The El'brus-3 and MARS-M: Recent advances in Russian high-performance computing

Mikhail N. Dorozhevets, Peter Wolcott

Research output: Contribution to journalArticle

8 Citations (Scopus)

Abstract

The El'brus-3 and MARS-M represent two recent efforts to address the Soviet Union's high-performance computing needs through original, indigenous development. The El'brus-3 extends very long instruction word (VLIW) concepts to a multiprocessor environment and offers features that increase performance and efficiency and decrease code size for both scientific and general-purpose applications. It incorporates procedure static and globally dynamic instruction scheduling, multiple, simultaneous branch path execution, and iteration frames for executing loops with recurrences and conditional branches. The MARS-M integrates VLIW, data flow, decoupled heterogeneous processors, and hierarchical systems into a unified framework. It also offers a combination of static and dynamic VLIW scheduling. While the viability of these machines has been demonstrated, significant barriers to their production and use remain.

Original languageEnglish (US)
Pages (from-to)5-48
Number of pages44
JournalThe Journal of Supercomputing
Volume6
Issue number1
DOIs
StatePublished - Mar 1 1992

Fingerprint

Branch
High Performance
Scheduling
Instruction Scheduling
Hierarchical systems
Hierarchical Systems
Dynamic Scheduling
Computing
Multiprocessor
Data Flow
Viability
Recurrence
Union
Integrate
Iteration
Decrease
Path
Concepts
Framework

Keywords

  • El'brus-3
  • MARS-M
  • VLIW
  • high-performance computing
  • parallel processing

ASJC Scopus subject areas

  • Software
  • Theoretical Computer Science
  • Information Systems
  • Hardware and Architecture

Cite this

The El'brus-3 and MARS-M : Recent advances in Russian high-performance computing. / Dorozhevets, Mikhail N.; Wolcott, Peter.

In: The Journal of Supercomputing, Vol. 6, No. 1, 01.03.1992, p. 5-48.

Research output: Contribution to journalArticle

@article{dbcbd256bc634780bc498630d38de77c,
title = "The El'brus-3 and MARS-M: Recent advances in Russian high-performance computing",
abstract = "The El'brus-3 and MARS-M represent two recent efforts to address the Soviet Union's high-performance computing needs through original, indigenous development. The El'brus-3 extends very long instruction word (VLIW) concepts to a multiprocessor environment and offers features that increase performance and efficiency and decrease code size for both scientific and general-purpose applications. It incorporates procedure static and globally dynamic instruction scheduling, multiple, simultaneous branch path execution, and iteration frames for executing loops with recurrences and conditional branches. The MARS-M integrates VLIW, data flow, decoupled heterogeneous processors, and hierarchical systems into a unified framework. It also offers a combination of static and dynamic VLIW scheduling. While the viability of these machines has been demonstrated, significant barriers to their production and use remain.",
keywords = "El'brus-3, MARS-M, VLIW, high-performance computing, parallel processing",
author = "Dorozhevets, {Mikhail N.} and Peter Wolcott",
year = "1992",
month = "3",
day = "1",
doi = "10.1007/BF00128641",
language = "English (US)",
volume = "6",
pages = "5--48",
journal = "Journal of Supercomputing",
issn = "0920-8542",
publisher = "Springer Netherlands",
number = "1",

}

TY - JOUR

T1 - The El'brus-3 and MARS-M

T2 - Recent advances in Russian high-performance computing

AU - Dorozhevets, Mikhail N.

AU - Wolcott, Peter

PY - 1992/3/1

Y1 - 1992/3/1

N2 - The El'brus-3 and MARS-M represent two recent efforts to address the Soviet Union's high-performance computing needs through original, indigenous development. The El'brus-3 extends very long instruction word (VLIW) concepts to a multiprocessor environment and offers features that increase performance and efficiency and decrease code size for both scientific and general-purpose applications. It incorporates procedure static and globally dynamic instruction scheduling, multiple, simultaneous branch path execution, and iteration frames for executing loops with recurrences and conditional branches. The MARS-M integrates VLIW, data flow, decoupled heterogeneous processors, and hierarchical systems into a unified framework. It also offers a combination of static and dynamic VLIW scheduling. While the viability of these machines has been demonstrated, significant barriers to their production and use remain.

AB - The El'brus-3 and MARS-M represent two recent efforts to address the Soviet Union's high-performance computing needs through original, indigenous development. The El'brus-3 extends very long instruction word (VLIW) concepts to a multiprocessor environment and offers features that increase performance and efficiency and decrease code size for both scientific and general-purpose applications. It incorporates procedure static and globally dynamic instruction scheduling, multiple, simultaneous branch path execution, and iteration frames for executing loops with recurrences and conditional branches. The MARS-M integrates VLIW, data flow, decoupled heterogeneous processors, and hierarchical systems into a unified framework. It also offers a combination of static and dynamic VLIW scheduling. While the viability of these machines has been demonstrated, significant barriers to their production and use remain.

KW - El'brus-3

KW - MARS-M

KW - VLIW

KW - high-performance computing

KW - parallel processing

UR - http://www.scopus.com/inward/record.url?scp=0009361641&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0009361641&partnerID=8YFLogxK

U2 - 10.1007/BF00128641

DO - 10.1007/BF00128641

M3 - Article

AN - SCOPUS:0009361641

VL - 6

SP - 5

EP - 48

JO - Journal of Supercomputing

JF - Journal of Supercomputing

SN - 0920-8542

IS - 1

ER -