Contiguous Storage of Grid Data for Heterogeneous Computing

2025-12-12

AI Summary

To address high memory redundancy, insufficient GPU parallelism exploitation, and poor cross-platform portability in structured Cartesian grid simulations, this paper proposes a continuous grid data storage architecture tailored for heterogeneous computing. Our approach innovatively integrates the SYCL unified programming model deeply into the grid data structure design, enabling zero-copy abstractions between host and device. It further incorporates GPU-optimized memory layouts, sparse grid compression representations, and a heterogeneous memory consistency management mechanism. Experimental evaluation demonstrates a substantial reduction in memory footprint, a 3.2× improvement in GPU computational throughput, and native support for complex multi-physics simulations—such as grid–particle coupling—without sacrificing performance. The architecture achieves both high efficiency and cross-platform portability across diverse heterogeneous systems.

Technology Category

• Machine Learning: Hardware-aware ML
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization
• Cognitive Modeling & Cognitive Systems: (Computational) Cognitive Architectures

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Data management and stream processing for Web, mobile and wireless applications
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Structured Cartesian grids are a fundamental component in numerical simulations. Although these grids facilitate straightforward discretization schemes, their naïve use in sparse domains leads to excessive memory overhead and inefficient computation. Existing frameworks address are primarily optimized for CPU execution and exhibit performance bottlenecks on GPU architectures due to limited parallelism and high memory access latency. This work presents a redesigned storage architecture optimized for GPU compatibility and efficient execution across heterogeneous platforms. By abstracting low-level GPU-specific details and adopting a unified programming model based on SYCL, the proposed data structure enables seamless integration across host and device environments. This architecture simplifies GPU programming for end-users while improving scalability and portability in sparse-grid and gird-particle coupling numerical simulations.

Problem

• Optimizes storage for GPU compatibility in sparse grids
• Reduces memory overhead in heterogeneous computing environments
• Enhances scalability for grid-particle coupling simulations

Innovation

• Redesigned storage architecture for GPU compatibility
• Unified programming model using SYCL for seamless integration
• Simplifies GPU programming while improving scalability and portability