Examples of scaling in large GA applications

Performance of the Columbus multireference configuration code on the IBM SP

The code is solving the eigenvalue problem of a large sparse matrix (dimension >100 million) using the Davidson method. Speedup calculated relative to the 64-processor execution. Courtesy of Holger Dachsel.

Performance of molecular dynamics simulation on the 1300-node Cray T3E-900 (a teraflop) system

Classical molecular dynamics simulation of liquid octanol. The system size is 216,000 atoms. Speedup calculated relative to the 125-processor execution. Courtesy of T.P. Straatsma.