Search

Graphs and matrices are widely used in algorithms for social network analyses. Since the number of interactions is much less than the possible number of interactions, the graphs and matrices used in the analyses are usually sparse. In this paper, we propose an efficient implementation of a sparse-matrix computation which arises in our publicly available citation recommendation service called the advisor. The recommendation algorithm uses a sparse matrix generated from the citation graph. We observed that the nonzero pattern of this matrix is highly irregular and the computation suffers from high number of cache misses. We propose techniques for storing the matrix in memory efficiently and reducing the number of cache misses. Experimental results show that our techniques are highly efficient on reducing the query processing time which is highly crucial for a web service. [ABSTRACT FROM PUBLISHER]

We discuss efficient shared memory parallelization of sparse matrix computations whose main traits resemble to those of the sparse matrix-vector multiply operation. Such computations are difficult to parallelize because of the relatively small computational granularity characterized by small number of operations per each data access. Our main application is a sparse matrix scaling algorithm which is more memory bound than the sparse matrix vector multiplication operation. We take the application and parallelize it using the standard OpenMP programming principles. Apart from the common race condition avoiding constructs, we do not reorganize the algorithm. Rather, we identify associated performance metrics and describe models to optimize them. By using these models, we implement parallel matrix scaling algorithms for two well-known sparse matrix storage formats. Experimental results show that simple parallelization attempts which leave data/work partitioning to the runtime scheduler can suffer from the overhead of avoiding race conditions especially when the number of threads increases. The proposed algorithms perform better than these algorithms by optimizing the identified performance metrics and reducing the overhead. [ABSTRACT FROM PUBLISHER]