Matrix-Free Finite Volume Kernels on a Dataflow Architecture

Fast and accurate numerical simulations are crucial for designing large-scale geological carbon storage projects ensuring safe long-term CO 2 containment as a climate change mitigation strategy. These simulations involve solving numerous large and complex linear systems arising from the implicit Fin...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:SC24: International Conference for High Performance Computing, Networking, Storage and Analysis S. 1 - 11
Hauptverfasser: Sai, Ryuichi, Hamon, Francois P., Mellor-Crummey, John, Araya-Polo, Mauricio
Format: Tagungsbericht
Sprache:Englisch
Veröffentlicht: IEEE 17.11.2024
Schlagworte:
Bandwidth
Climate change
Computational modeling
dataflow architecture
distributed memory
energy
finite-volume
high-performance computing
In-memory computing
Linear systems
matrix-free linear solver
Numerical models
Numerical simulation
Prevention and mitigation
Solid modeling
wafer-scale engine
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Fast and accurate numerical simulations are crucial for designing large-scale geological carbon storage projects ensuring safe long-term CO 2 containment as a climate change mitigation strategy. These simulations involve solving numerous large and complex linear systems arising from the implicit Finite Volume (FV) discretization of PDEs governing subsurface fluid flow. Compounded with highly detailed geomodels, solving linear systems is computationally and memory expensive, and accounts for the majority of the simulation time. Modern memory hierarchies are insufficient to meet the latency and bandwidth needs of large-scale numerical simulations. Therefore, exploring algorithms that can leverage alternative and balanced paradigms such as dataflow and in-memory computing is crucial. This work introduces a matrix-free algorithm to solve FV-based linear systems using a dataflow architecture to significantly minimize memory latency and bandwidth bottlenecks. Our implementation achieves two orders of magnitude speedup compared to a GPGPU-based reference implementation, and up to 1.2 PFlops on a single dataflow device.
DOI:10.1109/SC41406.2024.00034