MatRaptor: A Sparse-Sparse Matrix Multiplication Accelerator Based on Row-Wise Product
MatRaptor: A Sparse-Sparse Matrix Multiplication Accelerator Based on Row-Wise Product
复制标题
DOI：
10.1109/micro50266.2020.00068
复制
发表时间：
2020-10
期刊：
影响因子：
--
通讯作者：
Nitish Srivastava;Hanchen Jin;Jie Liu-;D. Albonesi;Zhiru Zhang
中科院分区：
文献类型：
--
作者：
Nitish Srivastava;Hanchen Jin;Jie Liu-;D. Albonesi;Zhiru Zhang
稀疏 - 矩阵乘法（SPGEMM）是一种在许多应用领域中广泛使用的计算内核，例如数据分析，图形处理和科学计算。与使用内部或外部产品作为用于矩阵乘法的元操作的常规方法不同，我们的方法基于行明，该产品在行列中提供了更好的权衡数据重复使用和片上记忆要求，并为大型稀疏材料实现更高的性能。高利用记忆带宽，我们在我们的实验中使用GEM5模拟我们的加速器体系结构。与Outerspace相比，单线读取的CPU，GPU上的8.8倍加速度和1.8倍加速器（OuterSpace）上的1.8倍加速器（Outerspace）也具有7.2倍较低的功率消耗和31.3×较小的区域。
Sparse-sparse matrix multiplication (SpGEMM) is a computation kernel widely used in numerous application domains such as data analytics, graph processing, and scientific computing. In this work we propose MatRaptor, a novel SpGEMM accelerator that is high performance and highly resource efficient. Unlike conventional methods using inner or outer product as the meta operation for matrix multiplication, our approach is based on row-wise product, which offers a better tradeoff in terms of data reuse and on-chip memory requirements, and achieves higher performance for large sparse matrices. We further propose a new hardware-friendly sparse storage format, which allows parallel compute engines to access the sparse data in a vectorized and streaming fashion, leading to high utilization of memory bandwidth. We prototype and simulate our accelerator architecture using gem5 on a diverse set of matrices. Our experiments show that MatRaptor achieves 129.2× speedup over single-threaded CPU, 8.8× speedup over GPU and 1.8× speedup over the state-of-the-art SpGEMM accelerator (OuterSPACE). MatRaptor also has 7.2× lower power consumption and 31.3× smaller area compared to OuterSPACE.