FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3016169946> ?p ?o ?g. }

• W3016169946 abstract "Conventional neural accelerators rely on isolated self-sufficient functional units that perform an atomic operation while communicating the results through an operand delivery-aggregation logic. Each single unit processes all the bits of their operands atomically and produce all the bits of the results in isolation. This paper explores a different design style, where each unit is only responsible for a slice of the bit-level operations to interleave and combine the benefits of bit-level parallelism with the abundant data-level parallelism in deep neural networks. A dynamic collection of these units cooperate at runtime to generate bits of the results, collectively. Such cooperation requires extracting new grouping between the bits, which is only possible if the operands and operations are vectorizable. The abundance of Data Level Parallelism and mostly repeated execution patterns, provides a unique opportunity to define and leverage this new dimension of Bit-Parallel Vector Composability. This design intersperses bit parallelism within data-level parallelism and dynamically interweaves the two together. As such, the building block of our neural accelerator is a Composable Vector Unit that is a collection of Narrower-Bitwidth Vector Engines, which are dynamically composed or decomposed at the bit granularity. Using six diverse CNN and LSTM deep networks, we evaluate this design style across four design points: with and without algorithmic bitwidth heterogeneity and with and without availability of a high-bandwidth off-chip memory. Across these four design points, Bit-Parallel Vector Composability brings (1.4x to 3.5x) speedup and (1.1x to 2.7x) energy reduction. We also comprehensively compare our design style to the Nvidia RTX 2080 TI GPU, which also supports INT-4 execution. The benefits range between 28.0x and 33.7x improvement in Performance-per-Watt." @default.
• W3016169946 created "2020-04-17" @default.
• W3016169946 creator A5035427716 @default.
• W3016169946 creator A5037648751 @default.
• W3016169946 creator A5082318034 @default.
• W3016169946 creator A5082499242 @default.
• W3016169946 creator A5084514143 @default.
• W3016169946 date "2020-04-11" @default.
• W3016169946 modified "2023-09-23" @default.
• W3016169946 title "Bit-Parallel Vector Composability for Neural Acceleration" @default.
• W3016169946 cites W2133834148 @default.
• W3016169946 cites W2285660444 @default.
• W3016169946 cites W2442974303 @default.
• W3016169946 cites W2508602506 @default.
• W3016169946 cites W2516141709 @default.
• W3016169946 cites W2518281301 @default.
• W3016169946 cites W2524428287 @default.
• W3016169946 cites W2563587242 @default.
• W3016169946 cites W2606722458 @default.
• W3016169946 cites W2625457103 @default.
• W3016169946 cites W2719597717 @default.
• W3016169946 cites W2751477244 @default.
• W3016169946 cites W2761132374 @default.
• W3016169946 cites W2786771851 @default.
• W3016169946 cites W2794141774 @default.
• W3016169946 cites W2909365574 @default.
• W3016169946 cites W2949870694 @default.
• W3016169946 cites W2963367920 @default.
• W3016169946 hasPublicationYear "2020" @default.
• W3016169946 type Work @default.
• W3016169946 sameAs 3016169946 @default.
• W3016169946 citedByCount "0" @default.
• W3016169946 crossrefType "posted-content" @default.
• W3016169946 hasAuthorship W3016169946A5035427716 @default.
• W3016169946 hasAuthorship W3016169946A5037648751 @default.
• W3016169946 hasAuthorship W3016169946A5082318034 @default.
• W3016169946 hasAuthorship W3016169946A5082499242 @default.
• W3016169946 hasAuthorship W3016169946A5084514143 @default.
• W3016169946 hasConcept C120314980 @default.
• W3016169946 hasConcept C154945302 @default.
• W3016169946 hasConcept C173608175 @default.
• W3016169946 hasConcept C2778814252 @default.
• W3016169946 hasConcept C2781172179 @default.
• W3016169946 hasConcept C41008148 @default.
• W3016169946 hasConcept C50644808 @default.
• W3016169946 hasConcept C55526617 @default.
• W3016169946 hasConcept C61483411 @default.
• W3016169946 hasConcept C68339613 @default.
• W3016169946 hasConcept C9390403 @default.
• W3016169946 hasConceptScore W3016169946C120314980 @default.
• W3016169946 hasConceptScore W3016169946C154945302 @default.
• W3016169946 hasConceptScore W3016169946C173608175 @default.
• W3016169946 hasConceptScore W3016169946C2778814252 @default.
• W3016169946 hasConceptScore W3016169946C2781172179 @default.
• W3016169946 hasConceptScore W3016169946C41008148 @default.
• W3016169946 hasConceptScore W3016169946C50644808 @default.
• W3016169946 hasConceptScore W3016169946C55526617 @default.
• W3016169946 hasConceptScore W3016169946C61483411 @default.
• W3016169946 hasConceptScore W3016169946C68339613 @default.
• W3016169946 hasConceptScore W3016169946C9390403 @default.
• W3016169946 hasOpenAccess W3016169946 @default.
• W3016169946 hasRelatedWork W1505735619 @default.
• W3016169946 hasRelatedWork W1531251462 @default.
• W3016169946 hasRelatedWork W1536960396 @default.
• W3016169946 hasRelatedWork W1974531365 @default.
• W3016169946 hasRelatedWork W2015443653 @default.
• W3016169946 hasRelatedWork W2049142403 @default.
• W3016169946 hasRelatedWork W2077569042 @default.
• W3016169946 hasRelatedWork W2135717971 @default.
• W3016169946 hasRelatedWork W2155772026 @default.
• W3016169946 hasRelatedWork W2262429236 @default.
• W3016169946 hasRelatedWork W2298144119 @default.
• W3016169946 hasRelatedWork W2359058895 @default.
• W3016169946 hasRelatedWork W2472255304 @default.
• W3016169946 hasRelatedWork W2581189719 @default.
• W3016169946 hasRelatedWork W2775543260 @default.
• W3016169946 hasRelatedWork W2788825194 @default.
• W3016169946 hasRelatedWork W2884836535 @default.
• W3016169946 hasRelatedWork W2896415993 @default.
• W3016169946 hasRelatedWork W3091813304 @default.
• W3016169946 hasRelatedWork W3171483929 @default.
• W3016169946 isParatext "false" @default.
• W3016169946 isRetracted "false" @default.
• W3016169946 magId "3016169946" @default.
• W3016169946 workType "article" @default.