From High-Level Specification to High-Performance Code

Franz Franchetti, Jose M.F. Moura, David A. Padua, Jack Dongarra

Research output: Contribution to journalReview article

Abstract

Computer architectures and systems are becoming ever more powerful but increasingly more complex. With the end of frequency scaling (about 2004) and the era of multicores/manycores/accelerators, it is exceedingly hard to extract the promised performance, in particular, at a reasonable energy budget. Only highly trained and educated experts can hope to conquer this barrier that, if not appropriately dealt with, can translate into multiple orders of magnitude of underutilization of computer systems when programmed by less specialized programmers or domain scientists. To overcome this challenge, the last ten years have seen a flurry of activity to automate the design and generation of highly efficient implementations for these multicore/ manycore architectures, and to translate high level descriptions of programs into high performance and power efficiency.

Original languageEnglish (US)
Article number8510981
Pages (from-to)1875-1878
Number of pages4
JournalProceedings of the IEEE
Volume106
Issue number11
DOIs
StatePublished - Nov 2018

Fingerprint

Computer systems
Specifications
Computer architecture
Particle accelerators

ASJC Scopus subject areas

  • Electrical and Electronic Engineering

Cite this

Franchetti, F., Moura, J. M. F., Padua, D. A., & Dongarra, J. (2018). From High-Level Specification to High-Performance Code. Proceedings of the IEEE, 106(11), 1875-1878. [8510981]. https://doi.org/10.1109/JPROC.2018.2875253

From High-Level Specification to High-Performance Code. / Franchetti, Franz; Moura, Jose M.F.; Padua, David A.; Dongarra, Jack.

In: Proceedings of the IEEE, Vol. 106, No. 11, 8510981, 11.2018, p. 1875-1878.

Research output: Contribution to journalReview article

Franchetti, F, Moura, JMF, Padua, DA & Dongarra, J 2018, 'From High-Level Specification to High-Performance Code', Proceedings of the IEEE, vol. 106, no. 11, 8510981, pp. 1875-1878. https://doi.org/10.1109/JPROC.2018.2875253
Franchetti, Franz ; Moura, Jose M.F. ; Padua, David A. ; Dongarra, Jack. / From High-Level Specification to High-Performance Code. In: Proceedings of the IEEE. 2018 ; Vol. 106, No. 11. pp. 1875-1878.
@article{f3780e3a15e048d195df9cbce95613ee,
title = "From High-Level Specification to High-Performance Code",
abstract = "Computer architectures and systems are becoming ever more powerful but increasingly more complex. With the end of frequency scaling (about 2004) and the era of multicores/manycores/accelerators, it is exceedingly hard to extract the promised performance, in particular, at a reasonable energy budget. Only highly trained and educated experts can hope to conquer this barrier that, if not appropriately dealt with, can translate into multiple orders of magnitude of underutilization of computer systems when programmed by less specialized programmers or domain scientists. To overcome this challenge, the last ten years have seen a flurry of activity to automate the design and generation of highly efficient implementations for these multicore/ manycore architectures, and to translate high level descriptions of programs into high performance and power efficiency.",
author = "Franz Franchetti and Moura, {Jose M.F.} and Padua, {David A.} and Jack Dongarra",
year = "2018",
month = "11",
doi = "10.1109/JPROC.2018.2875253",
language = "English (US)",
volume = "106",
pages = "1875--1878",
journal = "Proceedings of the IEEE",
issn = "0018-9219",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "11",

}

TY - JOUR

T1 - From High-Level Specification to High-Performance Code

AU - Franchetti, Franz

AU - Moura, Jose M.F.

AU - Padua, David A.

AU - Dongarra, Jack

PY - 2018/11

Y1 - 2018/11

N2 - Computer architectures and systems are becoming ever more powerful but increasingly more complex. With the end of frequency scaling (about 2004) and the era of multicores/manycores/accelerators, it is exceedingly hard to extract the promised performance, in particular, at a reasonable energy budget. Only highly trained and educated experts can hope to conquer this barrier that, if not appropriately dealt with, can translate into multiple orders of magnitude of underutilization of computer systems when programmed by less specialized programmers or domain scientists. To overcome this challenge, the last ten years have seen a flurry of activity to automate the design and generation of highly efficient implementations for these multicore/ manycore architectures, and to translate high level descriptions of programs into high performance and power efficiency.

AB - Computer architectures and systems are becoming ever more powerful but increasingly more complex. With the end of frequency scaling (about 2004) and the era of multicores/manycores/accelerators, it is exceedingly hard to extract the promised performance, in particular, at a reasonable energy budget. Only highly trained and educated experts can hope to conquer this barrier that, if not appropriately dealt with, can translate into multiple orders of magnitude of underutilization of computer systems when programmed by less specialized programmers or domain scientists. To overcome this challenge, the last ten years have seen a flurry of activity to automate the design and generation of highly efficient implementations for these multicore/ manycore architectures, and to translate high level descriptions of programs into high performance and power efficiency.

UR - http://www.scopus.com/inward/record.url?scp=85055790343&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85055790343&partnerID=8YFLogxK

U2 - 10.1109/JPROC.2018.2875253

DO - 10.1109/JPROC.2018.2875253

M3 - Review article

AN - SCOPUS:85055790343

VL - 106

SP - 1875

EP - 1878

JO - Proceedings of the IEEE

JF - Proceedings of the IEEE

SN - 0018-9219

IS - 11

M1 - 8510981

ER -