Matches in SemOpenAlex for { <https://semopenalex.org/work/W4313496554> ?p ?o ?g. }
Showing items 1 to 76 of
76
with 100 items per page.
- W4313496554 abstract "Convolutional Neural Networks (CNNs) inference on Heterogeneous Multi-Processor System-on-Chips (HMPSoCs) in edge devices represent cutting-edge embedded machine learning. Embedded CPU and GPU within an HMPSoC can both perform inference using CNNs. However, common practice is to run a CNN on the HMPSoC component (CPU or GPU) provides the best performance (lowest latency) for that CNN. CNNs are not monolithic and are composed of several layers of different types. Some of these layers have lower latency on the CPU, while others execute faster on the GPU. In this work, we investigate the reason behind this observation. We also propose an execution of CNN that switches between CPU and GPU at the layer granularity, wherein a CNN layer executes on the component that provides it with the lowest latency. Switching between the CPU and the GPU back and forth mid-inference introduces additional overhead (delay) in the inference. Regardless of overhead, we show in this work that a CPU-GPU layer switched execution results in, on average, having 4.72% lower CNN inference latency on the Khadas VIM 3 board with Amlogic A311D HMPSoC." @default.
- W4313496554 created "2023-01-06" @default.
- W4313496554 creator A5004356661 @default.
- W4313496554 creator A5023204068 @default.
- W4313496554 creator A5040204792 @default.
- W4313496554 creator A5067055700 @default.
- W4313496554 date "2022-08-01" @default.
- W4313496554 modified "2023-10-16" @default.
- W4313496554 title "CPU-GPU Layer-Switched Low Latency CNN Inference" @default.
- W4313496554 cites W2557406251 @default.
- W4313496554 cites W2626129225 @default.
- W4313496554 cites W2792243241 @default.
- W4313496554 cites W2922395136 @default.
- W4313496554 cites W2931092525 @default.
- W4313496554 cites W2931743911 @default.
- W4313496554 cites W2963163009 @default.
- W4313496554 cites W2984200518 @default.
- W4313496554 cites W3010466514 @default.
- W4313496554 cites W3034544214 @default.
- W4313496554 cites W3090924437 @default.
- W4313496554 cites W3095320840 @default.
- W4313496554 cites W3104393472 @default.
- W4313496554 cites W3105888187 @default.
- W4313496554 cites W3215253865 @default.
- W4313496554 cites W4236853429 @default.
- W4313496554 cites W4244330903 @default.
- W4313496554 doi "https://doi.org/10.1109/dsd57027.2022.00051" @default.
- W4313496554 hasPublicationYear "2022" @default.
- W4313496554 type Work @default.
- W4313496554 citedByCount "3" @default.
- W4313496554 countsByYear W43134965542023 @default.
- W4313496554 crossrefType "proceedings-article" @default.
- W4313496554 hasAuthorship W4313496554A5004356661 @default.
- W4313496554 hasAuthorship W4313496554A5023204068 @default.
- W4313496554 hasAuthorship W4313496554A5040204792 @default.
- W4313496554 hasAuthorship W4313496554A5067055700 @default.
- W4313496554 hasConcept C111919701 @default.
- W4313496554 hasConcept C154945302 @default.
- W4313496554 hasConcept C173608175 @default.
- W4313496554 hasConcept C177774035 @default.
- W4313496554 hasConcept C180613757 @default.
- W4313496554 hasConcept C2776214188 @default.
- W4313496554 hasConcept C41008148 @default.
- W4313496554 hasConcept C49154492 @default.
- W4313496554 hasConcept C76155785 @default.
- W4313496554 hasConcept C81363708 @default.
- W4313496554 hasConcept C82876162 @default.
- W4313496554 hasConcept C9390403 @default.
- W4313496554 hasConceptScore W4313496554C111919701 @default.
- W4313496554 hasConceptScore W4313496554C154945302 @default.
- W4313496554 hasConceptScore W4313496554C173608175 @default.
- W4313496554 hasConceptScore W4313496554C177774035 @default.
- W4313496554 hasConceptScore W4313496554C180613757 @default.
- W4313496554 hasConceptScore W4313496554C2776214188 @default.
- W4313496554 hasConceptScore W4313496554C41008148 @default.
- W4313496554 hasConceptScore W4313496554C49154492 @default.
- W4313496554 hasConceptScore W4313496554C76155785 @default.
- W4313496554 hasConceptScore W4313496554C81363708 @default.
- W4313496554 hasConceptScore W4313496554C82876162 @default.
- W4313496554 hasConceptScore W4313496554C9390403 @default.
- W4313496554 hasLocation W43134965541 @default.
- W4313496554 hasOpenAccess W4313496554 @default.
- W4313496554 hasPrimaryLocation W43134965541 @default.
- W4313496554 hasRelatedWork W1847937745 @default.
- W4313496554 hasRelatedWork W2058541071 @default.
- W4313496554 hasRelatedWork W206617434 @default.
- W4313496554 hasRelatedWork W2092389159 @default.
- W4313496554 hasRelatedWork W2092999126 @default.
- W4313496554 hasRelatedWork W2145974411 @default.
- W4313496554 hasRelatedWork W2551154568 @default.
- W4313496554 hasRelatedWork W2554083593 @default.
- W4313496554 hasRelatedWork W2955891667 @default.
- W4313496554 hasRelatedWork W2998804831 @default.
- W4313496554 isParatext "false" @default.
- W4313496554 isRetracted "false" @default.
- W4313496554 workType "article" @default.