Matches in SemOpenAlex for { <https://semopenalex.org/work/W3100550724> ?p ?o ?g. }
- W3100550724 endingPage "2250" @default.
- W3100550724 startingPage "2232" @default.
- W3100550724 abstract "Advances in deep neural networks (DNNs) and the availability of massive real-world data have enabled superhuman levels of accuracy on many AI tasks and ushered the explosive growth of AI workloads across the spectrum of computing devices. However, their superior accuracy comes at a high computational cost, which necessitates approaches beyond traditional computing paradigms to improve their operational efficiency. Leveraging the application-level insight of error resilience, we demonstrate how approximate computing (AxC) can significantly boost the efficiency of AI platforms and play a pivotal role in the broader adoption of AI-based applications and services. To this end, we present RaPiD, a multi-tera operations per second (TOPS) AI hardware accelerator core (fabricated at 14-nm technology) that we built from the ground-up using AxC techniques across the stack including algorithms, architecture, programmability, and hardware. We highlight the workload-guided systematic explorations of AxC techniques for AI, including custom number representations, quantization/pruning methodologies, mixed-precision architecture design, instruction sets, and compiler technologies with quality programmability, employed in the RaPiD accelerator." @default.
- W3100550724 created "2020-11-23" @default.
- W3100550724 creator A5000627792 @default.
- W3100550724 creator A5001435691 @default.
- W3100550724 creator A5004766079 @default.
- W3100550724 creator A5006712350 @default.
- W3100550724 creator A5010094713 @default.
- W3100550724 creator A5011627573 @default.
- W3100550724 creator A5012357226 @default.
- W3100550724 creator A5012408072 @default.
- W3100550724 creator A5013070084 @default.
- W3100550724 creator A5016098233 @default.
- W3100550724 creator A5022726678 @default.
- W3100550724 creator A5026584653 @default.
- W3100550724 creator A5030672284 @default.
- W3100550724 creator A5033365504 @default.
- W3100550724 creator A5034251568 @default.
- W3100550724 creator A5036093986 @default.
- W3100550724 creator A5039169370 @default.
- W3100550724 creator A5041602011 @default.
- W3100550724 creator A5041762641 @default.
- W3100550724 creator A5042802099 @default.
- W3100550724 creator A5042998510 @default.
- W3100550724 creator A5044747691 @default.
- W3100550724 creator A5046238836 @default.
- W3100550724 creator A5046597133 @default.
- W3100550724 creator A5047980521 @default.
- W3100550724 creator A5053640868 @default.
- W3100550724 creator A5054907336 @default.
- W3100550724 creator A5056502880 @default.
- W3100550724 creator A5058932447 @default.
- W3100550724 creator A5062648446 @default.
- W3100550724 creator A5066505135 @default.
- W3100550724 creator A5066614300 @default.
- W3100550724 creator A5072965417 @default.
- W3100550724 creator A5078440061 @default.
- W3100550724 creator A5079524744 @default.
- W3100550724 creator A5080337591 @default.
- W3100550724 creator A5081231681 @default.
- W3100550724 creator A5082043392 @default.
- W3100550724 creator A5086674756 @default.
- W3100550724 creator A5089478486 @default.
- W3100550724 date "2020-12-01" @default.
- W3100550724 modified "2023-10-01" @default.
- W3100550724 title "Efficient AI System Design With Cross-Layer Approximate Computing" @default.
- W3100550724 cites W1996901117 @default.
- W3100550724 cites W1998917233 @default.
- W3100550724 cites W2009832130 @default.
- W3100550724 cites W2010468809 @default.
- W3100550724 cites W2013305145 @default.
- W3100550724 cites W2016053056 @default.
- W3100550724 cites W2020517863 @default.
- W3100550724 cites W2026005150 @default.
- W3100550724 cites W2043607059 @default.
- W3100550724 cites W2053323369 @default.
- W3100550724 cites W2064675550 @default.
- W3100550724 cites W2097117768 @default.
- W3100550724 cites W2118683741 @default.
- W3100550724 cites W2125203716 @default.
- W3100550724 cites W2139501017 @default.
- W3100550724 cites W2142883190 @default.
- W3100550724 cites W2145539833 @default.
- W3100550724 cites W2152839228 @default.
- W3100550724 cites W2156425544 @default.
- W3100550724 cites W2160815625 @default.
- W3100550724 cites W2187230075 @default.
- W3100550724 cites W2267244539 @default.
- W3100550724 cites W2285660444 @default.
- W3100550724 cites W2513554817 @default.
- W3100550724 cites W2516141709 @default.
- W3100550724 cites W2518511512 @default.
- W3100550724 cites W2553417306 @default.
- W3100550724 cites W2586654419 @default.
- W3100550724 cites W2606722458 @default.
- W3100550724 cites W2625264446 @default.
- W3100550724 cites W2625457103 @default.
- W3100550724 cites W2653504903 @default.
- W3100550724 cites W2657126969 @default.
- W3100550724 cites W2743069335 @default.
- W3100550724 cites W2748818695 @default.
- W3100550724 cites W2782511028 @default.
- W3100550724 cites W2785141883 @default.
- W3100550724 cites W2792893539 @default.
- W3100550724 cites W2809371234 @default.
- W3100550724 cites W2883542588 @default.
- W3100550724 cites W2893813411 @default.
- W3100550724 cites W2898913080 @default.
- W3100550724 cites W2945652047 @default.
- W3100550724 cites W2946574806 @default.
- W3100550724 cites W2963122961 @default.
- W3100550724 cites W2963125010 @default.
- W3100550724 cites W2963480671 @default.
- W3100550724 cites W2964080840 @default.
- W3100550724 cites W2980034233 @default.
- W3100550724 cites W2981680926 @default.
- W3100550724 cites W2981686346 @default.
- W3100550724 cites W2982144630 @default.
- W3100550724 cites W3005538384 @default.