Matches in SemOpenAlex for { <https://semopenalex.org/work/W4320509743> ?p ?o ?g. }
Showing items 1 to 85 of
85
with 100 items per page.
- W4320509743 abstract "Edge computing has been emerging as a popular scenario for model inference. However, the inference performance on edge devices (e.g., Multi-Core DSP, FGPA, etc.) suffers from inefficiency due to the lack of highly optimized inference frameworks. Previous model inference frameworks are mainly developed in an operator-centric way, which provides insufficient acceleration to edge-based inference. Besides, the operator-centric framework incurs significant costs for continuous development and maintenance. In this paper, we propose Xenos, which can automatically conduct dataflow-centric optimization of the computation graph and accelerate inference in two dimensions. Vertically, Xenos develops operator linking technique to improve data locality by restructuring the inter-operator dataflow. Horizontally, Xenos develops DSP-aware operator split technique to enable higher parallelism across multiple DSP units. Our evaluation proves the effectiveness of vertical and horizontal dataflow optimization, which reduce the inference time by 21.2%--84.9% and 17.9%--96.2% , respectively. Besides, Xenos also outperforms the widely-used TVM by 3.22$times$--17.92$times$. Moreover, we extend Xenos to a distributed solution, which we call d-Xenos. d-Xenos employs multiple edge devices to jointly conduct the inference task and achieves a speedup of 3.68x--3.78x compared with the single device." @default.
- W4320509743 created "2023-02-14" @default.
- W4320509743 creator A5000548162 @default.
- W4320509743 creator A5022526821 @default.
- W4320509743 creator A5035707673 @default.
- W4320509743 creator A5041238529 @default.
- W4320509743 creator A5048928060 @default.
- W4320509743 creator A5061453835 @default.
- W4320509743 creator A5066504632 @default.
- W4320509743 creator A5078941654 @default.
- W4320509743 creator A5080610727 @default.
- W4320509743 creator A5080819165 @default.
- W4320509743 date "2023-02-01" @default.
- W4320509743 modified "2023-10-18" @default.
- W4320509743 title "Xenos: Dataflow-Centric Optimization to Accelerate Model Inference on Edge Devices" @default.
- W4320509743 doi "https://doi.org/10.48550/arxiv.2302.00282" @default.
- W4320509743 hasPublicationYear "2023" @default.
- W4320509743 type Work @default.
- W4320509743 citedByCount "0" @default.
- W4320509743 crossrefType "posted-content" @default.
- W4320509743 hasAuthorship W4320509743A5000548162 @default.
- W4320509743 hasAuthorship W4320509743A5022526821 @default.
- W4320509743 hasAuthorship W4320509743A5035707673 @default.
- W4320509743 hasAuthorship W4320509743A5041238529 @default.
- W4320509743 hasAuthorship W4320509743A5048928060 @default.
- W4320509743 hasAuthorship W4320509743A5061453835 @default.
- W4320509743 hasAuthorship W4320509743A5066504632 @default.
- W4320509743 hasAuthorship W4320509743A5078941654 @default.
- W4320509743 hasAuthorship W4320509743A5080610727 @default.
- W4320509743 hasAuthorship W4320509743A5080819165 @default.
- W4320509743 hasBestOaLocation W43205097431 @default.
- W4320509743 hasConcept C104317684 @default.
- W4320509743 hasConcept C138885662 @default.
- W4320509743 hasConcept C154945302 @default.
- W4320509743 hasConcept C158448853 @default.
- W4320509743 hasConcept C162307627 @default.
- W4320509743 hasConcept C17020691 @default.
- W4320509743 hasConcept C173608175 @default.
- W4320509743 hasConcept C176727019 @default.
- W4320509743 hasConcept C185592680 @default.
- W4320509743 hasConcept C2776214188 @default.
- W4320509743 hasConcept C2778456923 @default.
- W4320509743 hasConcept C2779808786 @default.
- W4320509743 hasConcept C41008148 @default.
- W4320509743 hasConcept C41895202 @default.
- W4320509743 hasConcept C55493867 @default.
- W4320509743 hasConcept C68339613 @default.
- W4320509743 hasConcept C80444323 @default.
- W4320509743 hasConcept C86339819 @default.
- W4320509743 hasConcept C96324660 @default.
- W4320509743 hasConceptScore W4320509743C104317684 @default.
- W4320509743 hasConceptScore W4320509743C138885662 @default.
- W4320509743 hasConceptScore W4320509743C154945302 @default.
- W4320509743 hasConceptScore W4320509743C158448853 @default.
- W4320509743 hasConceptScore W4320509743C162307627 @default.
- W4320509743 hasConceptScore W4320509743C17020691 @default.
- W4320509743 hasConceptScore W4320509743C173608175 @default.
- W4320509743 hasConceptScore W4320509743C176727019 @default.
- W4320509743 hasConceptScore W4320509743C185592680 @default.
- W4320509743 hasConceptScore W4320509743C2776214188 @default.
- W4320509743 hasConceptScore W4320509743C2778456923 @default.
- W4320509743 hasConceptScore W4320509743C2779808786 @default.
- W4320509743 hasConceptScore W4320509743C41008148 @default.
- W4320509743 hasConceptScore W4320509743C41895202 @default.
- W4320509743 hasConceptScore W4320509743C55493867 @default.
- W4320509743 hasConceptScore W4320509743C68339613 @default.
- W4320509743 hasConceptScore W4320509743C80444323 @default.
- W4320509743 hasConceptScore W4320509743C86339819 @default.
- W4320509743 hasConceptScore W4320509743C96324660 @default.
- W4320509743 hasLocation W43205097431 @default.
- W4320509743 hasOpenAccess W4320509743 @default.
- W4320509743 hasPrimaryLocation W43205097431 @default.
- W4320509743 hasRelatedWork W1509211761 @default.
- W4320509743 hasRelatedWork W1819211029 @default.
- W4320509743 hasRelatedWork W2073056184 @default.
- W4320509743 hasRelatedWork W2171993104 @default.
- W4320509743 hasRelatedWork W2173304276 @default.
- W4320509743 hasRelatedWork W2582456645 @default.
- W4320509743 hasRelatedWork W2968111836 @default.
- W4320509743 hasRelatedWork W3092708771 @default.
- W4320509743 hasRelatedWork W4247683689 @default.
- W4320509743 hasRelatedWork W4254078218 @default.
- W4320509743 isParatext "false" @default.
- W4320509743 isRetracted "false" @default.
- W4320509743 workType "article" @default.