EigenKernel - A middleware for parallel generalized eigenvalue solvers to attain high scalability and usability
| Autor: | Tomoya Fukumoto, Kazuyuki Tanaka, Takeo Hoshi, Akiyoshi Kuwata, Yusaku Yamamoto, Yuki Harada, Hiroto Imachi, Takeshi Fukaya |
|---|---|
| Jazyk: | angličtina |
| Rok vydání: | 2018 |
| Předmět: |
Computer science
FOS: Physical sciences 010103 numerical & computational mathematics Parallel computing 01 natural sciences symbols.namesake FOS: Mathematics Performance prediction Mathematics - Numerical Analysis 0101 mathematics Condensed Matter - Materials Science ScaLAPACK Applied Mathematics General Engineering Materials Science (cond-mat.mtrl-sci) Markov chain Monte Carlo Numerical Analysis (math.NA) Computational Physics (physics.comp-ph) Solver Supercomputer 010101 applied mathematics Middleware Scalability Benchmark (computing) symbols Computer Science::Mathematical Software Physics - Computational Physics |
| Popis: | An open-source middleware EigenKernel was developed for use with parallel generalized eigenvalue solvers or large-scale electronic state calculation to attain high scalability and usability. The middleware enables the users to choose the optimal solver, among the three parallel eigenvalue libraries of ScaLAPACK, ELPA, EigenExa and hybrid solvers constructed from them, according to the problem specification and the target architecture. The benchmark was carried out on the Oakforest-PACS supercomputer and reveals that ELPA, EigenExa and their hybrid solvers show better performance, when compared with pure ScaLAPACK solvers. The benchmark on the K computer is also used for discussion. In addition, a preliminary research for the performance prediction was investigated, so as to predict the elapsed time T as the function of the number of used nodes P (T=T(P)). The prediction is based on Bayesian inference using the Markov Chain Monte Carlo (MCMC) method and the test calculation indicates that the method is applicable not only to performance interpolation but also to extrapolation. Such a middleware is of crucial importance for application-algorithm-architecture co-design among the current, next-generation (exascale), and future-generation (post-Moore era) supercomputers. 25 pages, 8 figures |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|