國立高雄大學圖資館 |

Efficient execution of irregular dataflow graphshardware/software co-optimization for probabilistic AI and sparse linear algebra /

Record Type:	Electronic resources : Monograph/item
Title/Author:	Efficient execution of irregular dataflow graphsby Nimish Shah, Wannes Meert, Marian Verhelst.
Reminder of title:	hardware/software co-optimization for probabilistic AI and sparse linear algebra /
Author:	Shah, Nimish.
other author:	Meert, Wannes.
Published:	Cham :Springer Nature Switzerland :2023.
Description:	xxi, 143 p. :ill., digital ;24 cm.
Contained By:	Springer Nature eBook
Subject:	Electronic circuitsData processing.
Online resource:	https://doi.org/10.1007/978-3-031-33136-7
ISBN:	9783031331367$q(electronic bk.)

Efficient execution of irregular dataflow graphshardware/software co-optimization for probabilistic AI and sparse linear algebra /
Shah, Nimish.

Efficient execution of irregular dataflow graphshardware/software co-optimization for probabilistic AI and sparse linear algebra /[electronic resource] :by Nimish Shah, Wannes Meert, Marian Verhelst. - Cham :Springer Nature Switzerland :2023. - xxi, 143 p. :ill., digital ;24 cm.

Chapter 1. Irregular workloads at risk of losing the hardware lottery -- Chapter 2. Suitable data representation: A study of fixed point, floating point,and positTM formats for probabilistic AI -- Chapter 3. GraphOpt: constrained-optimization-based parallelization of irregular workloads for multicore processors -- Chapter 4. DAG Processing Unit version 1 (DPU): Efficient execution of irregular workloads on a multicore processor -- Chapter 5. DAG Processing Unit version 2 (DPU-v2): Efficient execution of irregular workloads on a spatial datapath -- Chapter 6. Conclusions and future work.

This book focuses on the acceleration of emerging irregular sparse workloads, posed by novel artificial intelligent (AI) models and sparse linear algebra. Specifically, the book outlines several co-optimized hardware-software solutions for a highly promising class of emerging sparse AI models called Probabilistic Circuit (PC) and a similar sparse matrix workload for triangular linear systems (SpTRSV) The authors describe optimizations for the entire stack, targeting applications, compilation, hardware architecture and silicon implementation, resulting in orders of magnitude higher performance and energy-efficiency compared to the existing state-of-the-art solutions. Thus, this book provides important building blocks for the upcoming generation of edge AI platforms. Analyzes the key bottlenecks in the existing platforms for these sparse and irregular AI and linear algebra algorithms; Discusses an emerging set of AI workloads that rely on sparse matrix operations and graph-based computations; Shows how to address the execution challenges of this novel class of algorithms through hardware-software codesign.

ISBN: 9783031331367$q(electronic bk.)

Standard No.: 10.1007/978-3-031-33136-7doiSubjects--Topical Terms:

182262
Electronic circuits
--Data processing.

LC Class. No.: TK7867 / .S53 2023

Dewey Class. No.: 621.3815028563

Efficient execution of irregular dataflow graphshardware/software co-optimization for probabilistic AI and sparse linear algebra /
LDR:02796nmm a2200325 a 4500 001 648597
003 DE-He213
005 20230713134339.0
006 m d
007 cr nn 008maaau
008 240116s2023 sz s 0 eng d
020 $a 9783031331367$q(electronic bk.)
020 $a 9783031331350$q(paper)
024 7 $a 10.1007/978-3-031-33136-7 $2 doi
035 $a 978-3-031-33136-7
040 $a GP $c GP
041 0 $a eng
050 4 $a TK7867 $b .S53 2023
072 7 $a TJFC $2 bicssc
072 7 $a TEC008010 $2 bisacsh
072 7 $a TJFC $2 thema
082 0 4 $a 621.3815028563 $2 23
090 $a TK7867 $b .S525 2023
100 1 $a Shah, Nimish. $3 958394
245 1 0 $a Efficient execution of irregular dataflow graphs $h [electronic resource] : $b hardware/software co-optimization for probabilistic AI and sparse linear algebra / $c by Nimish Shah, Wannes Meert, Marian Verhelst.
260 $a Cham : $b Springer Nature Switzerland : $b Imprint: Springer, $c 2023.
300 $a xxi, 143 p. : $b ill., digital ; $c 24 cm.
505 0 $a Chapter 1. Irregular workloads at risk of losing the hardware lottery -- Chapter 2. Suitable data representation: A study of fixed point, floating point,and positTM formats for probabilistic AI -- Chapter 3. GraphOpt: constrained-optimization-based parallelization of irregular workloads for multicore processors -- Chapter 4. DAG Processing Unit version 1 (DPU): Efficient execution of irregular workloads on a multicore processor -- Chapter 5. DAG Processing Unit version 2 (DPU-v2): Efficient execution of irregular workloads on a spatial datapath -- Chapter 6. Conclusions and future work.
520 $a This book focuses on the acceleration of emerging irregular sparse workloads, posed by novel artificial intelligent (AI) models and sparse linear algebra. Specifically, the book outlines several co-optimized hardware-software solutions for a highly promising class of emerging sparse AI models called Probabilistic Circuit (PC) and a similar sparse matrix workload for triangular linear systems (SpTRSV) The authors describe optimizations for the entire stack, targeting applications, compilation, hardware architecture and silicon implementation, resulting in orders of magnitude higher performance and energy-efficiency compared to the existing state-of-the-art solutions. Thus, this book provides important building blocks for the upcoming generation of edge AI platforms. Analyzes the key bottlenecks in the existing platforms for these sparse and irregular AI and linear algebra algorithms; Discusses an emerging set of AI workloads that rely on sparse matrix operations and graph-based computations; Shows how to address the execution challenges of this novel class of algorithms through hardware-software codesign.
650 0 $a Electronic circuits $x Data processing. $3 182262
650 0 $a Electronic circuit design $x Mathematical models. $3 485459
650 0 $a Embedded computer systems. $3 184402
650 1 4 $a Electronic Circuits and Systems. $3 913394
650 2 4 $a Embedded Systems. $3 919564
650 2 4 $a Machine Learning. $3 833608
700 1 $a Meert, Wannes. $3 892425
700 1 $a Verhelst, Marian. $3 375734
710 2 $a SpringerLink (Online service) $3 273601
773 0 $t Springer Nature eBook
856 4 0 $u https://doi.org/10.1007/978-3-031-33136-7
950 $a Engineering (SpringerNature-11647)