equations - Part I of II
- Lessons Learned and Open Issues from the Development of the Proteus Toolkit for Coastal and Hydraulics Modeling
- Chris Kees and Matthew Farthing, U.S. Army Engineer Research and Development Center, USA
- FEniCS: An Attempt to Combine Simplicity, Generality, Efficiency and Reliability
- Kent-Andre Mardal, University of Oslo, Norway; Hans Petter Langtangen, Simula Research Laboratory and University of Oslo, Norway; Anders Logg, Simula Research Laboratory, Norway; Garth N. Wells, University of Cambridge, United Kingdom
MS: Python-based Software for Solving Partial Differential Equations - Part II of II
- Making a Python based PDE Solver Work Efficiently in Parallel with the Available Open Source Interfaces to MPI
- Daniel Wheeler and Jonathan Guyer, National Institute of Standards and Technology, USA; James O’Beirne, George Mason University, USA
- Python, Clawpack, and PyClaw
- Randall J. LeVeque and Kyle T. Mandli, University of Washington, USA
- Mpi4py and Petsc4py: Using Python to develop Scalable PDE Solvers
- Lisandro Dalcin, Centro Int. de Métodos Computacionales en Ingeniería, Argentina; Chris Kees, U.S. Army Engineer Research and Development Center, USA
- Paper to GPU: Optimizing and Executing Discontinuous Galerkin Operators in Python
- Andreas Kloeckner, Courant Institute of Mathematical Sciences, New York University, USA
MS: Python Software for Numerical Optimization
- A Flexible Python Environment for PDE-Constrained Optimization
- Dominique Orban, École Polytechnique de Montréal, Canada; Nick Gould and Sue Thorne, Rutherford Appleton Laboratory, United Kingdom
- Analyzing Optimization Models Developed with the Pyomo Modeling Software (slides pending review by Sandia NL for release)
- William E. Hart and Jean-Paul Watson, Sandia National Laboratories, USA; David Woodruff, University of California, Davis, USA
- Stochastic Nonlinear Programming with Pyomo (slides pending review by Sandia NL for release)
- Carl Laird, Texas A&M University, USA; Jean-Paul Watson, Sandia National Laboratories, USA
- Algorithmic Differentiation in Python
- Sebastian F. Walter, Humboldt University Berlin, Germany; Bradley M. Bell, University of Washington, USA
MS: FEniCS: Automated Solution of Differential Equations
- On the Construction of Preconditioners for Systems of PDEs
- Kent-Andre Mardal, University of Oslo, Norway
- Automated Goal-Oriented Error Control
- Marie E. Rognes and Anders Logg, Simula Research Laboratory, Norway
MS: CSE Education
- Sculpture, Geometry and Computer Science
- Randy Heiland, Indiana University, USA; Charles Perry, www.charlesperry.com, USA; Barbara Ream, International School of Columbus, USA; Andrew Lumsdaine, Indiana University, USA
MS: Verifiable, Reproducible Research and Computational Science
- Reproducible Research, Lessons from the Madagascar Project
- Sergey Fomel, University of Texas at Austin, USA (lead of the Magadascar project).
- Design and Preliminary Results for PIC on GPUs with Python
- Min Ragan-Kelley, University of California, Berkeley, USA; John Verboncoeur, University of California, USA
MS: Advanced Algorithms on GPUs
- A Case Study of GPUs in Scientific Computing: Low-Order FEM
- Matthew G. Knepley, Argonne National Laboratory, USA
- High-Order Discontinuous Galerkin Methods by GPU Metaprogramming
- Andreas Kloeckner, Courant Institute of Mathematical Sciences, New York University, USA; Timothy Warburton, Rice University, USA; Jan S. Hesthaven, Brown University, USA