FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3125773586> ?p ?o ?g. }

• W3125773586 abstract "With the increasing adoption of graph neural networks (GNNs) in the machine learning community, GPUs have become an essential tool to accelerate GNN training. However, training GNNs on very large graphs that do not fit in GPU memory is still a challenging task. Unlike conventional neural networks, mini-batching input samples in GNNs requires complicated tasks such as traversing neighboring nodes and gathering their feature values. While this process accounts for a significant portion of the training time, we find existing GNN implementations using popular deep neural network (DNN) libraries such as PyTorch are limited to a CPU-centric approach for the entire data preparation step. This all-in-CPU approach has negative impact on the overall GNN training performance as it over-utilizes CPU resources and hinders GPU acceleration of GNN training. To overcome such limitations, we introduce PyTorch-Direct, which enables a GPU-centric data accessing paradigm for GNN training. In PyTorch-Direct, GPUs are capable of efficiently accessing complicated data structures in host memory directly without CPU intervention. Our microbenchmark and end-to-end GNN training results show that PyTorch-Direct reduces data transfer time by 47.1% on average and speeds up GNN training by up to 1.6x. Furthermore, by reducing CPU utilization, PyTorch-Direct also saves system power by 12.4% to 17.5% during training. To minimize programmer effort, we introduce a new unified type along with necessary changes to the PyTorch memory allocator, dispatch logic, and placement rules. As a result, users need to change at most two lines of their PyTorch GNN training code for each tensor object to take advantage of PyTorch-Direct." @default.
• W3125773586 created "2021-02-01" @default.
• W3125773586 creator A5024169920 @default.
• W3125773586 creator A5030156276 @default.
• W3125773586 creator A5037706821 @default.
• W3125773586 creator A5040404999 @default.
• W3125773586 creator A5050532440 @default.
• W3125773586 creator A5056321228 @default.
• W3125773586 creator A5073106039 @default.
• W3125773586 creator A5089789557 @default.
• W3125773586 date "2021-01-20" @default.
• W3125773586 modified "2023-09-27" @default.
• W3125773586 title "PyTorch-Direct: Enabling GPU Centric Data Access for Very Large Graph Neural Network Training with Irregular Accesses." @default.
• W3125773586 cites W131619556 @default.
• W3125773586 cites W1532854728 @default.
• W3125773586 cites W1994727615 @default.
• W3125773586 cites W2101196063 @default.
• W3125773586 cites W2194775991 @default.
• W3125773586 cites W2624431344 @default.
• W3125773586 cites W2905224888 @default.
• W3125773586 cites W2918342466 @default.
• W3125773586 cites W2961295589 @default.
• W3125773586 cites W2962767366 @default.
• W3125773586 cites W2963858333 @default.
• W3125773586 cites W2964015378 @default.
• W3125773586 cites W2970971581 @default.
• W3125773586 cites W3013163801 @default.
• W3125773586 cites W3021182071 @default.
• W3125773586 cites W3035128007 @default.
• W3125773586 cites W3036016037 @default.
• W3125773586 cites W3080555959 @default.
• W3125773586 cites W3084983693 @default.
• W3125773586 cites W3100848837 @default.
• W3125773586 cites W3021975806 @default.
• W3125773586 hasPublicationYear "2021" @default.
• W3125773586 type Work @default.
• W3125773586 sameAs 3125773586 @default.
• W3125773586 citedByCount "0" @default.
• W3125773586 crossrefType "posted-content" @default.
• W3125773586 hasAuthorship W3125773586A5024169920 @default.
• W3125773586 hasAuthorship W3125773586A5030156276 @default.
• W3125773586 hasAuthorship W3125773586A5037706821 @default.
• W3125773586 hasAuthorship W3125773586A5040404999 @default.
• W3125773586 hasAuthorship W3125773586A5050532440 @default.
• W3125773586 hasAuthorship W3125773586A5056321228 @default.
• W3125773586 hasAuthorship W3125773586A5073106039 @default.
• W3125773586 hasAuthorship W3125773586A5089789557 @default.
• W3125773586 hasConcept C108583219 @default.
• W3125773586 hasConcept C111919701 @default.
• W3125773586 hasConcept C113775141 @default.
• W3125773586 hasConcept C118524514 @default.
• W3125773586 hasConcept C132525143 @default.
• W3125773586 hasConcept C149635348 @default.
• W3125773586 hasConcept C154945302 @default.
• W3125773586 hasConcept C173608175 @default.
• W3125773586 hasConcept C199360897 @default.
• W3125773586 hasConcept C26713055 @default.
• W3125773586 hasConcept C2778514511 @default.
• W3125773586 hasConcept C41008148 @default.
• W3125773586 hasConcept C50644808 @default.
• W3125773586 hasConcept C80444323 @default.
• W3125773586 hasConcept C98045186 @default.
• W3125773586 hasConceptScore W3125773586C108583219 @default.
• W3125773586 hasConceptScore W3125773586C111919701 @default.
• W3125773586 hasConceptScore W3125773586C113775141 @default.
• W3125773586 hasConceptScore W3125773586C118524514 @default.
• W3125773586 hasConceptScore W3125773586C132525143 @default.
• W3125773586 hasConceptScore W3125773586C149635348 @default.
• W3125773586 hasConceptScore W3125773586C154945302 @default.
• W3125773586 hasConceptScore W3125773586C173608175 @default.
• W3125773586 hasConceptScore W3125773586C199360897 @default.
• W3125773586 hasConceptScore W3125773586C26713055 @default.
• W3125773586 hasConceptScore W3125773586C2778514511 @default.
• W3125773586 hasConceptScore W3125773586C41008148 @default.
• W3125773586 hasConceptScore W3125773586C50644808 @default.
• W3125773586 hasConceptScore W3125773586C80444323 @default.
• W3125773586 hasConceptScore W3125773586C98045186 @default.
• W3125773586 hasLocation W31257735861 @default.
• W3125773586 hasOpenAccess W3125773586 @default.
• W3125773586 hasPrimaryLocation W31257735861 @default.
• W3125773586 hasRelatedWork W2533980075 @default.
• W3125773586 hasRelatedWork W2798508334 @default.
• W3125773586 hasRelatedWork W2904283553 @default.
• W3125773586 hasRelatedWork W2907996334 @default.
• W3125773586 hasRelatedWork W2912509813 @default.
• W3125773586 hasRelatedWork W2938991217 @default.
• W3125773586 hasRelatedWork W2981237538 @default.
• W3125773586 hasRelatedWork W2989605642 @default.
• W3125773586 hasRelatedWork W2991040477 @default.
• W3125773586 hasRelatedWork W3021234081 @default.
• W3125773586 hasRelatedWork W3029054683 @default.
• W3125773586 hasRelatedWork W3037521362 @default.
• W3125773586 hasRelatedWork W3116115872 @default.
• W3125773586 hasRelatedWork W3130911940 @default.
• W3125773586 hasRelatedWork W3137648508 @default.
• W3125773586 hasRelatedWork W3176615255 @default.
• W3125773586 hasRelatedWork W3189583948 @default.
• W3125773586 hasRelatedWork W3195988899 @default.
• W3125773586 hasRelatedWork W3208040518 @default.
• W3125773586 hasRelatedWork W2740948200 @default.

• next