Application of compiler-assisted multiple-instruction retry to VLIW architectures

Shyh Kwei Chen, W. Kent Fuchs, Wen-Mei W Hwu

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Two compiler-assisted multiple-instruction word retry schemes for very long instruction word (VLIW) architectures are described. The first scheme compacts the compiler-generated hazard-free code with different degrees of rollback capability for scalar processors [14], and inserts no-ops in the scheduled code words. The second scheme employs a hardware read buffer [2] to resolve frequently occurring data hazards, and utilizes the compiler to resolve the remaining hazards.

Original languageEnglish (US)
Title of host publicationProceedings of the Conference on Fault-Tolerant Parallel and Distributed Systems
PublisherIEEE
Pages51-58
Number of pages8
StatePublished - 1995
EventProceedings of the 1995 Fault-Tolerant Parallel and Distributed Systems - Galveston, TX, USA
Duration: Jun 13 1994Jun 14 1994

Other

OtherProceedings of the 1995 Fault-Tolerant Parallel and Distributed Systems
CityGalveston, TX, USA
Period6/13/946/14/94

Fingerprint

Very long instruction word architecture
Hazards
Hardware

ASJC Scopus subject areas

  • Computer Science(all)

Cite this

Chen, S. K., Fuchs, W. K., & Hwu, W-M. W. (1995). Application of compiler-assisted multiple-instruction retry to VLIW architectures. In Proceedings of the Conference on Fault-Tolerant Parallel and Distributed Systems (pp. 51-58). IEEE.

Application of compiler-assisted multiple-instruction retry to VLIW architectures. / Chen, Shyh Kwei; Fuchs, W. Kent; Hwu, Wen-Mei W.

Proceedings of the Conference on Fault-Tolerant Parallel and Distributed Systems. IEEE, 1995. p. 51-58.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Chen, SK, Fuchs, WK & Hwu, W-MW 1995, Application of compiler-assisted multiple-instruction retry to VLIW architectures. in Proceedings of the Conference on Fault-Tolerant Parallel and Distributed Systems. IEEE, pp. 51-58, Proceedings of the 1995 Fault-Tolerant Parallel and Distributed Systems, Galveston, TX, USA, 6/13/94.
Chen SK, Fuchs WK, Hwu W-MW. Application of compiler-assisted multiple-instruction retry to VLIW architectures. In Proceedings of the Conference on Fault-Tolerant Parallel and Distributed Systems. IEEE. 1995. p. 51-58
Chen, Shyh Kwei ; Fuchs, W. Kent ; Hwu, Wen-Mei W. / Application of compiler-assisted multiple-instruction retry to VLIW architectures. Proceedings of the Conference on Fault-Tolerant Parallel and Distributed Systems. IEEE, 1995. pp. 51-58
@inproceedings{feac673ba626495c84107ab1f9bce64e,
title = "Application of compiler-assisted multiple-instruction retry to VLIW architectures",
abstract = "Two compiler-assisted multiple-instruction word retry schemes for very long instruction word (VLIW) architectures are described. The first scheme compacts the compiler-generated hazard-free code with different degrees of rollback capability for scalar processors [14], and inserts no-ops in the scheduled code words. The second scheme employs a hardware read buffer [2] to resolve frequently occurring data hazards, and utilizes the compiler to resolve the remaining hazards.",
author = "Chen, {Shyh Kwei} and Fuchs, {W. Kent} and Hwu, {Wen-Mei W}",
year = "1995",
language = "English (US)",
pages = "51--58",
booktitle = "Proceedings of the Conference on Fault-Tolerant Parallel and Distributed Systems",
publisher = "IEEE",

}

TY - GEN

T1 - Application of compiler-assisted multiple-instruction retry to VLIW architectures

AU - Chen, Shyh Kwei

AU - Fuchs, W. Kent

AU - Hwu, Wen-Mei W

PY - 1995

Y1 - 1995

N2 - Two compiler-assisted multiple-instruction word retry schemes for very long instruction word (VLIW) architectures are described. The first scheme compacts the compiler-generated hazard-free code with different degrees of rollback capability for scalar processors [14], and inserts no-ops in the scheduled code words. The second scheme employs a hardware read buffer [2] to resolve frequently occurring data hazards, and utilizes the compiler to resolve the remaining hazards.

AB - Two compiler-assisted multiple-instruction word retry schemes for very long instruction word (VLIW) architectures are described. The first scheme compacts the compiler-generated hazard-free code with different degrees of rollback capability for scalar processors [14], and inserts no-ops in the scheduled code words. The second scheme employs a hardware read buffer [2] to resolve frequently occurring data hazards, and utilizes the compiler to resolve the remaining hazards.

UR - http://www.scopus.com/inward/record.url?scp=0029491036&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0029491036&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:0029491036

SP - 51

EP - 58

BT - Proceedings of the Conference on Fault-Tolerant Parallel and Distributed Systems

PB - IEEE

ER -