Parallel Computation of 7-Point Stencil via 3D Decomposition - MPI

Description

Implementation of a parallel algorithm using 7 point stencil computation is performed using MPI.

Compilation

• 3d_slice.cpp: $ mpic++ -Wall 3d_slice.cpp -o slice

• 3d_serial.cpp: $ gcc -o 3d_serial serial.cpp

Usage

$ mpirun -np "p" ./slice "width" "height" "depth" "iteration" where:

• "p": number of processees
• "width": width of the 3d structure
• "height": height of the 3d structure
• "depth": depth of the 3d structure
• ("width"+2) must be divisible by "p" and result must be bigger than 2

e.g. $ mpirun -np 4 ./slice 126 62 62 100 $ ./serial 126 62 62 100