15-440-lab4
K-Means

Boya Yang (boyay) Shikun Zhang (shikunz)

Submission incluse:

• pdf file:

  • report.pdf

• shell scripts:

  • runSample runs a sample data of size 5000 with k=5 using both algorithms (use 8 processors for parallel)

  First Step: generate Data

  • dataGenerator generates datasets for both DNA and Points using a combination of different k's and p's, saves file into folder /data/ k ranges from 2 to 10; p is in [10, 100, 1000, 10000, 100000] also generate the solution file, the original k centroids

  • runSequential runs sequential algorithms on two types of datasets generated by the dataGenerator with profiling information

  • runParallel runs parallel algorithms on two types of datasets generated by the dataGenerator by using 2, 4, 8 or 12 processors with profiling information

• python files (used in the shell scripts):

  -k # of clusters -e value of epsilon -i input file -o output file -p number of points aruond centroid -l length of DNA strand, default = 50 -r range of coordinates, default = 1000000

  • PointGenerator.py -k -p -o [-r ]

  • PointSequential.py -k -e -i -o

  • PointParallel.py -k -e -i -o

  • DNAGenerator.py -k -p -o [-l ]

  • DNASequential.py -k -e -i -o

  • DNAParallel.py -k -e -i -o ======================================================================