Matches in SemOpenAlex for { <https://semopenalex.org/work/W4384705419> ?p ?o ?g. }
Showing items 1 to 88 of
88
with 100 items per page.
- W4384705419 abstract "Graph Neural Networks (GNNs) have shown success in many real-world applications that involve graph-structured data. Most of the existing single-node GNN training systems are capable of training medium-scale graphs with tens of millions of edges; however, scaling them to large-scale graphs with billions of edges remains challenging. In addition, it is challenging to map GNN training algorithms onto a computation node as state-of-the-art machines feature heterogeneous architecture consisting of multiple processors and a variety of accelerators.We propose HyScale-GNN, a novel system to train GNN models on a single-node heterogeneous architecture. HyScale-GNN performs hybrid training which utilizes both the processors and the accelerators to train a model collaboratively. Our system design overcomes the memory size limitation of existing works and is optimized for training GNNs on large-scale graphs. We propose a two-stage data pre-fetching scheme to reduce the communication overhead during GNN training. To improve task mapping efficiency, we propose a dynamic resource management mechanism, which adjusts the workload assignment and resource allocation during runtime. We evaluate HyScale-GNN on a CPU-GPU and a CPU-FPGA heterogeneous architecture. Using several large-scale datasets and two widely-used GNN models, we compare the performance of our design with a multi-GPU baseline implemented in PyTorch-Geometric. The CPU-GPU design and the CPU-FPGA design achieve up to 2.08× speedup and 12.6× speedup, respectively. Compared with the state-of-the-art large-scale multi-node GNN training systems such as P <sup xmlns:mml=http://www.w3.org/1998/Math/MathML xmlns:xlink=http://www.w3.org/1999/xlink>3</sup> and DistDGL, our CPU-FPGA design achieves up to 5.27× speedup using a single node." @default.
- W4384705419 created "2023-07-20" @default.
- W4384705419 creator A5033166029 @default.
- W4384705419 creator A5081791851 @default.
- W4384705419 date "2023-05-01" @default.
- W4384705419 modified "2023-09-26" @default.
- W4384705419 title "HyScale-GNN: A Scalable Hybrid GNN Training System on Single-Node Heterogeneous Architecture" @default.
- W4384705419 cites W2606722458 @default.
- W4384705419 cites W2886438137 @default.
- W4384705419 cites W2970929262 @default.
- W4384705419 cites W3008905965 @default.
- W4384705419 cites W3096566397 @default.
- W4384705419 cites W3100848837 @default.
- W4384705419 cites W3106125969 @default.
- W4384705419 cites W3132185085 @default.
- W4384705419 cites W3159109662 @default.
- W4384705419 cites W3159953606 @default.
- W4384705419 cites W3200354549 @default.
- W4384705419 cites W4200504397 @default.
- W4384705419 cites W4211165432 @default.
- W4384705419 cites W4221155008 @default.
- W4384705419 cites W4281708879 @default.
- W4384705419 cites W4290944486 @default.
- W4384705419 cites W4293025067 @default.
- W4384705419 cites W4312325851 @default.
- W4384705419 cites W4319074048 @default.
- W4384705419 cites W4321637087 @default.
- W4384705419 doi "https://doi.org/10.1109/ipdps54959.2023.00062" @default.
- W4384705419 hasPublicationYear "2023" @default.
- W4384705419 type Work @default.
- W4384705419 citedByCount "0" @default.
- W4384705419 crossrefType "proceedings-article" @default.
- W4384705419 hasAuthorship W4384705419A5033166029 @default.
- W4384705419 hasAuthorship W4384705419A5081791851 @default.
- W4384705419 hasConcept C111919701 @default.
- W4384705419 hasConcept C11413529 @default.
- W4384705419 hasConcept C118524514 @default.
- W4384705419 hasConcept C120314980 @default.
- W4384705419 hasConcept C127413603 @default.
- W4384705419 hasConcept C132525143 @default.
- W4384705419 hasConcept C149635348 @default.
- W4384705419 hasConcept C173608175 @default.
- W4384705419 hasConcept C2778476105 @default.
- W4384705419 hasConcept C2779960059 @default.
- W4384705419 hasConcept C41008148 @default.
- W4384705419 hasConcept C42935608 @default.
- W4384705419 hasConcept C45374587 @default.
- W4384705419 hasConcept C48044578 @default.
- W4384705419 hasConcept C62611344 @default.
- W4384705419 hasConcept C66938386 @default.
- W4384705419 hasConcept C68339613 @default.
- W4384705419 hasConcept C77088390 @default.
- W4384705419 hasConcept C80444323 @default.
- W4384705419 hasConceptScore W4384705419C111919701 @default.
- W4384705419 hasConceptScore W4384705419C11413529 @default.
- W4384705419 hasConceptScore W4384705419C118524514 @default.
- W4384705419 hasConceptScore W4384705419C120314980 @default.
- W4384705419 hasConceptScore W4384705419C127413603 @default.
- W4384705419 hasConceptScore W4384705419C132525143 @default.
- W4384705419 hasConceptScore W4384705419C149635348 @default.
- W4384705419 hasConceptScore W4384705419C173608175 @default.
- W4384705419 hasConceptScore W4384705419C2778476105 @default.
- W4384705419 hasConceptScore W4384705419C2779960059 @default.
- W4384705419 hasConceptScore W4384705419C41008148 @default.
- W4384705419 hasConceptScore W4384705419C42935608 @default.
- W4384705419 hasConceptScore W4384705419C45374587 @default.
- W4384705419 hasConceptScore W4384705419C48044578 @default.
- W4384705419 hasConceptScore W4384705419C62611344 @default.
- W4384705419 hasConceptScore W4384705419C66938386 @default.
- W4384705419 hasConceptScore W4384705419C68339613 @default.
- W4384705419 hasConceptScore W4384705419C77088390 @default.
- W4384705419 hasConceptScore W4384705419C80444323 @default.
- W4384705419 hasLocation W43847054191 @default.
- W4384705419 hasOpenAccess W4384705419 @default.
- W4384705419 hasPrimaryLocation W43847054191 @default.
- W4384705419 hasRelatedWork W1582107257 @default.
- W4384705419 hasRelatedWork W1595151633 @default.
- W4384705419 hasRelatedWork W2021313767 @default.
- W4384705419 hasRelatedWork W2040758634 @default.
- W4384705419 hasRelatedWork W2064105003 @default.
- W4384705419 hasRelatedWork W2086205578 @default.
- W4384705419 hasRelatedWork W2384714939 @default.
- W4384705419 hasRelatedWork W2470589840 @default.
- W4384705419 hasRelatedWork W94000989 @default.
- W4384705419 hasRelatedWork W2516240846 @default.
- W4384705419 isParatext "false" @default.
- W4384705419 isRetracted "false" @default.
- W4384705419 workType "article" @default.