Abstract:

Whole Device Modeling (WDM) is generally described as assembling physics models that provide an integrated simulation of the plasma. The integrated WDM requires a fidelity hierarchy of multiphysics, multi-scale computational models. The High-Fidelity Whole Device Model of Magnetically Confined Fusion Plasma (WDMApp) is a prominent application in the DOE Exascale Computing Project (ECP). The 10-year problem target of the project is the high-fidelity simulation of whole device burning plasmas applicable to an advanced tokamak regime (specifically, an ITER steady-state plasmas with ten-fold energy gain), integrating the effects of energetic particles, plasma-material interactions, heating, and current drive. The most important step in the project, and one that involves the highest risk, is the coupling of two existing, well-established, extreme-scale gyrokinetic codes – the GENE continuum code for the core plasma, and the XGC particle-in-cell (PIC) code for the boundary plasma. The GEM PIC code is also used in the core as a test of flexibility of the extensible framework EFFIS 2.0 (End-to-End Framework for Fusion Integrated Simulations 2.0) and for risk mitigation. We have accomplished this challenging milestone for the first time in the magnetic fusion community for electrostatic plasmas by developing and implementing novel algorithms for both GENE-XGC and GEM-XGC coupling. The execution of this challenge problem required the optimization of the WDMApp codes (GENE, GEM and XGC) on pre-exascale and exascale computers, leveraging the ECP Co-Design and Software Technologies projects for portability and performance. WDMApp is a DOE 413.3b project with the objective of realizing by 2023 the demostration of pedestal formation on an experimentally realistic time-scale for ITER, and a Figure of Merit exceeding 300 on exascale platforms, accomplished through algorthmic advances, performance engineering, and hardware improvements. Building on the legacy of WDMApp and the Simons Collaboration on “Hidden Symmetries and Fusion Energy,” we are well-positioned to carry out stellarator high-fidelity simulations over approximately the next five years. This research is supported by the DOE Office of Advanced Scientific Computing Research under the auspices of the ECP.