Experience with adapting sequential programs to a parallel environment is shared with the reader. Our programs are used in quantum-chemical calculations but certain parts of them are of general application and our results can be adapted to other types of problems. Several PC nodes are connected through a fast network and consolidated to a cluster. Our applications make use of the Message Passing Interface environment. Encouraging results concerning speedup and efficiency have been obtained. Experiments leading to a superlinear speedup using the hyperthreading technology are also reported.
|Efficiency of Matrix Elements Computations on Parallel Systems||2014-07-30|