Using Python for large scale linear algebra applications

Authors:
Oliver Bröker;Oscar Chinellato;Roman Geus
Affiliations:
Computational Laboratory (CoLab), Hirschengraben 84, ETH, Zurich 8092, Switzerland;Institute of Computational Science (ICoS), ETH, Zurich, Switzerland;Paul-Scherrer-Institute (PSI), 5232 Villingen, Switzerland
Venue:
Future Generation Computer Systems - Special section: Complex problem-solving environments for grid computing
Year:
2005

Citing 4
Cited 1

Parallel Preconditioning with Sparse Approximate Inverses

SIAM Journal on Scientific Computing
LAPACK Users' guide (third ed.)

LAPACK Users' guide (third ed.)
Multigrid

Multigrid
An Empirical Comparison of Seven Programming Languages

Computer

Object-oriented electrodynamic S-matrix code with modern applications

Journal of Computational Physics

Quantified Score

Hi-index	0.01

Visualization

Abstract

Software used in scientific computing is traditionally developed using compiled languages for the sake of maximal performance. However, for most applications, the time-critical portion of the code that requires the efficiency of a compiled language, is confined to a small set of well-defined functions. Implementing the remaining part of the application using an interactive and interpreted high-level language offers many advantages without a big performance degradation tradeoff. This paper describes the Pythonic approach, a mixed language approach combining the Python programming language with near operating-system level languages. We demonstrate the effectiveness of the Pythonic approach by showing a few small examples and fragments of two large scale linear algebra applications. Essential advantages of the Pythonic mixed language approach is the combination of flexible, readable, shorter, and most importantly less error-prone syntax with performance similar to pure Fortran or C/C++ implementations.