Parallel Climate Data Assimilation PSAS Package Achieves 18 GFLOPS on 512-node Intel Paragon

We have designed and implemented a set of highly efficient and highly scalable algorithms for an unstructured computational package, the PSAS data assimilation package, as demonstrated by detailed performance analysis of systematic runs on up to 512-node Intel Paragon. The preconditioned Conjugate Gradient solver achieves a sustained 18 Gflops performance. As the results, we achieved an unprecedented 100 -fold solution time reduction on the Intel Paragon parallel platform over Cray C90. This not only meets and exceeds the DAO daily computing requirements, but also made it possible to explore much larger and challenging data assimilation problems which were unthinkable on a traditional computer platform such as Cray C90.