September2008Meeting

This talk will be an introduction to doing scientific computing with Perl and Math::GSL.

• features:

  • numerical derivatives and integration
  • special functions, statistics, permutations/combinations
  • random number generators
  • Linear algebra (BLAS)
  • Fast Fourier Transform
  • Wavelets
  • Splines/Interpolation
  • Histograms
  • 1D Curve Fitting
  • 1D Root Solving
  • Multidim. Root Solving (soon)
  • Ordinary Differential Equation Solver (soon)
  • Multidimensional curvefitting (soon)
  • Multidimensional minimization (soon)

• use Perl+Math::GSL as glue between sensors/data source and computer algebra system (like Matlab/Mathematica/Maple/Macsyma ... )

  • why isn't there a Perl CAS or something like sciPy.org ?

• why it is better than pure Perl

  • 2.5x faster numeric sort, sort_k_largest, sort_k_smallest functions

• why Math::GSL + Perl beats "whatever you're doing now"

  • throwaway analysis code is 80% data translation, 20% heavy-lifting
  • easier to work with sensor data, networking
  • somewhere between matlab and a monolithic fortran solver on a cluster of mainframes

Podcast Slides