Parallel Implementations of the k-Means Algorithm

Sequential, MPI, and OpenCL implementations of the k-means clustering algorithm.

Prerequisites

• OpenMPI
• OpenCL
• Python 2.7 (or greater)

Input format

A text file for the x coordinates and another for the y, example inputs are in the Data folder

Run instructions

Compile the programs with

make

Note: The default OpenCL compiler is nvcc, edit the makefile for AMD/Intel GPUs

Sequential

./kMeans [n] [k] [x file] [y file]

Example:

./kMeans 1000 8 Data/X1000.txt Data/Y1000.txt

MPI

./kMeanMPI [n] [k] [x file] [y file]

OpenCL

./kMeansCL [n] [k] [x file] [y file]

Visualisation

Sequential

python plot.py [x file] [y file]

MPI

python plotMPI.py [x file] [y file]

OpenCL

python plotOCL.py [x file] [y file]