Parallelism strategies for finite difference acoustic wave simulation operators on GPUs

• Grant number: 22/00434-0
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: March 01, 2022
• End date: June 30, 2023
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Hermes Senger
• Grantee: Letícia Suellen Farias Machado
• Host Institution: Centro de Ciências Exatas e de Tecnologia (CCET). Universidade Federal de São Carlos (UFSCAR). São Carlos , SP, Brazil
• Associated scholarship(s): 22/11070-9 - Optimization strategies for the acoustic wave stencil on multicore systems, BE.EP.IC

Abstract

Algorithms for numerically solving partial differential equations (PDEs) are used in different application domains to approximate the physics of interest. In the context of seismic applications, the process has a high computational cost as they constitute the computational core of inverse problems involving computation in the order of Exascale. The demand for high performance implies that every code must be highly optimized for diverse and rapidly evolving computer architectures. The general objective of this research is to investigate operator code optimization strategies for numerical simulation of acoustic wave propagation for execution on GPUs. Such codes implement well-known stencil patterns that can benefit from the massive parallelism of GPUs but lead to a performance bottleneck in memory accesses.(AU)

Scientific publications

(References retrieved automatically from Web of Science and SciELO through information on FAPESP grants and their corresponding numbers as mentioned in the publications by the authors)

MACHADO, LETICIA S. F.; TADONKI, CLAUDE; SENGER, HERMES. A Source-to-source NUMA Profiling Approach. 2023 INTERNATIONAL SYMPOSIUM ON COMPUTER ARCHITECTURE AND HIGH PERFORMANCE COMPUTING WORKSHOPS, SBAC-PADW, v. N/A, p. 6-pg., 2023-01-01. (19/26702-8, 22/00434-0, 22/11070-9, 23/00566-6)