A Scalability Study of the Ice-sheet and Sea-level System Model (ISSM, Version 4.18)

Abstract

Abstract. Accurately modeling the contribution of Greenland and Antarctica to sea level rise requires to solve partial differential equations at a high spatial resolution. It is important to test the scalability of existing ice sheet models in order to assess whether they are ready to take advantage of new cluster architectures. In this paper, we discuss the overall scaling of the Ice-sheet and Sea-level System Model (ISSM) applied to the Greenland ice sheet. The model setup used as benchmark problem comprises a variety of modules with different levels of complexity and computational demands. The core builds the so-called stress balance module, which uses the higher-order approximation (or Blatter-Pattyn) of the Stokes equations and a mesh of linear prismatic finite elements to compute the ice flow. We develop a detailed user-oriented, yet low-overhead performance instrumentation tailored to the requirements of earth system models and run scaling tests up to 6 144 MPI processes. The results show that the computation of the Greenland model scales overall well up to 3 072 MPI processes, but is eventually slowed down by matrix assembly, the output handling, and lower-dimensional problems that employ lower numbers of unknowns per MPI process. We also discuss improvements of the scaling and identify further improvements needed for climate research. The instrumented version of ISSM, thus, not only identifies potential performance bottlenecks that were not present at lower core counts but also provides the capability to continually monitor the performance of ISSM code basis. This is of long-term significance as the overall performance of ISSM model depends on the subtle interplay between algorithms, their implementation, underlying libraries, compilers, run-time systems and hardware characteristics, all of which are in a constant state of flux.

Performance portable ice-sheet modeling with MALI

“…The computational component of the hybrid "SSA + SIA" model in PISM is found to scale well on up to 1024 cores in Dickens (2015) but the I/O component is found to scale poorly. In Fischler et al (2021), low-overhead performance instrumentation is developed for the "Blatter-Pattyn" model in ISSM and good scaling is found on up to 3072 cores. The study finds that matrix assembly and I/O begins to scale poorly and highlights the importance of continuous performance monitoring.…”