CoSPARSE: A Software and Hardware Reconfigurable SpMV Framework for Graph Analytics
CoSPARSE: A Software and Hardware Reconfigurable SpMV Framework for Graph Analytics
复制标题
DOI：
10.1109/dac18074.2021.9586114
复制
发表时间：
2021-12
期刊：
影响因子：
--
通讯作者：
Siying Feng;Jiawen Sun;S. Pal;Xin He;Kuba Kaszyk;Dong-hyeon Park;J. Morton;T. Mudge;M. Cole;M. O’Boyle;C. Chakrabarti;R. Dreslinski
中科院分区：
文献类型：
--
作者：
Siying Feng;Jiawen Sun;S. Pal;Xin He;Kuba Kaszyk;Dong-hyeon Park;J. Morton;T. Mudge;M. Cole;M. O’Boyle;C. Chakrabarti;R. Dreslinski
稀疏矩阵向量乘法（SpMV）是迭代图分析算法的关键构建块。通常，这样的算法在迭代中具有变化的活动顶点集。这种可变性已被用于通过在迭代（软件）之间动态切换算法或为图形分析算法设计定制加速器（硬件）来提高性能。在这项工作中，我们提出了一个新的框架，CoSPARSE，采用硬件和软件重新配置作为一个协同的解决方案，以加速基于SpMV的图分析算法。基于先前提出的通用可重构硬件，我们实现了CoSPARSE作为一个软件层，抽象的硬件作为一个专门的SpMV加速器。CoSPARSE为每次迭代动态选择软件和硬件配置，与没有重新配置的简单实现相比，最大加速比为2.0倍。在一套图形算法中，CoSPARSE在Xeon CPU上的性能优于最先进的共享内存框架Ligra，性能提高了3.51倍，能效提高了877倍。
Sparse matrix-vector multiplication (SpMV) is a critical building block for iterative graph analytics algorithms. Typically, such algorithms have a varying active vertex set across iterations. This variability has been used to improve performance by either dynamically switching algorithms between iterations (software) or designing custom accelerators (hardware) for graph analytics algorithms. In this work, we propose a novel framework, CoSPARSE, that employs hardware and software reconfiguration as a synergistic solution to accelerate SpMV-based graph analytics algorithms. Building on previously proposed general-purpose reconfigurable hardware, we implement CoSPARSE as a software layer, abstracting the hardware as a specialized SpMV accelerator. CoSPARSE dynamically selects software and hardware configurations for each iteration and achieves a maximum speedup of 2.0 × compared to the naïve implementation with no reconfiguration. Across a suite of graph algorithms, CoSPARSE outperforms a state-of-the-art shared memory framework, Ligra, on a Xeon CPU with up to 3.51 × better performance and 877 × better energy efficiency.