MultithreadingBenchmarks.jl

Author RelationalAI-oss
Popularity
5 Stars
Updated Last
1 Year Ago
Started In
August 2019

MultithreadingBenchmarks

Build Status

Results

The results from different experiments are written up as Issues in this package. The experiment writeups are all linked from the top-level experiment tracking issue, here:

Instructions for Running the Benchmark

The benchmark suite is set up to run via Pkg.test():

(MultithreadingBenchmarks) pkg> test

The testfile simply invokes test/runbench.jl which can also be invoked manually.

Configuration

However, currently, which experiments are run is configured in test/runbench.jl, and is currently controlled by just commenting out the various experiments in that file (Sorry!). So just make sure when you run test/runbench.jl that the experiments are configured how you want them to be! :)

The test should automatically scale to the number of cores on your machine, though you may want to edit the const NUM_DATAPOINTS = 9 to be smaller if you have fewer cores.