FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3207203625> ?p ?o ?g. }

• W3207203625 abstract "Graph neural networks (GNNs) have been emerging as powerful learning tools for unstructured data and successfully applied to many graph-based application domains. Sampling-based graph training is commonly used in existing GNN training frameworks to handle large-scale graphs. However, this type of approach is restricted by the problems of long memory accessing latency, neighborhood explosion during mini-batch sampling, and inefficient loading of vertex features from CPU to GPU. In this paper, we propose 2PGraph, a system that supports high-speed locality-aware mini-batch sampling and GNN layer-aware feature caching. 2PGraph significantly reduces sampling time by vertex-cluster sampling, which improves the locality of vertex access and limits the range of neighborhood expansion. To further reduce the sampling time in a distributed environment, we renumber the vertex numbers in subgraphs after graph partition, which improves the data locality of each partition. 2PGraph also avoids abundant data transfer between CPU and GPU through the feature data caching on available GPU resources with a hit rate of 100%. Furthermore, 2PGraph develops a GNN layer-aware feature caching policy during data parallel training and achieves better cache efficiency and memory utilization. We evaluate 2PGraph against two state-of-the-art industrial GNN frameworks, i.e., PyG and DGL, on a diverse array of benchmarks. Experimental results show that 2PGraph reduces up to 90% mini-batch sampling and 99% data loading time, and achieves up to 8.7 × performance speedup over the state-of-the-art baselines on an 8-GPU cluster." @default.
• W3207203625 created "2021-10-25" @default.
• W3207203625 creator A5031722471 @default.
• W3207203625 creator A5033969180 @default.
• W3207203625 creator A5041959417 @default.
• W3207203625 creator A5055426388 @default.
• W3207203625 creator A5060470951 @default.
• W3207203625 creator A5076345724 @default.
• W3207203625 date "2021-09-01" @default.
• W3207203625 modified "2023-09-26" @default.
• W3207203625 title "2PGraph: Accelerating GNN Training over Large Graphs on GPU Clusters" @default.
• W3207203625 cites W1514171102 @default.
• W3207203625 cites W1991418309 @default.
• W3207203625 cites W1994727615 @default.
• W3207203625 cites W2011963492 @default.
• W3207203625 cites W2070232376 @default.
• W3207203625 cites W2096029105 @default.
• W3207203625 cites W2116341502 @default.
• W3207203625 cites W2135957668 @default.
• W3207203625 cites W2558748708 @default.
• W3207203625 cites W2584785188 @default.
• W3207203625 cites W2767104566 @default.
• W3207203625 cites W2898565872 @default.
• W3207203625 cites W2945827377 @default.
• W3207203625 cites W2963406064 @default.
• W3207203625 cites W2963601856 @default.
• W3207203625 cites W2964571482 @default.
• W3207203625 cites W2982536084 @default.
• W3207203625 cites W3002924435 @default.
• W3207203625 cites W3011667710 @default.
• W3207203625 cites W3037471945 @default.
• W3207203625 cites W3042370959 @default.
• W3207203625 cites W3086238199 @default.
• W3207203625 cites W3096566397 @default.
• W3207203625 cites W3152893301 @default.
• W3207203625 cites W3159953606 @default.
• W3207203625 cites W4210257598 @default.
• W3207203625 cites W4297971002 @default.
• W3207203625 doi "https://doi.org/10.1109/cluster48925.2021.00036" @default.
• W3207203625 hasPublicationYear "2021" @default.
• W3207203625 type Work @default.
• W3207203625 sameAs 3207203625 @default.
• W3207203625 citedByCount "3" @default.
• W3207203625 countsByYear W32072036252022 @default.
• W3207203625 countsByYear W32072036252023 @default.
• W3207203625 crossrefType "proceedings-article" @default.
• W3207203625 hasAuthorship W3207203625A5031722471 @default.
• W3207203625 hasAuthorship W3207203625A5033969180 @default.
• W3207203625 hasAuthorship W3207203625A5041959417 @default.
• W3207203625 hasAuthorship W3207203625A5055426388 @default.
• W3207203625 hasAuthorship W3207203625A5060470951 @default.
• W3207203625 hasAuthorship W3207203625A5076345724 @default.
• W3207203625 hasConcept C114614502 @default.
• W3207203625 hasConcept C115537543 @default.
• W3207203625 hasConcept C132525143 @default.
• W3207203625 hasConcept C138885662 @default.
• W3207203625 hasConcept C173608175 @default.
• W3207203625 hasConcept C2778119891 @default.
• W3207203625 hasConcept C2779808786 @default.
• W3207203625 hasConcept C2781335571 @default.
• W3207203625 hasConcept C33923547 @default.
• W3207203625 hasConcept C41008148 @default.
• W3207203625 hasConcept C41895202 @default.
• W3207203625 hasConcept C42812 @default.
• W3207203625 hasConcept C48903430 @default.
• W3207203625 hasConcept C68339613 @default.
• W3207203625 hasConcept C80444323 @default.
• W3207203625 hasConcept C80899671 @default.
• W3207203625 hasConceptScore W3207203625C114614502 @default.
• W3207203625 hasConceptScore W3207203625C115537543 @default.
• W3207203625 hasConceptScore W3207203625C132525143 @default.
• W3207203625 hasConceptScore W3207203625C138885662 @default.
• W3207203625 hasConceptScore W3207203625C173608175 @default.
• W3207203625 hasConceptScore W3207203625C2778119891 @default.
• W3207203625 hasConceptScore W3207203625C2779808786 @default.
• W3207203625 hasConceptScore W3207203625C2781335571 @default.
• W3207203625 hasConceptScore W3207203625C33923547 @default.
• W3207203625 hasConceptScore W3207203625C41008148 @default.
• W3207203625 hasConceptScore W3207203625C41895202 @default.
• W3207203625 hasConceptScore W3207203625C42812 @default.
• W3207203625 hasConceptScore W3207203625C48903430 @default.
• W3207203625 hasConceptScore W3207203625C68339613 @default.
• W3207203625 hasConceptScore W3207203625C80444323 @default.
• W3207203625 hasConceptScore W3207203625C80899671 @default.
• W3207203625 hasFunder F4320321001 @default.
• W3207203625 hasLocation W32072036251 @default.
• W3207203625 hasOpenAccess W3207203625 @default.
• W3207203625 hasPrimaryLocation W32072036251 @default.
• W3207203625 hasRelatedWork W108745714 @default.
• W3207203625 hasRelatedWork W1498943610 @default.
• W3207203625 hasRelatedWork W1509211761 @default.
• W3207203625 hasRelatedWork W1968003516 @default.
• W3207203625 hasRelatedWork W2082176405 @default.
• W3207203625 hasRelatedWork W2238531340 @default.
• W3207203625 hasRelatedWork W2268138364 @default.
• W3207203625 hasRelatedWork W2582456645 @default.
• W3207203625 hasRelatedWork W3207203625 @default.
• W3207203625 hasRelatedWork W4294734199 @default.
• W3207203625 isParatext "false" @default.
• W3207203625 isRetracted "false" @default.

• next