This paper focuses on how to optimize the cache performance of sparse matrix-matrix multiplication(SpGEMM).It classifies the cache misses into two categories;one is caused by the irregular distribution pattern of the ...This paper focuses on how to optimize the cache performance of sparse matrix-matrix multiplication(SpGEMM).It classifies the cache misses into two categories;one is caused by the irregular distribution pattern of the multiplier-matrix,and the other is caused by the multiplicand.For each of them,the paper puts forward an optimization method respectively.The first hash based method removes cache misses of the 1 st category effectively,and improves the performance by a factor of 6 on an Intel 8-core CPU for the best cases.For cache misses of the 2nd category,it proposes a new cache replacement algorithm,which achieves a cache hit rate much higher than other historical knowledge based algorithms,and the algorithm is applicable on CELL and GPU.To further verify the effectiveness of our methods,we implement our algorithm on GPU,and the performance perfectly scales with the size of on-chip storage.展开更多
矩阵转置是矩阵运算的基本操作,广泛应用于信号处理、科学计算以及深度学习等各种领域。随着国防科技大学自主研制的飞腾异构多核数字信号处理器(digital signal processor, DSP)在各种领域中的推广应用,对高性能矩阵转置实现提出了强...矩阵转置是矩阵运算的基本操作,广泛应用于信号处理、科学计算以及深度学习等各种领域。随着国防科技大学自主研制的飞腾异构多核数字信号处理器(digital signal processor, DSP)在各种领域中的推广应用,对高性能矩阵转置实现提出了强烈需求。针对飞腾异构多核DSP的体系结构特征与矩阵转置操作的特点,提出了一种适配不同数据位宽(8 B、4 B以及2 B)矩阵的并行矩阵转置算法ftmMT。该算法基于DSP中向量处理单元的Load/Store部件实现了向量化,同时基于矩阵分块实现了多个DSP核的并行处理,通过隐式乒乓设计实现了片上向量化转置与片外访存的重叠以及访存性能的大幅提升。实验结果表明,ftmMT能够显著加快矩阵转置操作,与CPU上的开源转置库HPTT相比,可获得高达8.99倍的性能加速。展开更多
基金Supported by the National High Technology Research and Development Programme of China(No.2010AA012302,2009AA01 A134)Tsinghua National Laboratory for Information Science and Technology(TNList)Cross-discipline Foundation
文摘This paper focuses on how to optimize the cache performance of sparse matrix-matrix multiplication(SpGEMM).It classifies the cache misses into two categories;one is caused by the irregular distribution pattern of the multiplier-matrix,and the other is caused by the multiplicand.For each of them,the paper puts forward an optimization method respectively.The first hash based method removes cache misses of the 1 st category effectively,and improves the performance by a factor of 6 on an Intel 8-core CPU for the best cases.For cache misses of the 2nd category,it proposes a new cache replacement algorithm,which achieves a cache hit rate much higher than other historical knowledge based algorithms,and the algorithm is applicable on CELL and GPU.To further verify the effectiveness of our methods,we implement our algorithm on GPU,and the performance perfectly scales with the size of on-chip storage.
