Applied Mathematics 483/583

Slides from lectures:

The 3-up version has 3 slides per page and room for notes.

These slides are being made freely available and are released under the Creative Commons CC BY license. This means that you are welcome to use them and quote from them as long as you give appropriate attribution.

• Lecture 1 ... 3-up version

  Goals of course, overview, requirements

• Lecture 2 ... 3-up version

  Binary and floating point, version control

• Lecture 3 ... 3-up version

  Python and git demo on mysqrt

• Lecture 4 ... 3-up version

  More demo of Python, git

• Lecture 5 ... 3-up version

  Python concepts, data types, modules, IPython notebook

• Lecture 6 ... 3-up version

  NumPy arrays, functions, timing Python

• Lecture 7 ... 3-up version

  Debugging python, intro to Fortran 90

• Lecture 8 ... 3-up version

  Fortran functions, subroutines, arrays

• Lecture 9 ... 3-up version

  Multi-file programs, Makefiles and make

• Lecture 10 ... 3-up version

  Fortran modules

• Lecture 11 ... 3-up version

  computer architecture, cache, optimization

• Lecture 12 ... 3-up version

  Computer arithmetic, optimization flags, parallel computing

• Lecture 13 ... 3-up version

  Amdahl's law, strong and weak scaling, OpenMP

• Lecture 14 ... 3-up version

  OpenMP: critical sections, parallel do loops

• Lecture 15 ... 3-up version

  Amazon Web services, timing Fortran

• Lecture 16 ... 3-up version

  OpenMP: loop dependencies, threadsafe, other directives

• Lecture 17 ... 3-up version

  Fine grain vs. coarse grain, normalize and umax1.f90 examples

• Lecture 18 ... 3-up version

  Introduction to MPI

• Lecture 19: Guest lecture by Fernando Perez (UC Berkeley) on parallel computing in IPython.

  Rather than slides, Fernando used an IPython Notebook that was adapted from one developed by Skipper Seabold. You can view this notebook online or download this and other examples by cloning the git respository https://github.com/jseabold/zorro

• Lecture 20 ... 3-up version

  Adaptive quadrature, recursion, and load balancing

• Lecture 21 ... 3-up version

  MPI send and receive, master-worker paradigm

• Lecture 22 ... 3-up version

  MPI master-worker paradigm, LAPACK, BLAS

• Lecture 23 ... 3-up version

  LU factorization, condition number, heat equation discretization, Jacobi

• Lecture 24 ... 3-up version

  OpenMP and MPI versions of Jacobi, Gauss-Seidel, SOR

• Lecture 25

  Guest lecture by Brad Chamberlain from Cray on "Chapel: A Next-Generation Partitioned Global Address Space (PGAS) Language"

• Lecture 26 ... 3-up version

  Monte Carlo methods and random numbers

• Lecture 27 ... 3-up version

  Monte Carlo methods for Laplace's equation, f2py

• Lecture 28 ... 3-up version

  Numba, JIT, reading and writing binary files, Summary