Matches in SemOpenAlex for { <https://semopenalex.org/work/W4384705370> ?p ?o ?g. }
Showing items 1 to 81 of
81
with 100 items per page.
- W4384705370 abstract "Graph Neural Network (GNN) inference is used in many real-world applications. Data sparsity in GNN inference, including sparsity in the input graph and the GNN model, offer opportunities to further speed up inference. Also, many pruning techniques have been proposed for model compression that increase the data sparsity of GNNs.We propose Dynasparse, a comprehensive hardware-software codesign on FPGA to accelerate GNN inference through dynamic sparsity exploitation. For this, we decouple the GNN computation kernels from the basic computation primitives, and explore hardware-software codesign as follows: 1) Hardware design: We propose a novel unified accelerator design on FPGA to efficiently execute various computation primitives. We develop a customized soft processor that is tightly coupled with the accelerator to execute a runtime system. Moreover, we develop efficient hardware mechanisms to profile the data sparsity and perform on-the-fly data format transformation to prepare the input data for various computation primitives; 2) Software design: We develop a runtime system that works synergistically with the accelerator to perform dynamic kernel-to-primitive mapping based on data sparsity. We implement Dynasparse on a state-of-the-art FPGA platform, Xilinx Alveo U250, and evaluate the design using widely used GNN models (GCN, GraphSAGE, GIN and SGC). For the above GNN models and various input graphs, the proposed accelerator and dynamic kernel-to-primitive mapping reduces the inference latency by 3.73× on the average compared with the static mapping strategies employed in the state-of-the-art GNN accelerators. Compared with state-of-the-art CPU (GPU) implementations, Dynasparse achieves up to 56.9× (2.37×) speedup in end-to-end latency. Compared with state-of-the-art FPGA implementations, Dynasparse achieves 2.7× speedup in accelerator execution latency." @default.
- W4384705370 created "2023-07-20" @default.
- W4384705370 creator A5033166029 @default.
- W4384705370 creator A5075652202 @default.
- W4384705370 date "2023-05-01" @default.
- W4384705370 modified "2023-10-17" @default.
- W4384705370 title "Dynasparse: Accelerating GNN Inference through Dynamic Sparsity Exploitation" @default.
- W4384705370 cites W2030081822 @default.
- W4384705370 cites W3011374900 @default.
- W4384705370 cites W3017228913 @default.
- W4384705370 cites W3046757167 @default.
- W4384705370 cites W3105753905 @default.
- W4384705370 cites W3111579839 @default.
- W4384705370 cites W3129894558 @default.
- W4384705370 cites W3172512547 @default.
- W4384705370 cites W3195824306 @default.
- W4384705370 cites W4211165432 @default.
- W4384705370 cites W4381327585 @default.
- W4384705370 doi "https://doi.org/10.1109/ipdps54959.2023.00032" @default.
- W4384705370 hasPublicationYear "2023" @default.
- W4384705370 type Work @default.
- W4384705370 citedByCount "0" @default.
- W4384705370 crossrefType "proceedings-article" @default.
- W4384705370 hasAuthorship W4384705370A5033166029 @default.
- W4384705370 hasAuthorship W4384705370A5075652202 @default.
- W4384705370 hasConcept C113775141 @default.
- W4384705370 hasConcept C11413529 @default.
- W4384705370 hasConcept C114614502 @default.
- W4384705370 hasConcept C13164978 @default.
- W4384705370 hasConcept C132525143 @default.
- W4384705370 hasConcept C149635348 @default.
- W4384705370 hasConcept C154945302 @default.
- W4384705370 hasConcept C173608175 @default.
- W4384705370 hasConcept C199360897 @default.
- W4384705370 hasConcept C2776214188 @default.
- W4384705370 hasConcept C2777904410 @default.
- W4384705370 hasConcept C33923547 @default.
- W4384705370 hasConcept C41008148 @default.
- W4384705370 hasConcept C42935608 @default.
- W4384705370 hasConcept C45374587 @default.
- W4384705370 hasConcept C68339613 @default.
- W4384705370 hasConcept C74193536 @default.
- W4384705370 hasConcept C80444323 @default.
- W4384705370 hasConcept C9390403 @default.
- W4384705370 hasConceptScore W4384705370C113775141 @default.
- W4384705370 hasConceptScore W4384705370C11413529 @default.
- W4384705370 hasConceptScore W4384705370C114614502 @default.
- W4384705370 hasConceptScore W4384705370C13164978 @default.
- W4384705370 hasConceptScore W4384705370C132525143 @default.
- W4384705370 hasConceptScore W4384705370C149635348 @default.
- W4384705370 hasConceptScore W4384705370C154945302 @default.
- W4384705370 hasConceptScore W4384705370C173608175 @default.
- W4384705370 hasConceptScore W4384705370C199360897 @default.
- W4384705370 hasConceptScore W4384705370C2776214188 @default.
- W4384705370 hasConceptScore W4384705370C2777904410 @default.
- W4384705370 hasConceptScore W4384705370C33923547 @default.
- W4384705370 hasConceptScore W4384705370C41008148 @default.
- W4384705370 hasConceptScore W4384705370C42935608 @default.
- W4384705370 hasConceptScore W4384705370C45374587 @default.
- W4384705370 hasConceptScore W4384705370C68339613 @default.
- W4384705370 hasConceptScore W4384705370C74193536 @default.
- W4384705370 hasConceptScore W4384705370C80444323 @default.
- W4384705370 hasConceptScore W4384705370C9390403 @default.
- W4384705370 hasFunder F4320306076 @default.
- W4384705370 hasFunder F4320314786 @default.
- W4384705370 hasLocation W43847053701 @default.
- W4384705370 hasOpenAccess W4384705370 @default.
- W4384705370 hasPrimaryLocation W43847053701 @default.
- W4384705370 hasRelatedWork W2115380918 @default.
- W4384705370 hasRelatedWork W2218038495 @default.
- W4384705370 hasRelatedWork W2518118925 @default.
- W4384705370 hasRelatedWork W2532502681 @default.
- W4384705370 hasRelatedWork W3035662153 @default.
- W4384705370 hasRelatedWork W3103034165 @default.
- W4384705370 hasRelatedWork W3154092384 @default.
- W4384705370 hasRelatedWork W3159273459 @default.
- W4384705370 hasRelatedWork W4280636456 @default.
- W4384705370 hasRelatedWork W4319952061 @default.
- W4384705370 isParatext "false" @default.
- W4384705370 isRetracted "false" @default.
- W4384705370 workType "article" @default.