Sparse Matrix Storage Format
نویسندگان
چکیده
Operations on Sparse Matrices are the key computational kernels in many scientific and engineering applications. They are characterized with poor substantiated performance. It is not uncommon for microprocessors to gain only 10-20% of their peak floating-point performance when doing sparse matrix computations even when special vector processors have been added as coprocessor facilities. In this paper we present new data format for sparse matrix storage. This format facilitates the continuous reuse of elements in the processing array. In comparison to other formats we achieve lower storage efficiency (only an extra bit per non-zero elements). A conjuncture of the proposed approach is that the hardware execution efficiency on sparse matrices can be improved. Keywords—Sparse Matrix Formats, Operation Efficiency, Hardware
منابع مشابه
A Hilbert-order multiplication scheme for unstructured sparse matrices
We investigate a new storage format for unstructured sparse matrices, based on the space filling Hilbert curve. Numerical tests with matrix-vector multiplication show the potential of the fractal storage format (FS) in comparison to the traditional compressed row storage format (CRS). The FS format outperforms the CRS format by up to 50% for matrix-vector multiplications with multiple right han...
متن کاملData Structures and Algorithms for Distributed Sparse Matrix Operations
We propose extensions of the classical row compressed storage format for sparse matrices. The extensions are designed to accomodate distributed storage of the matrix. We outline an implementation of the matrix-vector product using this distributed storage format, and give algorithms for building and using the communication structure between processors.
متن کاملA Hierarchical Sparse Matrix Storage Format for Vector Processors
We describe and evaluate a Hierarchical Sparse Matrix (HiSM) storage format designed to be a unified format for sparse matrix applications on vector processors. The advantages that the format offers are low storage requirements, a flexible structure for element manipulations and allowing for efficient operations. To take full advantage of the format we also propose a vector architecture extensi...
متن کاملVectorized Sparse Matrix Multiply for Compressed Row Storage Format
The innovation of this work is a simple vectorizable algorithm for performing sparse matrix vector multiply in compressed sparse row (CSR) storage format. Unlike the vectorizable jagged diagonal format (JAD), this algorithm requires no data rearrangement and can be easily adapted to a sophisticated library framework such as PETSc. Numerical experiments on the Cray X1 show an order of magnitude ...
متن کاملPerformance Comparison of Storage Formats for Sparse Matrices
The sparse data structure represents a matrix in space proportional to the number of non-zero entries. Many storage formats have been proposed to represent sparse matrices. In this paper we evaluate and compare the storage efficiency of various sparse matrix storage formats, and consider the performance results of matrix-vector multiplication using these storage formats.
متن کامل