Distributed K-Nearest-Neighbors using MPI and a Vantage Point Tree

Datasets used for benchmarking

• corel
• fma
• miniboone
• tv_news_com

Build instructions

You will need an mpi compatible compiler (mpic++) to compile and run (mpirun) the application.

You can find more details about compilation inside the Makefile.

make

Instructions for dataset paths

Place each file named $fileName that belongs to the $dataset in the directory with the following structure:

$repoRootDirectory/dataset/$dataset/$fileName

For example, for the file ColorHistogram.asc from corel dataset use

./dataset/corel/ColorHistogram.asc

Note: Some files are cleaned using a python script before processing, beware.