Matches in SemOpenAlex for { <https://semopenalex.org/work/W2003840611> ?p ?o ?g. }
- W2003840611 abstract "Sparse matrix-sparse/dense matrix multiplications, spgemm and csrmm, respectively, among other applications find usage in various matrix formulations of graph problems. Considering the difficulties in executing graph problems and the duality between graphs and matrices, computations such as spgemm and csrmm have recently caught the attention of HPC community. These computations pose challenges such as load balancing, irregular nature of the computation, and difficulty in predicting the output size. It is even more challenging when combined with the GPU architectural constraints such as memory accesses, limited shared memory, strict SIMD and thread execution. To address these challenges on a GPU, we evaluate three possible variations of matrix multiplication (Row-Column, Column-Row, Row-Row) and perform suitable optimizations targeted at sparse matrices. Our experiments indicate that the Row-Row formulation, which mostly outperforms the other formulations, is 3.5x faster on average compared to an optimized multi-core implementation in the Intel MKL library. We extend the Row-Row formulation to a CPU+GPU hybrid algorithm that simultaneously utilizes the CPU also. In this direction, we present heuristics to find the right amount of work division between the CPU and the GPU. Our hybrid row-row formulation of the spgemm operation performs 5.5x faster on average when compared to the optimized multi-core implementation in the Intel MKL library. Our experience indicates that it is difficult to identify right amount of work division between the CPU and the GPU. We therefore investigate a subclass of sparse matrices, band matrices, and present an analytical method to identify a good work division when multiplying two band matrices. Our GPU csrmm operation performs 2.5x faster on average when compared to a corresponding implementation in the cusparse library, which outperforms the Intel MKL library implementation." @default.
- W2003840611 created "2016-06-24" @default.
- W2003840611 creator A5011731495 @default.
- W2003840611 creator A5023412602 @default.
- W2003840611 creator A5076139739 @default.
- W2003840611 date "2012-12-01" @default.
- W2003840611 modified "2023-09-30" @default.
- W2003840611 title "Sparse matrix-matrix multiplication on modern architectures" @default.
- W2003840611 cites W1520511539 @default.
- W2003840611 cites W1974610639 @default.
- W2003840611 cites W1984419037 @default.
- W2003840611 cites W1990832096 @default.
- W2003840611 cites W1998992875 @default.
- W2003840611 cites W2007259720 @default.
- W2003840611 cites W2029311758 @default.
- W2003840611 cites W2043670592 @default.
- W2003840611 cites W2049500052 @default.
- W2003840611 cites W2056999868 @default.
- W2003840611 cites W2089437293 @default.
- W2003840611 cites W2089939118 @default.
- W2003840611 cites W2118558147 @default.
- W2003840611 cites W2120119701 @default.
- W2003840611 cites W2122948532 @default.
- W2003840611 cites W2128853364 @default.
- W2003840611 cites W2159481344 @default.
- W2003840611 cites W2162322364 @default.
- W2003840611 cites W2166955231 @default.
- W2003840611 cites W2169150754 @default.
- W2003840611 cites W2544965371 @default.
- W2003840611 cites W3141650078 @default.
- W2003840611 cites W4301491118 @default.
- W2003840611 doi "https://doi.org/10.1109/hipc.2012.6507483" @default.
- W2003840611 hasPublicationYear "2012" @default.
- W2003840611 type Work @default.
- W2003840611 sameAs 2003840611 @default.
- W2003840611 citedByCount "34" @default.
- W2003840611 countsByYear W20038406112014 @default.
- W2003840611 countsByYear W20038406112015 @default.
- W2003840611 countsByYear W20038406112016 @default.
- W2003840611 countsByYear W20038406112017 @default.
- W2003840611 countsByYear W20038406112018 @default.
- W2003840611 countsByYear W20038406112019 @default.
- W2003840611 countsByYear W20038406112020 @default.
- W2003840611 countsByYear W20038406112021 @default.
- W2003840611 countsByYear W20038406112023 @default.
- W2003840611 crossrefType "proceedings-article" @default.
- W2003840611 hasAuthorship W2003840611A5011731495 @default.
- W2003840611 hasAuthorship W2003840611A5023412602 @default.
- W2003840611 hasAuthorship W2003840611A5076139739 @default.
- W2003840611 hasConcept C106487976 @default.
- W2003840611 hasConcept C111919701 @default.
- W2003840611 hasConcept C11413529 @default.
- W2003840611 hasConcept C114614502 @default.
- W2003840611 hasConcept C121332964 @default.
- W2003840611 hasConcept C126042441 @default.
- W2003840611 hasConcept C138101251 @default.
- W2003840611 hasConcept C150552126 @default.
- W2003840611 hasConcept C159985019 @default.
- W2003840611 hasConcept C163716315 @default.
- W2003840611 hasConcept C17349429 @default.
- W2003840611 hasConcept C173608175 @default.
- W2003840611 hasConcept C192562407 @default.
- W2003840611 hasConcept C2778119891 @default.
- W2003840611 hasConcept C2780551164 @default.
- W2003840611 hasConcept C2780595030 @default.
- W2003840611 hasConcept C33923547 @default.
- W2003840611 hasConcept C41008148 @default.
- W2003840611 hasConcept C45374587 @default.
- W2003840611 hasConcept C56372850 @default.
- W2003840611 hasConcept C60798267 @default.
- W2003840611 hasConcept C62520636 @default.
- W2003840611 hasConcept C68339613 @default.
- W2003840611 hasConcept C76155785 @default.
- W2003840611 hasConcept C83283714 @default.
- W2003840611 hasConcept C84114770 @default.
- W2003840611 hasConcept C94375191 @default.
- W2003840611 hasConceptScore W2003840611C106487976 @default.
- W2003840611 hasConceptScore W2003840611C111919701 @default.
- W2003840611 hasConceptScore W2003840611C11413529 @default.
- W2003840611 hasConceptScore W2003840611C114614502 @default.
- W2003840611 hasConceptScore W2003840611C121332964 @default.
- W2003840611 hasConceptScore W2003840611C126042441 @default.
- W2003840611 hasConceptScore W2003840611C138101251 @default.
- W2003840611 hasConceptScore W2003840611C150552126 @default.
- W2003840611 hasConceptScore W2003840611C159985019 @default.
- W2003840611 hasConceptScore W2003840611C163716315 @default.
- W2003840611 hasConceptScore W2003840611C17349429 @default.
- W2003840611 hasConceptScore W2003840611C173608175 @default.
- W2003840611 hasConceptScore W2003840611C192562407 @default.
- W2003840611 hasConceptScore W2003840611C2778119891 @default.
- W2003840611 hasConceptScore W2003840611C2780551164 @default.
- W2003840611 hasConceptScore W2003840611C2780595030 @default.
- W2003840611 hasConceptScore W2003840611C33923547 @default.
- W2003840611 hasConceptScore W2003840611C41008148 @default.
- W2003840611 hasConceptScore W2003840611C45374587 @default.
- W2003840611 hasConceptScore W2003840611C56372850 @default.
- W2003840611 hasConceptScore W2003840611C60798267 @default.
- W2003840611 hasConceptScore W2003840611C62520636 @default.
- W2003840611 hasConceptScore W2003840611C68339613 @default.
- W2003840611 hasConceptScore W2003840611C76155785 @default.