BLAS for GPUs

Rajib Nath; Stanimire Tomov; Jack Dongarra

doi:10.1201/b10376

BLAS for GPUs

Rajib Nath, Stanimire Tomov, Jack Dongarra

Computer Science and Mathematics Div

Research output: Chapter in Book/Report/Conference proceeding › Chapter › peer-review

7 Scopus citations

Abstract

Recent activities of major chip manufacturers, such as Intel, AMD, IBM and NVIDIA, make it more evident than ever that future designs of microprocessors and large HPC systems will be hybrid/heterogeneous in nature, relying on the integration (in varying proportions) of two major types of components: 1. Multi-/many-cores CPU technology, where the number of cores will continue to escalate while avoiding the power wall, instruction level parallelism wall, and the memory wall [1]; and 2. Special purpose hardware and accelerators, especially GPUs, which are in commodity production, have outpaced standard CPUs in performance, and have become as easy-if not easier-to program than multicore CPUs.

Original language	English
Title of host publication	Scientific Computing with Multicore and Accelerators
Publisher	CRC Press
Pages	57-80
Number of pages	24
ISBN (Electronic)	9781439825372
ISBN (Print)	9781439825365
DOIs	https://doi.org/10.1201/b10376
State	Published - Jan 1 2010

Access to Document

10.1201/b10376

Cite this

@inbook{67092b3b0e67435dba92771184a6ab01,

title = "BLAS for GPUs",

abstract = "Recent activities of major chip manufacturers, such as Intel, AMD, IBM and NVIDIA, make it more evident than ever that future designs of microprocessors and large HPC systems will be hybrid/heterogeneous in nature, relying on the integration (in varying proportions) of two major types of components: 1. Multi-/many-cores CPU technology, where the number of cores will continue to escalate while avoiding the power wall, instruction level parallelism wall, and the memory wall [1]; and 2. Special purpose hardware and accelerators, especially GPUs, which are in commodity production, have outpaced standard CPUs in performance, and have become as easy-if not easier-to program than multicore CPUs.",

author = "Rajib Nath and Stanimire Tomov and Jack Dongarra",

note = "Publisher Copyright: {\textcopyright} 2011 by Taylor and Francis Group, LLC.",

year = "2010",

month = jan,

day = "1",

doi = "10.1201/b10376",

language = "English",

isbn = "9781439825365",

pages = "57--80",

booktitle = "Scientific Computing with Multicore and Accelerators",

publisher = "CRC Press",

}

TY - CHAP

T1 - BLAS for GPUs

AU - Nath, Rajib

AU - Tomov, Stanimire

AU - Dongarra, Jack

PY - 2010/1/1

Y1 - 2010/1/1

N2 - Recent activities of major chip manufacturers, such as Intel, AMD, IBM and NVIDIA, make it more evident than ever that future designs of microprocessors and large HPC systems will be hybrid/heterogeneous in nature, relying on the integration (in varying proportions) of two major types of components: 1. Multi-/many-cores CPU technology, where the number of cores will continue to escalate while avoiding the power wall, instruction level parallelism wall, and the memory wall [1]; and 2. Special purpose hardware and accelerators, especially GPUs, which are in commodity production, have outpaced standard CPUs in performance, and have become as easy-if not easier-to program than multicore CPUs.

AB - Recent activities of major chip manufacturers, such as Intel, AMD, IBM and NVIDIA, make it more evident than ever that future designs of microprocessors and large HPC systems will be hybrid/heterogeneous in nature, relying on the integration (in varying proportions) of two major types of components: 1. Multi-/many-cores CPU technology, where the number of cores will continue to escalate while avoiding the power wall, instruction level parallelism wall, and the memory wall [1]; and 2. Special purpose hardware and accelerators, especially GPUs, which are in commodity production, have outpaced standard CPUs in performance, and have become as easy-if not easier-to program than multicore CPUs.

UR - http://www.scopus.com/inward/record.url?scp=84969913998&partnerID=8YFLogxK

U2 - 10.1201/b10376

DO - 10.1201/b10376

M3 - Chapter

AN - SCOPUS:84969913998

SN - 9781439825365

SP - 57

EP - 80

BT - Scientific Computing with Multicore and Accelerators

PB - CRC Press

ER -

BLAS for GPUs

Abstract

Access to Document

Other files and links

Fingerprint

Cite this