This assignment has two goals:
- it coaches you in good Julia practices, avoiding traps that can kill performance. By and large, this is an exercise in avoiding patterns that result in poor type inference, and in experimenting with tools that help you detect, diagnose, and resolve inference problems.
- it introduces a few concepts from the analysis of algorithms, specifically the "big-O" notation and its use in evaluating implementations.
For this homework, start from the test/runtests.jl file, which mostly just includes a number of individual test files.
Open them one-by-one and read the comments that describe the problems to be solved.
These individual test files start with one or more links, usually to sections of the Julia manual's Performance tips page. Start by reading those pages/sections for background information that will help in solving this problem.
Solve the problems by modifying the file of the same name in src/, not by modifying the tests (unless explicitly instructed to so do).
One exception to this rule is the ntests_finished at the top of runtests.jl: if you're bothered by failures of the
tests you haven't yet fixed, just set this to the number of files whose tests you want to run. For example, to run
just the tests in test/globals.jl, set ntests_finished = 1. If you do modify ntests_finished,
make sure you put it back (or set ntests_finished = typemax(Int)) to ensure they all run before
you declare victory with this assignment.
Some problems might say "See ?somefunction," which means you should read the help on somefunction.
Finally, a couple of notes about this repository:
- when creating it, I deliberately left out the
DocumenterandCodecovplugins since neither is needed for this assignment. - I listed
BenchmarkToolsas a depenency of this repository. This is for your convenience, but any real package generally shouldn't include such dependencies. Put them in your default environment instead, reserving the package environment for things actually needed by the package.