valdojoao / Distributed-Systems

04/2016: Implementation of a high performance parallel Fish and Shark application targeting a multi-core architecture (AMD Opteron 2.3GHz processor with 12 cores and 12 contexts)

Distributed-Systems: 04/2016

The goal of the project was to implement a parallel Fish and Shark application targeting on multi-cores. The project was developed in Java starting from a simple sequential solution then using Skandium framework was possible to build the parallel application. All tests were run on an AMD Opteron (2.3GHz) processor with 12 cores and 12 contexts. More details fan be found at Joao_Report.PDF

Requirements

Skandium http://backus.di.unipi.it/~marcod/SPM1011/Skandium.pdf

Manual for Fish and Sharks application

How to compile
javac -cp /usr/local/Skandium/skandium-1.0b2.jar src/fish_and_sharks /*.java

How to run
java -cp /usr/local/Skandium/skandium-1.0b2.jar: fish_and_sharks.Run #threads #matrixSize #generations.
E.g: java -cp /usr/local/Skandium/skandium-1.0b2.jar: fish_and_sharks.Run 1 2000 50
The above example will run the application with 1 thread, matrix size of 2000 for 50 generations.

PS: In order to see the value of the original matrix in the constructor of the class Matrix you need to uncomment the method printMatrix() which is on line 46.
In order to see the value of the updated matrix after the given iteration, in the class Run you need to uncomment the method matrix.printMatrix() which is on line 57.