Exciting news! We're transitioning to the Statewide California Earthquake Center. Our new website is under construction, but we'll continue using this website for SCEC business in the meantime. We're also archiving the Southern Center site to preserve its rich history. A new and improved platform is coming soon!

Porting AWP-ODC to next generation architectures: the Characteristic Scientific Application project

Yifeng Cui, Scott Callaghan, Yiming Ren, Lars Koesterke, Wenyang Zhang, Mattew Norman, Daniel Roten, Philip J. Maechling, & Christine A. Goulet

Submitted September 11, 2022, SCEC Contribution #12322, 2022 SCEC Annual Meeting Poster #231

AWP-ODC is a SCEC community software program designed, and widely used, to perform large-scale, highly-parallel, physics-based earthquake rupture and wave propagation simulations. AWP-ODC currently achieves excellent speedup and efficiency on multi-core CPU and GPU-based architectures. The wave propagation solver in AWP-ODC faces significant challenges on new and emerging supercomputers due to the increased parallelism of the complex next generation processor and accelerator architectures that require substantial code structure changes to obtain excellent performance. This poster introduces an early science project for the Leadership-Class Computing Facility (LCCF) at Texas Advanced Computing Center (TACC), funded by NSF’s Characteristic Science Applications (CSA) program. The goal of our SDSC and SCEC CSA project is to improve the performance of the AWP-ODC software on new types of computing hardware and to generate benchmarks for the next generation LCCF architectures at TACC. We will also present a prototype version of GPU-based CyberShake AWP-ODC-GPU-SGT, in C that is designed for AMD’s next-generation architecture. This HIP-based code is designed to target Oak Ridge National Laboratory (OLCF)’s Frontier, an Exascale system based on the AMD Radeon Instinct GPUs and EPYC CPUs. We demonstrate the studies of the scaling and current performance status.

Key Words
AWP-ODC, LCCF, CSA, TACC, High performance computing, wave propagation

Citation
Cui, Y., Callaghan, S., Ren, Y., Koesterke, L., Zhang, W., Norman, M., Roten, D., Maechling, P. J., & Goulet, C. A. (2022, 09). Porting AWP-ODC to next generation architectures: the Characteristic Scientific Application project. Poster Presentation at 2022 SCEC Annual Meeting.


Related Projects & Working Groups
Ground Motions