Matches in SemOpenAlex for { <https://semopenalex.org/work/W4386859295> ?p ?o ?g. }
Showing items 1 to 68 of
68
with 100 items per page.
- W4386859295 abstract "Graph neural networks (GNN) are vital for analytics on real-world problems with graph models. This work develops a multicore GNN training accelerator and develops multicore-specific optimizations for superior performance. It uses enhanced multicore-specific dynamic caching to circumvent the costs of irregular DRAM access patterns of graph-structured data. A novel feature vector segmentation approach is used to maximize on-chip data reuse with high on-chip computation per memory access, reducing data access latency, using a machine learning model for optimal performance. The work presents a major advance over prior FPGA/ASIC GNN accelerators by handling significantly larger datasets (with up to 8.6M vertices) on a variety of GNN models. On average, training speedup of 17× and energy efficiency improvement of 322× is achieved over DGL on a GPU; a speedup of 14× with 268× lower energy is shown over GPU-based GNNAdvisor; and 11× and 24× speedups are obtained over ASIC-based Rubik and FPGA-based GraphACT." @default.
- W4386859295 created "2023-09-20" @default.
- W4386859295 creator A5002036272 @default.
- W4386859295 creator A5022067893 @default.
- W4386859295 creator A5063963932 @default.
- W4386859295 creator A5076919222 @default.
- W4386859295 creator A5090685279 @default.
- W4386859295 date "2023-08-07" @default.
- W4386859295 modified "2023-10-03" @default.
- W4386859295 title "A Multicore GNN Training Accelerator" @default.
- W4386859295 cites W1990505905 @default.
- W4386859295 cites W2070232376 @default.
- W4386859295 cites W2761132374 @default.
- W4386859295 cites W3017228913 @default.
- W4386859295 cites W3096566397 @default.
- W4386859295 cites W3106125969 @default.
- W4386859295 cites W3172512547 @default.
- W4386859295 cites W3213228850 @default.
- W4386859295 cites W3214431650 @default.
- W4386859295 cites W3214705685 @default.
- W4386859295 cites W4226137788 @default.
- W4386859295 cites W4293024973 @default.
- W4386859295 cites W4295934676 @default.
- W4386859295 cites W4312726386 @default.
- W4386859295 doi "https://doi.org/10.1109/islped58423.2023.10244283" @default.
- W4386859295 hasPublicationYear "2023" @default.
- W4386859295 type Work @default.
- W4386859295 citedByCount "0" @default.
- W4386859295 crossrefType "proceedings-article" @default.
- W4386859295 hasAuthorship W4386859295A5002036272 @default.
- W4386859295 hasAuthorship W4386859295A5022067893 @default.
- W4386859295 hasAuthorship W4386859295A5063963932 @default.
- W4386859295 hasAuthorship W4386859295A5076919222 @default.
- W4386859295 hasAuthorship W4386859295A5090685279 @default.
- W4386859295 hasConcept C118524514 @default.
- W4386859295 hasConcept C149635348 @default.
- W4386859295 hasConcept C173608175 @default.
- W4386859295 hasConcept C2780513914 @default.
- W4386859295 hasConcept C41008148 @default.
- W4386859295 hasConcept C42935608 @default.
- W4386859295 hasConcept C68339613 @default.
- W4386859295 hasConcept C77390884 @default.
- W4386859295 hasConcept C78766204 @default.
- W4386859295 hasConceptScore W4386859295C118524514 @default.
- W4386859295 hasConceptScore W4386859295C149635348 @default.
- W4386859295 hasConceptScore W4386859295C173608175 @default.
- W4386859295 hasConceptScore W4386859295C2780513914 @default.
- W4386859295 hasConceptScore W4386859295C41008148 @default.
- W4386859295 hasConceptScore W4386859295C42935608 @default.
- W4386859295 hasConceptScore W4386859295C68339613 @default.
- W4386859295 hasConceptScore W4386859295C77390884 @default.
- W4386859295 hasConceptScore W4386859295C78766204 @default.
- W4386859295 hasLocation W43868592951 @default.
- W4386859295 hasOpenAccess W4386859295 @default.
- W4386859295 hasPrimaryLocation W43868592951 @default.
- W4386859295 hasRelatedWork W1566136542 @default.
- W4386859295 hasRelatedWork W1993191611 @default.
- W4386859295 hasRelatedWork W2062495483 @default.
- W4386859295 hasRelatedWork W2066442567 @default.
- W4386859295 hasRelatedWork W2100470915 @default.
- W4386859295 hasRelatedWork W2365995233 @default.
- W4386859295 hasRelatedWork W2366027386 @default.
- W4386859295 hasRelatedWork W2540591044 @default.
- W4386859295 hasRelatedWork W3010492628 @default.
- W4386859295 hasRelatedWork W2506672464 @default.
- W4386859295 isParatext "false" @default.
- W4386859295 isRetracted "false" @default.
- W4386859295 workType "article" @default.