mi-iro / MPI-Experiment

UESTC Course Experiment: Implements of Sieve of Eratosthenes, with optimizer and improvement

Parallel-and-Distributed-Computing-MPI-Experiment

Introduction

Parallel and Distributed Computing of UESTC

Course Experiment: Implements of Sieve of Eratosthenes, with optimizer and improvement

Usage

1. make
2. mpiexec -n <num_threads> ./optimizer1 <limit>
3. mpiexec -n <num_threads> ./optimizer2 <limit>
4. mpiexec -n <num_threads> ./optimizer3 <limit> [chunk_size]
5. mpiexec -n <num_threads> ./optimizer4 <limit> [chunk_size]
6. make clean

e.g.

make
mpiexec -n 4 ./optimizer1 1000000000
mpiexec -n 4 ./optimizer4 1000000000
mpiexec -n 4 ./optimizer4 1000000000 23333
make clean

Environment

Base on

• Microsoft Windows 10 Pro 10.0.19044 Build 19044
• WSL Ubuntu 18.04 bionic
• Intel Core i7-8565U @ 8x 1.992GHz (4 physical cores)
• MPICH 3.3a2

File Description

• base.cpp: The origin source code provided(corrected some bugs),
• optimizer1.cpp: base + 2's reduced,
• optimizer2.cpp: opt1 + broadcast reduced,
• optimizer3.cpp: opt2 + cache reduced,
• optimizer4.cpp: opt3 + 3's and 5's reduced + union + O3.