Principles of Performance

Today the quant finance and fintech sectors attract top talent from a wide range of quantitative academic backgrounds, often outside of traditional computer science. While this diversity has been advantageous in many respects, these students and professionals can often find themselves mired by performance related issues when they take their analyses from the classroom setting to the marketplace and have to interface with real world data structures at scale. This tutorial will cover the basic foundational concepts needed to effectively design and implement scalable algorithms like those common to quant finance or related financial technology applications.

Many common performance issues are rooted in a lack of knowledge about how a computer actually performs computation. We will begin by covering how modern computers physically execute code. This low level information will help us reason about the behavior of our high level code later. We will then look at Python’s execution model with our new understanding of the machine. We will then discuss how numpy allows us to take full advantage of the power of our computer while staying in Python. Finally, we will look at tools for analyzing the performance of Python programs and cover common issues and fixes.

By the end of this session, attendees will:

• Have a general understanding of how the processor works.
• Know about memory management and cache locality.
• Understand the reason Python is “slow”.
• Understand how numpy makes Python fast.
• Be familiar with using cProfile to analyze programs.

Install Steps

Prior to the tutorial, attendees need to install a git and Python 3.6. After that, run the following commands.

$ git clone --recursive https://github.com/llllllllll/principles-of-performance.git
$ cd principles-of-performance
$ source etc/setup-env

The setup-env script will attempt to download the needed packages. The setup-env should print a lot of stuff to the terminal. You can ignore most of it but the last line should be:

Environment is setup correctly!

Viewing the Tutorial

The tutorial is structured as a sphinx project. This allows the tutorial to be viewed from a standard browser or hosted online.

The material can be viewed in a browser by opening tutorial/build/html/index.html, for example:

$ ${BROWSER} tutorial/build/html/index.html

Acknowledgments

The numpy examples are sourced from Scott Sanderson’s https://github.com/ssanderson/foundations-of-numerical-computing

Contents:

• Bits
  • Representing Non-Negative Integers with Bits
  • Addition with Bits
  • Representing Negative Numbers
• Low Level Computation
  • Main Memory
  • Computation Circuits
  • Bitness
  • Sequencing Low Level Operations
• Arrays and Structs
  • Arrays
  • Structs
  • Multi-Dimensional Arrays
• Memory Locality
  • Cache Hierarchy
  • Multi-Dimensional Arrays
  • Example
• Memory Management
  • Virtual Memory
  • Two Tiers of Allocation
• How to Optimize Code
  • Algorithmic Complexity
  • Allocations and Copies
  • The Fastest Operation
• Python Overview
  • Python Objects
• Numpy Overview
  • ndarray
  • Operations
  • Broadcasting
• Profiling
  • cProfile
  • pstats
• Appendix
  • Processor
  • Memory
  • Data Structures
  • Miscellaneous

Indices and tables

• Index
• Module Index
• Search Page