Book Image

Scientific Computing with Python - Second Edition

By : Claus Führer, Jan Erik Solem, Olivier Verdier
Book Image

Scientific Computing with Python - Second Edition

By: Claus Führer, Jan Erik Solem, Olivier Verdier

Overview of this book

Python has tremendous potential within the scientific computing domain. This updated edition of Scientific Computing with Python features new chapters on graphical user interfaces, efficient data processing, and parallel computing to help you perform mathematical and scientific computing efficiently using Python. This book will help you to explore new Python syntax features and create different models using scientific computing principles. The book presents Python alongside mathematical applications and demonstrates how to apply Python concepts in computing with the help of examples involving Python 3.8. You'll use pandas for basic data analysis to understand the modern needs of scientific computing, and cover data module improvements and built-in features. You'll also explore numerical computation modules such as NumPy and SciPy, which enable fast access to highly efficient numerical algorithms. By learning to use the plotting module Matplotlib, you will be able to represent your computational results in talks and publications. A special chapter is devoted to SymPy, a tool for bridging symbolic and numerical computations. By the end of this Python book, you'll have gained a solid understanding of task automation and how to implement and test mathematical algorithms within the realm of scientific computing.
Table of Contents (23 chapters)
20
About Packt
22
References

18.2.1 Prerequisites

You need to install this module first by executing the following in a terminal window:

conda install mpi4py

The module is imported by adding the following line to your Python script:

import mpi4py as mpi

The execution of a parallelized code is done from a terminal with the command mpiexec. Assuming that your code is stored in the file script.py, executing this code on a computer with a four-core CPU is done in the terminal window by running the following command:

mpiexec -n 4 python script.py

Alternatively, to execute the same script on a cluster with two computers, run the following in a terminal window:

mpiexec --hostfile=hosts.txt python script.py

You have to provide a file hosts.txt containing the names or IP addresses of the computers with the number of their cores you want to bind to a cluster:

# Content of hosts.txt
192.168.1.25 :4 # master computer with 4 cores
192.168.1.101:2 # worker computer with 2 cores

The Python script, here script.py, has to be copied...