arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2405.15380cs.ARcs.AI
收藏

RISC-V系统机器学习推理工作负载的全栈评估

Full-stack evaluation of Machine Learning inference workloads for RISC-V systems

imec(微电子中心)

Debjyoti Bhattacharjee, Anmol, Tommaso Marinelli, Karan Pathak, Peter Kourzanov

AI总结:

本研究利用gem5模拟器和基于MLIR的编译工具链,对RISC-V架构上的机器学习推理工作负载进行了全栈性能评估,并指出了gem5的现有局限。

摘要:

架构模拟器在RISC-V研究中扮演着至关重要的角色,为工作负载评估提供了关键平台,无需昂贵的物理原型。它们作为探索创新架构概念的动态环境,能够实现快速迭代和性能指标的深入分析。随着深度学习算法日益普及,使用机器学习工作负载对新架构进行基准测试至关重要。深度学习算法中使用的多样化计算内核凸显了构建全面编译工具链以映射到目标硬件平台的必要性。本研究使用gem5(一个开源架构模拟器)评估了RISC-V架构上多种机器学习工作负载的性能。利用基于多级中间表示(MLIR)的开源编译工具链,研究提供了专门针对深度学习推理工作负载的基准测试结果。此外,研究还揭示了当前gem5在模拟RISC-V架构时的局限性,为未来的开发和完善提供了见解。

英文摘要:

Architectural simulators hold a vital role in RISC-V research, providing a crucial platform for workload evaluation without the need for costly physical prototypes. They serve as a dynamic environment for exploring innovative architectural concepts, enabling swift iteration and thorough analysis of performance metrics. As deep learning algorithms become increasingly pervasive, it is essential to benchmark new architectures with machine learning workloads. The diverse computational kernels used in deep learning algorithms highlight the necessity for a comprehensive compilation toolchain to map to target hardware platforms. This study evaluates the performance of a wide array of machine learning workloads on RISC-V architectures using gem5, an open-source architectural simulator. Leveraging an open-source compilation toolchain based on Multi-Level Intermediate Representation (MLIR), the research presents benchmarking results specifically focused on deep learning inference workloads. Additionally, the study sheds light on current limitations of gem5 when simulating RISC-V architectures, offering insights for future development and refinement.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑