國立虎尾科技大學 |

Architecting High Performance Silicon Systems for Accurate and Efficient On-Chip Deep Learning.

Record Type:	Language materials, manuscript : Monograph/item
Title/Author:	Architecting High Performance Silicon Systems for Accurate and Efficient On-Chip Deep Learning./
Author:	Tambe, Thierry.
Description:	1 online resource (137 pages)
Notes:	Source: Dissertations Abstracts International, Volume: 84-12, Section: B.
Contained By:	Dissertations Abstracts International84-12B.
Subject:	Electrical engineering. -
Online resource:	click for full text (PQDT)
ISBN:	9798379613839

Architecting High Performance Silicon Systems for Accurate and Efficient On-Chip Deep Learning.
Tambe, Thierry.

Architecting High Performance Silicon Systems for Accurate and Efficient On-Chip Deep Learning. - 1 online resource (137 pages)

Source: Dissertations Abstracts International, Volume: 84-12, Section: B.

Thesis (Ph.D.)--Harvard University, 2023.

Includes bibliographical references

The unabated pursuit of omniscient and omnipotent AI is levying hefty latency, memory, and energy taxes at all computing scales. At the same time, the twilight of Dennard scaling means traditional performance gains are no longer proportionally attained with reduction in transistor feature size -- compelling a global trend towards application-based hardware specialization.Over the course of my PhD, I have built a heterogeneity of solutions co-optimized across the algorithm, architecture, and silicon stack to generate breakthrough advances in arithmetic performance, compute density and flexibility, and energy efficiency for on-chip machine learning (ML), and natural language processing (NLP) in particular. My work aims to significantly increase the application space of embedded ML computing, in both the inference and training regimes, by coalescing innovative vectors spanning the algorithm, memory subsystem, hardware architecture, and circuit layers, while tuning their designs and inter-dependencies to promote greater performance, energy efficiency, and reliability within a silicon chip system.In the algorithm front, this thesis discusses best paper award-winning work on a novel floating-point based data type, AdaptivFloat, which enables resilient quantized AI computations; and is particularly suitable for NLP networks with large parameter distribution. To evaluate AdaptivFloat impact on a real system, this thesis describes a 16nm chip prototype that integrates FlexASR, a programmable hardware accelerator with AdaptivFloat-based processing elements, and specialized for attention-based recurrent neural networks used in speech and machine translation AI workloads. We further verify FlexASR fidelity to the front-end AI application via a formal hardware/software compiler interface. Towards the goal of lowering the prohibitive energy cost of inferencing large language models on TinyML devices, this dissertation describes a principled algorithm-hardware co-design solution, validated in a 12nm chip tapeout, that accelerates Transformer workloads by tailoring the accelerator's latency and energy expenditures according to the complexity of the input query it processes. Finally, recognizing that the overwhelming majority of the data generated during the deep learning training process exhibits a very short-lived lifetime, this thesis proposes leveraging non-conventional embedded dynamic RAMs (eDRAMs) as the main on-chip storage medium for ML training data -- which, along with a tightly-coupled offering of algorithmic alterations and custom hardware specialization, yields significant energy efficiency advantages over conventional SRAMs.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2024

Mode of access: World Wide Web

ISBN: 9798379613839Subjects--Topical Terms:

596380
Electrical engineering.
Subjects--Index Terms:

Deep learningIndex Terms--Genre/Form:

554714
Electronic books.

Architecting High Performance Silicon Systems for Accurate and Efficient On-Chip Deep Learning.
LDR:04060ntm a22003857 4500 001 1144266
005 20240531083814.5
006 m o d
007 cr mn ---uuuuu
008 250605s2023 xx obm 000 0 eng d
020 $a 9798379613839
035 $a (MiAaPQ)AAI30492097
035 $a AAI30492097
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Tambe, Thierry. $3 1469237
245 1 0 $a Architecting High Performance Silicon Systems for Accurate and Efficient On-Chip Deep Learning.
264 0 $c 2023
300 $a 1 online resource (137 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 84-12, Section: B.
500 $a Advisor: Wei, Gu-Yeon;Brooks, David.
502 $a Thesis (Ph.D.)--Harvard University, 2023.
504 $a Includes bibliographical references
520 $a The unabated pursuit of omniscient and omnipotent AI is levying hefty latency, memory, and energy taxes at all computing scales. At the same time, the twilight of Dennard scaling means traditional performance gains are no longer proportionally attained with reduction in transistor feature size -- compelling a global trend towards application-based hardware specialization.Over the course of my PhD, I have built a heterogeneity of solutions co-optimized across the algorithm, architecture, and silicon stack to generate breakthrough advances in arithmetic performance, compute density and flexibility, and energy efficiency for on-chip machine learning (ML), and natural language processing (NLP) in particular. My work aims to significantly increase the application space of embedded ML computing, in both the inference and training regimes, by coalescing innovative vectors spanning the algorithm, memory subsystem, hardware architecture, and circuit layers, while tuning their designs and inter-dependencies to promote greater performance, energy efficiency, and reliability within a silicon chip system.In the algorithm front, this thesis discusses best paper award-winning work on a novel floating-point based data type, AdaptivFloat, which enables resilient quantized AI computations; and is particularly suitable for NLP networks with large parameter distribution. To evaluate AdaptivFloat impact on a real system, this thesis describes a 16nm chip prototype that integrates FlexASR, a programmable hardware accelerator with AdaptivFloat-based processing elements, and specialized for attention-based recurrent neural networks used in speech and machine translation AI workloads. We further verify FlexASR fidelity to the front-end AI application via a formal hardware/software compiler interface. Towards the goal of lowering the prohibitive energy cost of inferencing large language models on TinyML devices, this dissertation describes a principled algorithm-hardware co-design solution, validated in a 12nm chip tapeout, that accelerates Transformer workloads by tailoring the accelerator's latency and energy expenditures according to the complexity of the input query it processes. Finally, recognizing that the overwhelming majority of the data generated during the deep learning training process exhibits a very short-lived lifetime, this thesis proposes leveraging non-conventional embedded dynamic RAMs (eDRAMs) as the main on-chip storage medium for ML training data -- which, along with a tightly-coupled offering of algorithmic alterations and custom hardware specialization, yields significant energy efficiency advantages over conventional SRAMs.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2024
538 $a Mode of access: World Wide Web
650 4 $a Electrical engineering. $3 596380
650 4 $a Computer science. $3 573171
653 $a Deep learning
653 $a Hardware accelerators
653 $a Natural language processing
653 $a Number systems
653 $a System-on-chips
655 7 $a Electronic books. $2 local $3 554714
690 $a 0544
690 $a 0984
710 2 $a ProQuest Information and Learning Co. $3 1178819
710 2 $a Harvard University. $b Engineering and Applied Sciences - Engineering Sciences. $3 1193033
773 0 $t Dissertations Abstracts International $g 84-12B.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=30492097 $z click for full text (PQDT)