Static analysis of processor stall cycle aggregation

Jongeun Lee, Aviral Shrivastava

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

1 Citation (Scopus)

Abstract

Processor Idle Cycle Aggregation (PICA) is a promising approach for low power execution of processors, in which small memory stalls are aggregated to create a large one, and the processor is switched to low-power mode in it. We extend the previous proposed approach in two dimensions, i) We develop static analysis for the PICA technique and present optimum parameters for five common types of loops based on steady-state analysis, ii) We show that software only control is unable to guarantee its correctness in a varying runtime environment, potentially causing deadlocks. We enhance the robustness of PICA with minimal hardware extension, ensuring correct execution for any loops and parameters, which greatly facilitates exploration based parameter optimization. The combined use of our static analysis and exploration based fine-tuning makes the PICA technique applicable, to any memory-bound loop, with energy reduction. We validate our analytical models against simulation based optimization and also show through our experiments on embedded application benchmarks, that our technique can be applied to a wide range of loops with average 20% energy reductions compared to executions without PICA.

Original languageEnglish (US)
Title of host publicationEmbedded Systems Week 2008 - Proceedings of the 6th IEEE/ACM/IFIP International Conference on Hardware/Software Codesign and System Synthesis, CODES+ISSS 2008
Pages25-30
Number of pages6
DOIs
StatePublished - 2008
EventEmbedded Systems Week 2008 - 6th IEEE/ACM/IFIP International Conference on Hardware/Software Codesign and System Synthesis, CODES+ISSS 2008 - Atlanta, GA, United States
Duration: Oct 19 2008Oct 24 2008

Other

OtherEmbedded Systems Week 2008 - 6th IEEE/ACM/IFIP International Conference on Hardware/Software Codesign and System Synthesis, CODES+ISSS 2008
CountryUnited States
CityAtlanta, GA
Period10/19/0810/24/08

Fingerprint

Static analysis
Program processors
Agglomeration
Data storage equipment
Analytical models
Tuning
Hardware
Experiments

Keywords

  • Code transformation
  • Embedded systems
  • Low power
  • Memory bound loops
  • Processor free time
  • Stall cycle aggregation

ASJC Scopus subject areas

  • Hardware and Architecture
  • Software
  • Electrical and Electronic Engineering

Cite this

Lee, J., & Shrivastava, A. (2008). Static analysis of processor stall cycle aggregation. In Embedded Systems Week 2008 - Proceedings of the 6th IEEE/ACM/IFIP International Conference on Hardware/Software Codesign and System Synthesis, CODES+ISSS 2008 (pp. 25-30) https://doi.org/10.1145/1450135.1450143

Static analysis of processor stall cycle aggregation. / Lee, Jongeun; Shrivastava, Aviral.

Embedded Systems Week 2008 - Proceedings of the 6th IEEE/ACM/IFIP International Conference on Hardware/Software Codesign and System Synthesis, CODES+ISSS 2008. 2008. p. 25-30.

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

Lee, J & Shrivastava, A 2008, Static analysis of processor stall cycle aggregation. in Embedded Systems Week 2008 - Proceedings of the 6th IEEE/ACM/IFIP International Conference on Hardware/Software Codesign and System Synthesis, CODES+ISSS 2008. pp. 25-30, Embedded Systems Week 2008 - 6th IEEE/ACM/IFIP International Conference on Hardware/Software Codesign and System Synthesis, CODES+ISSS 2008, Atlanta, GA, United States, 10/19/08. https://doi.org/10.1145/1450135.1450143
Lee J, Shrivastava A. Static analysis of processor stall cycle aggregation. In Embedded Systems Week 2008 - Proceedings of the 6th IEEE/ACM/IFIP International Conference on Hardware/Software Codesign and System Synthesis, CODES+ISSS 2008. 2008. p. 25-30 https://doi.org/10.1145/1450135.1450143
Lee, Jongeun ; Shrivastava, Aviral. / Static analysis of processor stall cycle aggregation. Embedded Systems Week 2008 - Proceedings of the 6th IEEE/ACM/IFIP International Conference on Hardware/Software Codesign and System Synthesis, CODES+ISSS 2008. 2008. pp. 25-30
@inproceedings{1619b8bf8f2144eca713ac23f28c87a2,
title = "Static analysis of processor stall cycle aggregation",
abstract = "Processor Idle Cycle Aggregation (PICA) is a promising approach for low power execution of processors, in which small memory stalls are aggregated to create a large one, and the processor is switched to low-power mode in it. We extend the previous proposed approach in two dimensions, i) We develop static analysis for the PICA technique and present optimum parameters for five common types of loops based on steady-state analysis, ii) We show that software only control is unable to guarantee its correctness in a varying runtime environment, potentially causing deadlocks. We enhance the robustness of PICA with minimal hardware extension, ensuring correct execution for any loops and parameters, which greatly facilitates exploration based parameter optimization. The combined use of our static analysis and exploration based fine-tuning makes the PICA technique applicable, to any memory-bound loop, with energy reduction. We validate our analytical models against simulation based optimization and also show through our experiments on embedded application benchmarks, that our technique can be applied to a wide range of loops with average 20{\%} energy reductions compared to executions without PICA.",
keywords = "Code transformation, Embedded systems, Low power, Memory bound loops, Processor free time, Stall cycle aggregation",
author = "Jongeun Lee and Aviral Shrivastava",
year = "2008",
doi = "10.1145/1450135.1450143",
language = "English (US)",
isbn = "9781605584706",
pages = "25--30",
booktitle = "Embedded Systems Week 2008 - Proceedings of the 6th IEEE/ACM/IFIP International Conference on Hardware/Software Codesign and System Synthesis, CODES+ISSS 2008",

}

TY - GEN

T1 - Static analysis of processor stall cycle aggregation

AU - Lee, Jongeun

AU - Shrivastava, Aviral

PY - 2008

Y1 - 2008

N2 - Processor Idle Cycle Aggregation (PICA) is a promising approach for low power execution of processors, in which small memory stalls are aggregated to create a large one, and the processor is switched to low-power mode in it. We extend the previous proposed approach in two dimensions, i) We develop static analysis for the PICA technique and present optimum parameters for five common types of loops based on steady-state analysis, ii) We show that software only control is unable to guarantee its correctness in a varying runtime environment, potentially causing deadlocks. We enhance the robustness of PICA with minimal hardware extension, ensuring correct execution for any loops and parameters, which greatly facilitates exploration based parameter optimization. The combined use of our static analysis and exploration based fine-tuning makes the PICA technique applicable, to any memory-bound loop, with energy reduction. We validate our analytical models against simulation based optimization and also show through our experiments on embedded application benchmarks, that our technique can be applied to a wide range of loops with average 20% energy reductions compared to executions without PICA.

AB - Processor Idle Cycle Aggregation (PICA) is a promising approach for low power execution of processors, in which small memory stalls are aggregated to create a large one, and the processor is switched to low-power mode in it. We extend the previous proposed approach in two dimensions, i) We develop static analysis for the PICA technique and present optimum parameters for five common types of loops based on steady-state analysis, ii) We show that software only control is unable to guarantee its correctness in a varying runtime environment, potentially causing deadlocks. We enhance the robustness of PICA with minimal hardware extension, ensuring correct execution for any loops and parameters, which greatly facilitates exploration based parameter optimization. The combined use of our static analysis and exploration based fine-tuning makes the PICA technique applicable, to any memory-bound loop, with energy reduction. We validate our analytical models against simulation based optimization and also show through our experiments on embedded application benchmarks, that our technique can be applied to a wide range of loops with average 20% energy reductions compared to executions without PICA.

KW - Code transformation

KW - Embedded systems

KW - Low power

KW - Memory bound loops

KW - Processor free time

KW - Stall cycle aggregation

UR - http://www.scopus.com/inward/record.url?scp=78650760534&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=78650760534&partnerID=8YFLogxK

U2 - 10.1145/1450135.1450143

DO - 10.1145/1450135.1450143

M3 - Conference contribution

AN - SCOPUS:78650760534

SN - 9781605584706

SP - 25

EP - 30

BT - Embedded Systems Week 2008 - Proceedings of the 6th IEEE/ACM/IFIP International Conference on Hardware/Software Codesign and System Synthesis, CODES+ISSS 2008

ER -