Applied parallel computing

Více o knize

The PARA workshops have focused on parallel computing methods in science and technology, with seven meetings held to date: PARA’94, PARA’95, PARA’96 in Lyngby, Denmark; PARA’98 in Umea, Sweden; PARA 2000 in Bergen, Norway; PARA 2002 in Espoo, Finland; and PARA 2004 again in Lyngby. The first six meetings featured lectures on modern numerical algorithms, computer science, engineering, and industrial applications in the context of scientific parallel computing. The PARA 2004 Workshop, titled “State of the Art in Scientific Computing,” took place from June 20–23, 2004, and was organized by Jack Dongarra from the University of Tennessee and Oak Ridge National Laboratory, alongside Kaj Madsen and Jerzy Was ´niewski from the Technical University of Denmark. This workshop shifted the focus to high-performance computing (HPC), highlighting the potential of advanced computers to tackle complex computational problems. However, the intricacies of modern computer architectures necessitate careful management to fully realize this potential. For instance, failing to utilize a computer's memory hierarchy can significantly hinder performance. A key concern in HPC is developing software that optimizes performance for specific systems. Additionally, the high cost of cutting-edge computers can be a barrier for many organizations, particularly those with only occasional HPC needs.