FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3173818854> ?p ?o ?g. }

• W3173818854 abstract "Energy-efficiency is a key concern for neural network applications. To alleviate this issue, hardware acceleration using FPGAs or GPUs can provide better energy-efficiency than general-purpose processors. However, further improvement of the energy-efficiency of such accelerators will be extremely beneficial specially to deploy neural network in power-constrained edge computing environments. In this paper, we experimentally explore the potential of device-level energy-efficiency techniques (e.g.,supply voltage underscaling, frequency scaling, and data quantization) for representative off-the-shelf FPGAs compared to GPUs. Frequency scaling in both platforms can improve the power and energy consumption but with performance overhead, e.g.,in GPUs it improves the power consumption and GOPs/J by up to 34% and 28%, respectively. However, leveraging reduced-precision instructions improves power (up to 13%), energy (up to 20%), and performance (up to 7%) simultaneously, with negligible reduction in accuracy of neural network accuracy." @default.
• W3173818854 created "2021-07-05" @default.
• W3173818854 creator A5004601038 @default.
• W3173818854 creator A5026986531 @default.
• W3173818854 date "2021-06-26" @default.
• W3173818854 modified "2023-09-27" @default.
• W3173818854 title "On the Impact of Device-Level Techniques on Energy-Efficiency of Neural Network Accelerators." @default.
• W3173818854 cites W1965193075 @default.
• W3173818854 cites W1994172067 @default.
• W3173818854 cites W2098163907 @default.
• W3173818854 cites W2103498248 @default.
• W3173818854 cites W2117539524 @default.
• W3173818854 cites W2130967222 @default.
• W3173818854 cites W2133156997 @default.
• W3173818854 cites W2155893237 @default.
• W3173818854 cites W2311426409 @default.
• W3173818854 cites W2402144811 @default.
• W3173818854 cites W2513554817 @default.
• W3173818854 cites W2562773490 @default.
• W3173818854 cites W2591922920 @default.
• W3173818854 cites W2612994737 @default.
• W3173818854 cites W2624299682 @default.
• W3173818854 cites W2625457103 @default.
• W3173818854 cites W2766888159 @default.
• W3173818854 cites W2769038840 @default.
• W3173818854 cites W2790957308 @default.
• W3173818854 cites W2801453235 @default.
• W3173818854 cites W2904147273 @default.
• W3173818854 cites W2905500442 @default.
• W3173818854 cites W2921944720 @default.
• W3173818854 cites W2933569951 @default.
• W3173818854 cites W2942130737 @default.
• W3173818854 cites W2945148926 @default.
• W3173818854 cites W2945449142 @default.
• W3173818854 cites W2949981335 @default.
• W3173818854 cites W2963125010 @default.
• W3173818854 cites W2963367920 @default.
• W3173818854 cites W2963640628 @default.
• W3173818854 cites W2967343999 @default.
• W3173818854 cites W3006088222 @default.
• W3173818854 cites W3046769567 @default.
• W3173818854 cites W753012316 @default.
• W3173818854 cites W2116400939 @default.
• W3173818854 hasPublicationYear "2021" @default.
• W3173818854 type Work @default.
• W3173818854 sameAs 3173818854 @default.
• W3173818854 citedByCount "0" @default.
• W3173818854 crossrefType "posted-content" @default.
• W3173818854 hasAuthorship W3173818854A5004601038 @default.
• W3173818854 hasAuthorship W3173818854A5026986531 @default.
• W3173818854 hasConcept C105795698 @default.
• W3173818854 hasConcept C111335779 @default.
• W3173818854 hasConcept C111919701 @default.
• W3173818854 hasConcept C11413529 @default.
• W3173818854 hasConcept C118993495 @default.
• W3173818854 hasConcept C119599485 @default.
• W3173818854 hasConcept C121332964 @default.
• W3173818854 hasConcept C127413603 @default.
• W3173818854 hasConcept C138236772 @default.
• W3173818854 hasConcept C149635348 @default.
• W3173818854 hasConcept C154945302 @default.
• W3173818854 hasConcept C157742956 @default.
• W3173818854 hasConcept C163258240 @default.
• W3173818854 hasConcept C186370098 @default.
• W3173818854 hasConcept C2524010 @default.
• W3173818854 hasConcept C26517878 @default.
• W3173818854 hasConcept C2742236 @default.
• W3173818854 hasConcept C2779960059 @default.
• W3173818854 hasConcept C2780165032 @default.
• W3173818854 hasConcept C28855332 @default.
• W3173818854 hasConcept C33923547 @default.
• W3173818854 hasConcept C38652104 @default.
• W3173818854 hasConcept C41008148 @default.
• W3173818854 hasConcept C42935608 @default.
• W3173818854 hasConcept C50644808 @default.
• W3173818854 hasConcept C62520636 @default.
• W3173818854 hasConcept C79974875 @default.
• W3173818854 hasConcept C99844830 @default.
• W3173818854 hasConceptScore W3173818854C105795698 @default.
• W3173818854 hasConceptScore W3173818854C111335779 @default.
• W3173818854 hasConceptScore W3173818854C111919701 @default.
• W3173818854 hasConceptScore W3173818854C11413529 @default.
• W3173818854 hasConceptScore W3173818854C118993495 @default.
• W3173818854 hasConceptScore W3173818854C119599485 @default.
• W3173818854 hasConceptScore W3173818854C121332964 @default.
• W3173818854 hasConceptScore W3173818854C127413603 @default.
• W3173818854 hasConceptScore W3173818854C138236772 @default.
• W3173818854 hasConceptScore W3173818854C149635348 @default.
• W3173818854 hasConceptScore W3173818854C154945302 @default.
• W3173818854 hasConceptScore W3173818854C157742956 @default.
• W3173818854 hasConceptScore W3173818854C163258240 @default.
• W3173818854 hasConceptScore W3173818854C186370098 @default.
• W3173818854 hasConceptScore W3173818854C2524010 @default.
• W3173818854 hasConceptScore W3173818854C26517878 @default.
• W3173818854 hasConceptScore W3173818854C2742236 @default.
• W3173818854 hasConceptScore W3173818854C2779960059 @default.
• W3173818854 hasConceptScore W3173818854C2780165032 @default.
• W3173818854 hasConceptScore W3173818854C28855332 @default.
• W3173818854 hasConceptScore W3173818854C33923547 @default.
• W3173818854 hasConceptScore W3173818854C38652104 @default.

• next