Matches in SemOpenAlex for { <https://semopenalex.org/work/W2904906706> ?p ?o ?g. }
Showing items 1 to 85 of
85
with 100 items per page.
- W2904906706 abstract "As graph applications become more popular and diverse, it is important to design efficient hardware architectures that maintain the flexibility of high-level graph programming frameworks. Prior works have identified the memory subsystem of traditional chip multiprocessors (CMPs) as a key source of inefficiency; however, these solutions do not exploit locality that exists in the structure of many real-world graphs. In this work, we target graphs that follow a power-law distribution, for which there is a unique opportunity to significantly boost the overall performance of the memory subsystem. We note that many natural graphs, derived from web, social networks, even biological networks, follow the power law, that is, 20% of the vertices are linked to 80% of the edges. Based on this observation, we propose a novel memory subsystem architecture that leverages this structural graph locality. Our architecture is based on a heterogeneous cache/scratchpad memory subsystem and a lightweight compute engine, to which the cores can offload atomic graph operations. Our architecture provides 2x speedup, on average, over a same-sized baseline CMP running the Ligra framework, while requiring no modification to the Ligra programming interface." @default.
- W2904906706 created "2018-12-22" @default.
- W2904906706 creator A5009899727 @default.
- W2904906706 creator A5019722520 @default.
- W2904906706 creator A5030335506 @default.
- W2904906706 creator A5040423477 @default.
- W2904906706 date "2018-09-01" @default.
- W2904906706 modified "2023-10-16" @default.
- W2904906706 title "Heterogeneous Memory Subsystem for Natural Graph Analytics" @default.
- W2904906706 cites W1482680420 @default.
- W2904906706 cites W1512159850 @default.
- W2904906706 cites W1783256592 @default.
- W2904906706 cites W1955594754 @default.
- W2904906706 cites W1975274780 @default.
- W2904906706 cites W2008620264 @default.
- W2904906706 cites W2013062050 @default.
- W2904906706 cites W2021425055 @default.
- W2904906706 cites W2044770804 @default.
- W2904906706 cites W2053076698 @default.
- W2904906706 cites W2060778826 @default.
- W2904906706 cites W2080958936 @default.
- W2904906706 cites W2086112773 @default.
- W2904906706 cites W2147657366 @default.
- W2904906706 cites W2166111585 @default.
- W2904906706 cites W2170382128 @default.
- W2904906706 cites W2170616854 @default.
- W2904906706 cites W2296407087 @default.
- W2904906706 cites W2529597500 @default.
- W2904906706 cites W2606413522 @default.
- W2904906706 cites W2612654866 @default.
- W2904906706 cites W2731857860 @default.
- W2904906706 cites W2766578007 @default.
- W2904906706 cites W2769133055 @default.
- W2904906706 cites W2790789009 @default.
- W2904906706 cites W2950627632 @default.
- W2904906706 cites W2962865652 @default.
- W2904906706 cites W2997134027 @default.
- W2904906706 cites W3023843293 @default.
- W2904906706 cites W3145162387 @default.
- W2904906706 cites W4234988573 @default.
- W2904906706 cites W4240241667 @default.
- W2904906706 cites W4241140669 @default.
- W2904906706 cites W4252563462 @default.
- W2904906706 doi "https://doi.org/10.1109/iiswc.2018.8573480" @default.
- W2904906706 hasPublicationYear "2018" @default.
- W2904906706 type Work @default.
- W2904906706 sameAs 2904906706 @default.
- W2904906706 citedByCount "24" @default.
- W2904906706 countsByYear W29049067062019 @default.
- W2904906706 countsByYear W29049067062020 @default.
- W2904906706 countsByYear W29049067062021 @default.
- W2904906706 countsByYear W29049067062022 @default.
- W2904906706 countsByYear W29049067062023 @default.
- W2904906706 crossrefType "proceedings-article" @default.
- W2904906706 hasAuthorship W2904906706A5009899727 @default.
- W2904906706 hasAuthorship W2904906706A5019722520 @default.
- W2904906706 hasAuthorship W2904906706A5030335506 @default.
- W2904906706 hasAuthorship W2904906706A5040423477 @default.
- W2904906706 hasConcept C132525143 @default.
- W2904906706 hasConcept C2522767166 @default.
- W2904906706 hasConcept C41008148 @default.
- W2904906706 hasConcept C79158427 @default.
- W2904906706 hasConcept C80444323 @default.
- W2904906706 hasConceptScore W2904906706C132525143 @default.
- W2904906706 hasConceptScore W2904906706C2522767166 @default.
- W2904906706 hasConceptScore W2904906706C41008148 @default.
- W2904906706 hasConceptScore W2904906706C79158427 @default.
- W2904906706 hasConceptScore W2904906706C80444323 @default.
- W2904906706 hasLocation W29049067061 @default.
- W2904906706 hasOpenAccess W2904906706 @default.
- W2904906706 hasPrimaryLocation W29049067061 @default.
- W2904906706 hasRelatedWork W2049367484 @default.
- W2904906706 hasRelatedWork W2339404707 @default.
- W2904906706 hasRelatedWork W2460089464 @default.
- W2904906706 hasRelatedWork W2777139086 @default.
- W2904906706 hasRelatedWork W2897678396 @default.
- W2904906706 hasRelatedWork W4317655900 @default.
- W2904906706 hasRelatedWork W4320063224 @default.
- W2904906706 hasRelatedWork W4320880550 @default.
- W2904906706 hasRelatedWork W804204030 @default.
- W2904906706 hasRelatedWork W2551093110 @default.
- W2904906706 isParatext "false" @default.
- W2904906706 isRetracted "false" @default.
- W2904906706 magId "2904906706" @default.
- W2904906706 workType "article" @default.