國立高雄大學圖資館 |

Language: English

Back

Automatic SIMD vectorization of SSA-...

Karrenberg, Ralf.

Automatic SIMD vectorization of SSA-based control flow graphs

Record Type:	Electronic resources : Monograph/item
Title/Author:	Automatic SIMD vectorization of SSA-based control flow graphsby Ralf Karrenberg.
Author:	Karrenberg, Ralf.
Published:	Wiesbaden :Springer Fachmedien Wiesbaden :2015.
Description:	xvi, 187 p. :ill., digital ;24 cm.
Contained By:	Springer eBooks
Subject:	Vector processing (Computer science)
Online resource:	http://dx.doi.org/10.1007/978-3-658-10113-8
ISBN:	9783658101138 (electronic bk.)

Automatic SIMD vectorization of SSA-based control flow graphs
Karrenberg, Ralf.

Automatic SIMD vectorization of SSA-based control flow graphs[electronic resource] /by Ralf Karrenberg. - Wiesbaden :Springer Fachmedien Wiesbaden :2015. - xvi, 187 p. :ill., digital ;24 cm.

Ralf Karrenberg presents Whole-Function Vectorization (WFV), an approach that allows a compiler to automatically create code that exploits data-parallelism using SIMD instructions. Data-parallel applications such as particle simulations, stock option price estimation, or video decoding require the same computations to be performed on huge amounts of data. Without WFV, one processor core executes a single instance of a data-parallel function. WFV transforms the function to execute multiple instances at once using SIMD instructions. The author describes an advanced WFV algorithm that includes a variety of analyses and code generation techniques. He shows that this approach improves the performance of the generated code in a variety of use cases. Contents Introduction, Foundations & Terminology, Related Work SIMD Property Analyses Whole-Function Vectorization Dynamic Code Variants, Evaluation, Conclusion, Outlook Target Groups Computer science researchers and students working in data-parallel computing Software and compiler engineers in the fields high-performance computing and compiler construction About the Author Ralf Karrenberg received his PhD in computer science at Saarland University in 2015. His seminal research on compilation techniques for SIMD architectures found wide recognition in both academia and the CPU and GPU industry. Currently, he is working for NVIDIA in Berlin. Prior to that, he contributed to research and development for visual effects in blockbuster movies at Weta Digital, New Zealand.

ISBN: 9783658101138 (electronic bk.)

Standard No.: 10.1007/978-3-658-10113-8doiSubjects--Topical Terms:

185817
Vector processing (Computer science)

LC Class. No.: QA76.5

Dewey Class. No.: 004.35

Automatic SIMD vectorization of SSA-based control flow graphs
LDR:02525nmm a2200325 a 4500 001 471285
003 DE-He213
005 20160115151321.0
006 m d
007 cr nn 008maaau
008 160223s2015 gw s 0 eng d
020 $a 9783658101138 (electronic bk.)
020 $a 9783658101121 (paper)
024 7 $a 10.1007/978-3-658-10113-8 $2 doi
035 $a 978-3-658-10113-8
040 $a GP $c GP
041 1 $a eng $b eng $b ger $h eng
050 4 $a QA76.5
072 7 $a UMX $2 bicssc
072 7 $a UMC $2 bicssc
072 7 $a COM051010 $2 bisacsh
072 7 $a COM010000 $2 bisacsh
082 0 4 $a 004.35 $2 23
090 $a QA76.5 $b .K18 2015
100 1 $a Karrenberg, Ralf. $3 726428
245 1 0 $a Automatic SIMD vectorization of SSA-based control flow graphs $h [electronic resource] / $c by Ralf Karrenberg.
260 $a Wiesbaden : $b Springer Fachmedien Wiesbaden : $b Imprint: Springer Vieweg, $c 2015.
300 $a xvi, 187 p. : $b ill., digital ; $c 24 cm.
520 $a Ralf Karrenberg presents Whole-Function Vectorization (WFV), an approach that allows a compiler to automatically create code that exploits data-parallelism using SIMD instructions. Data-parallel applications such as particle simulations, stock option price estimation, or video decoding require the same computations to be performed on huge amounts of data. Without WFV, one processor core executes a single instance of a data-parallel function. WFV transforms the function to execute multiple instances at once using SIMD instructions. The author describes an advanced WFV algorithm that includes a variety of analyses and code generation techniques. He shows that this approach improves the performance of the generated code in a variety of use cases. Contents Introduction, Foundations & Terminology, Related Work SIMD Property Analyses Whole-Function Vectorization Dynamic Code Variants, Evaluation, Conclusion, Outlook Target Groups Computer science researchers and students working in data-parallel computing Software and compiler engineers in the fields high-performance computing and compiler construction About the Author Ralf Karrenberg received his PhD in computer science at Saarland University in 2015. His seminal research on compilation techniques for SIMD architectures found wide recognition in both academia and the CPU and GPU industry. Currently, he is working for NVIDIA in Berlin. Prior to that, he contributed to research and development for visual effects in blockbuster movies at Weta Digital, New Zealand.
650 0 $a Vector processing (Computer science) $3 185817
650 0 $a Compilers (Computer programs) $3 219942
650 0 $a Parallel processing (Electronic computers) $3 184697
650 1 4 $a Computer Science. $3 212513
650 2 4 $a Programming Languages, Compilers, Interpreters. $3 274102
650 2 4 $a Computer Graphics. $3 274515
650 2 4 $a Appl.Mathematics/Computational Methods of Engineering. $3 273758
710 2 $a SpringerLink (Online service) $3 273601
773 0 $t Springer eBooks
856 4 0 $u http://dx.doi.org/10.1007/978-3-658-10113-8
950 $a Computer Science (Springer-11645)