DataOrientedDemo

This repository demonstrates performance differences between different implementations of a simple interpolation algorithm.

Clone the repository with the version of GSL which is included as a submodule:

git clone --recursive https://github.com/Shachlan/DataOrientedDemo.git

Compile the program:

make debug    # debug build
# or
make check    # optimized build

In order to run the program, call:

./check <number of models> <average number of animations per model> <number of repetitions>

For simplicity's sake, there's also the ability to use docker to run the demo:

make run-debug MODELS=<number of models> ANIMATIONS=<average number of animations per model> REPETITIONS=<number of repetitions>

or

make run-check MODELS=<number of models> ANIMATIONS=<average number of animations per model> REPETITIONS=<number of repetitions>

for example:

make run-check MODELS=5 ANIMATIONS=20 REPETITIONS=100

The program will create the models, interpolate them and verify that each implementation reached the same values, and then print the time each implementation took to complete.

PRs are welcome!