Session CM11: Mini-Conference on Integrated, Multiphysics, High-performance Computations for Magnetic Fusion Research

Integrated, multiphysics, high-performance computations on EFFIS framework in CPES

Overview of important multiscale physics results, enabled by the high-performance integrated simulation tool set EFFIS (End- to-end Framework for Fusion Integrated Simulation), will be presented. Three multiscale physics integration examples to be summarized are a) kinetic pedestal growth, magnetic equilibrium reconstruction, linear stability check, ELM crash, and divertor heat-load profile detection; b)kinetic transport modeling and electromagnetic micro-turbulences; and c) RMP penetration and kinetic pedestal transport. EFFIS performs the real-time integrated simulation with minimal intrusion into the component codes, and allows for independent choices of compilers by the component codes. The component codes can be of any kind, ranging from small scale memory intensive codes to extreme scale parallel processing codes. It can even be a binary executable. Memory-to-memory and file-to-file coupling can be mixed together within one real- time integrated simulation. The results are monitored in a web- based dashboard for collaborative physics analysis and validation, with all the provenance information automatically captured.   

Overview of the Simulation of Wave Interactions with MHD Project (SWIM)

The SWIM center has the scientific objectives of: improving our understanding of interactions that both RF wave and particle sources have on extended-MHD phenomena, improving our capability for predicting and optimizing the performance of burning plasmas, developing an integrated computational system for treating multi-physics phenomena with the required flexibility and extensibility to serve as a prototype for the Fusion Simulation Project, addressing mathematics issues related to the multi-scale, coupled physics of RF waves and extended MHD, and optimizing the integrated system on high performance computers. Our Center has now built an end-to-end computational system that allows existing physics codes to be able to function together in a parallel environment and connects them to utility software components and data management systems. We have used this framework to couple together state-of-the-art fusion energy codes to produce a unique and world-class simulation capability. A physicist's overview of the Integrated Plasma Simulator (IPS) will be given and applications described. For example the IPS is being employed to support ITER with operational scenario studies.   

FACETS for Multiphysics, Whole-Fusion-Device Modeling

FACETS is developing a multi-physics, multi-region computational application for whole-fusion-device modeling. The FACETS framework supports (1) concurrently executing components, (2) usability on platforms from laptops to Leadership Class Facilities (LCFs), (3) ability to reuse legacy code, (4) ability to include or exclude modules at link time, and (5) ability to instantiate particular implementations (e.g., for different transport fluxes) at runtime. FACETS has developed a parallel core solver (using nested iteration for improved convergence) that speeds up core transport calculations by 1.5 orders of magnitude. FACETS has multiple models for transport fluxes, which it obtains through the FMCFM generic interfaces. FACETS is currently using UEDGE for edge transport, with transport coefficients taken from experiment, and it is using NUBEAM for neutral beam injection. On the horizon are the inclusion of RF sources, free-boundary equilibrium, and wall modeling. FACETS has been benchmarked against ASTRA for core transport, and has now provided first results for core-edge simulations. The latter show early time experimental edge diffusivities can be used to predict self-consistent pedestal buildup.   

CASL: The Consortium for Advanced Simulation of Light Water Reactors

Like the fusion community, the nuclear engineering community is embarking on a new computational effort to create integrated, multiphysics simulations. The Consortium for Advanced Simulation of Light Water Reactors (CASL), one of 3 newly-funded DOE Energy Innovation Hubs, brings together an exceptionally capable team that will apply existing modeling and simulation capabilities and develop advanced capabilities to create a usable environment for predictive simulation of light water reactors (LWRs). This environment, designated the Virtual Reactor (VR), will: 1) Enable the use of leadership-class computing for engineering design and analysis to improve reactor capabilities, 2) Promote an enhanced scientific basis and understanding by replacing empirically based design and analysis tools with predictive capabilities, 3) Develop a highly integrated multiphysics environment for engineering analysis through increased fidelity methods, and 4) Incorporate UQ as a basis for developing priorities and supporting, application of the VR tools for predictive simulation. In this presentation, we present the plans for CASL and comment on the similarity and differences with the proposed Fusion Simulation Project (FSP).   

End-to-end Framework for Fusion Integrated Simulation (EFFIS)

EFFIS is a set of tools developed for working with large-scale simulations. EFFIS is used by researchers in the Center for Plasma Edge Simulation, as well as many other areas of science. EFFIS is composed of services including adaptable I/O, workflows, dashboards, visualization, code coupling, wide-area data movement, and provenance capturing. One of the unique aspects of EFFIS is that it transparently allows users to switch from code coupling on disk to coupling in memory, using the concept of a shared space in a staging area. The staging area is a small fraction of the compute nodes needed to run the large-scale simulation, but it is used for the construction of I/O pipelines and a code-coupling infrastructure. This allows the scientist to make minor changes for the code to work with ADIOS), and then with no changes perform complex transformations, and analytics, which all occur \textit{in situ} with the simulation. In this talk, we will focus on the technologies CPES uses, which are scalable and can be used on anything from workstations to petascale machines.   

Component Framework for Loosely Coupled High Performance Integrated Plasma Simulations

We present the design and implementation of a component-based simulation framework for the execution of coupled time-dependent plasma modeling codes. The Integrated Plasma Simulator (IPS) provides a flexible lightweight component model that streamlines the integration of stand alone codes into coupled simulations. Standalone codes are adapted to the IPS component interface specification using a thin wrapping layer implemented in the Python programming language. The framework provides services for inter-component method invocation, configuration, task, and data management, asynchronous event management, simulation monitoring, and checkpoint/restart capabilities. Services are invoked, as needed, by the computational components to coordinate the execution of different aspects of coupled simulations on Massive parallel Processing (MPP) machines. A common plasma state layer serves as the foundation for inter-component, file-based data exchange. The IPS design principles, implementation details, and execution model will be presented, along with an overview of several use cases.   

Progress in Parallel Implicit Methods for Tokamak Edge Plasma Modeling

Performance of prototype tokamak fusion devices depends sensitively on characteristics of the edge plasma between the hot core and surrounding walls. ~The edge plasma includes an especially wide range of physical time scales such that implicit numerical algorithms can substantially improve overall computational efficiency. Here some of the benefits and challenges of parallel implicit solution strategies are presented, with emphasis on preconditioned Newton-Krylov methods in the UEDGE and BOUT++ applications. ~Related multi-component issues are discussed in the context of the FACETS project, which is developing an integrated, parallel application to simulate physical processes from the material wall to the plasma core. ~It is demonstrated how this fusion research is motivating new capabilities in the PETSc equation solver library to better handle strong coupling between two or more distinct PDE-based mathematical models.