Review on "Earth System Modelling on System-level Heterogeneous Architectures: EMAC (version 2.42) on the Dynamical Exascale Entry Platform (DEEP)"
2016
The article reports about one part of the DEEP project examining the approach of heterogeneous cluster-computing for Earth System models.In particular, this article is about the GCCM EMAC.A performance analysis shows, that the chemistry submodel MECCA is the bottleneck of usual EMAC simulations.As MECCA is solving the chemistry within each grid box independently, it is embarrassing parallel and thus the ideal candidate for application on a Booster architecture.The article addresses a very important issue of Earth system modelling.The fact that these codes have a long history brings about that they are not at all optimised for modern computer architectures.Thus the refactoring of the code, to used the possibilities provided by current and future computing architectures is a very important issue.