NYCU Parallel Programming, 2024 Fall

• 📖 Parallel Programming, 2024 Fall, NYCU - https://nycu-sslab.github.io/PP-f24/
• 🧑‍🏫 You, Yi-Ping - https://ypyou.github.io/

Assignments

• 1 SIMD Programming - https://hackmd.io/@rjrGMU3mRO2bOlkt_7lsqA/ry9zuoFTR
• 2 Multi-thread Programming - https://hackmd.io/@rjrGMU3mRO2bOlkt_7lsqA/SkdxgX4JJe
• 3 OpenMP Programming
• 4 MPI Programming
• 5 CUDA Programming
• 6 OpenCL Programming - https://hackmd.io/w_711eM4TZeWuK27ar8HcA?view

Final Project - Acceleration-of-FP-Tree-construction (MPI)

This repository contains a C++11 implementation of the well-known FP-growth algorithm, published in the hope that it will be useful. I tested the code on three different samples and results were checked against this other implementation of the algorithm and use MPI method to construct FP-tree in parallel.

The files fptree.hpp and fptree.cpp contain the data structures and the algorithm, and main.cpp contains a few tests.

The file MPI_tree.cpp contains the FP-tree parallelized using the MPI method.

Compile the code using the appropriate options for C++11 (e.g. -std=c++11 using g++).

branch

• Use MPI to parallelize the FP-tree construction

• Parallelize the process by dividing it into two parts:

  • Build frequency item set

  • Tree construction

How to run the code

• Please ensure that your environment supports MPI and g++, then follow the instructions below

  • First: change the data file directory(data_folderPath) in MPI_main.cpp

  • Then

    # Navigate to the FP-growth directory
    cd src
    
    # Remove all compiled files
    make clean
    
    # Compile the program
    make
    
    # Run the parallel code on the dataset
    srun --mpi=pmix -n {number_of_processors} main