percMPI

Percolation for node-distributed data

How to run

The following packages are required

• CMake
• MPI (tested on MPICH2)

Visual Studio Code

Settings for Visual Studio Code are included (.vscode). Runs out-of-the-box on a linux machine with the following packages installed:

• visual studio code
• gcc
• gdb
• CMake
• MPI
• clang-format

Settings

For examples on how to build and run the code, see the visual studio tasks (.vscode/tasks.json)

Set the compile flag SINGLENODE for running sequentially

Set the compile flag COMMUNICATION for running parallel

Using shell-commands

When started from the main folder, scripts/build_make.sh builds and compiles different versions:

• parallel or multiple: Version for multiple nodes
• parallel-collectives or multiple-collectives: Version for multiple nodes with MPI Collectives where possible
• sequential or single: Version for a single node, the sequential algorithm

Data

A raw data set can be loaded with inputMode 3 (see task isotroph). An average and rms value need to be specified.

Random data is sampled for inputMode 20