Recursive statistical blockade: An enhanced technique for rare event simulation with application to SRAM circuit design

Amith Singhee; Jiajing Wang; Benton H. Calhoun; Rob A. Rutenbar

doi:10.1109/VLSI.2008.54

Recursive statistical blockade: An enhanced technique for rare event simulation with application to SRAM circuit design

Amith Singhee, Jiajing Wang, Benton H. Calhoun, Rob A. Rutenbar

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

Circuit reliability under statistical process variation is an area of growing concern. For highly replicated circuits such as SRAMs and flip flops, a rare statistical event for one circuit may induce a not-so-rare system failure. The authors of [1] proposed Statistical Blockade as a Monte Carlo technique that allows us to efficiently filter-to block-unwanted samples insufficiently rare in the tail distributions we seek. However, there are significant practical problems with the technique. In this work, we show common scenarios in SRAM design where these problems render Statistical Blockade ineffective. We then propose significant extensions to make Statistical Blockade practically usable in these common scenarios. We show speedups of 10²+ over standard Statistical Blockade and 10⁴+ over standard Monte Carlo, for an SRAM cell in an industrial 90nm technology.

Original language	English (US)
Title of host publication	Proceedings - 21st International Conference on VLSI Design, VLSI DESIGN 2008
Pages	131-136
Number of pages	6
DOIs	https://doi.org/10.1109/VLSI.2008.54
State	Published - 2008
Externally published	Yes
Event	21st International Conference on VLSI Design, VLSI DESIGN 2008 - Hyderabad, India Duration: Jan 4 2008 → Jan 8 2008

Publication series

Name	Proceedings of the IEEE International Frequency Control Symposium and Exposition

Other

Other	21st International Conference on VLSI Design, VLSI DESIGN 2008
Country/Territory	India
City	Hyderabad
Period	1/4/08 → 1/8/08

ASJC Scopus subject areas

General Engineering

Online availability

10.1109/VLSI.2008.54

Library availability

Discover UIUC Full Text

Cite this

Singhee, A., Wang, J., Calhoun, B. H., & Rutenbar, R. A. (2008). Recursive statistical blockade: An enhanced technique for rare event simulation with application to SRAM circuit design. In Proceedings - 21st International Conference on VLSI Design, VLSI DESIGN 2008 (pp. 131-136). Article 4450492 (Proceedings of the IEEE International Frequency Control Symposium and Exposition). https://doi.org/10.1109/VLSI.2008.54

Recursive statistical blockade: An enhanced technique for rare event simulation with application to SRAM circuit design. / Singhee, Amith; Wang, Jiajing; Calhoun, Benton H. et al.
Proceedings - 21st International Conference on VLSI Design, VLSI DESIGN 2008. 2008. p. 131-136 4450492 (Proceedings of the IEEE International Frequency Control Symposium and Exposition).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Singhee, A, Wang, J, Calhoun, BH & Rutenbar, RA 2008, Recursive statistical blockade: An enhanced technique for rare event simulation with application to SRAM circuit design. in Proceedings - 21st International Conference on VLSI Design, VLSI DESIGN 2008., 4450492, Proceedings of the IEEE International Frequency Control Symposium and Exposition, pp. 131-136, 21st International Conference on VLSI Design, VLSI DESIGN 2008, Hyderabad, India, 1/4/08. https://doi.org/10.1109/VLSI.2008.54

Singhee A, Wang J, Calhoun BH, Rutenbar RA. Recursive statistical blockade: An enhanced technique for rare event simulation with application to SRAM circuit design. In Proceedings - 21st International Conference on VLSI Design, VLSI DESIGN 2008. 2008. p. 131-136. 4450492. (Proceedings of the IEEE International Frequency Control Symposium and Exposition). doi: 10.1109/VLSI.2008.54

@inproceedings{0060dee2519d4f6d81304fcb4ad3cd9f,

title = "Recursive statistical blockade: An enhanced technique for rare event simulation with application to SRAM circuit design",

abstract = "Circuit reliability under statistical process variation is an area of growing concern. For highly replicated circuits such as SRAMs and flip flops, a rare statistical event for one circuit may induce a not-so-rare system failure. The authors of [1] proposed Statistical Blockade as a Monte Carlo technique that allows us to efficiently filter-to block-unwanted samples insufficiently rare in the tail distributions we seek. However, there are significant practical problems with the technique. In this work, we show common scenarios in SRAM design where these problems render Statistical Blockade ineffective. We then propose significant extensions to make Statistical Blockade practically usable in these common scenarios. We show speedups of 102+ over standard Statistical Blockade and 104+ over standard Monte Carlo, for an SRAM cell in an industrial 90nm technology.",

author = "Amith Singhee and Jiajing Wang and Calhoun, {Benton H.} and Rutenbar, {Rob A.}",

year = "2008",

doi = "10.1109/VLSI.2008.54",

language = "English (US)",

isbn = "0769530834",

series = "Proceedings of the IEEE International Frequency Control Symposium and Exposition",

pages = "131--136",

booktitle = "Proceedings - 21st International Conference on VLSI Design, VLSI DESIGN 2008",

}

TY - GEN

T1 - Recursive statistical blockade

T2 - 21st International Conference on VLSI Design, VLSI DESIGN 2008

AU - Singhee, Amith

AU - Wang, Jiajing

AU - Calhoun, Benton H.

AU - Rutenbar, Rob A.

PY - 2008

Y1 - 2008

N2 - Circuit reliability under statistical process variation is an area of growing concern. For highly replicated circuits such as SRAMs and flip flops, a rare statistical event for one circuit may induce a not-so-rare system failure. The authors of [1] proposed Statistical Blockade as a Monte Carlo technique that allows us to efficiently filter-to block-unwanted samples insufficiently rare in the tail distributions we seek. However, there are significant practical problems with the technique. In this work, we show common scenarios in SRAM design where these problems render Statistical Blockade ineffective. We then propose significant extensions to make Statistical Blockade practically usable in these common scenarios. We show speedups of 102+ over standard Statistical Blockade and 104+ over standard Monte Carlo, for an SRAM cell in an industrial 90nm technology.

AB - Circuit reliability under statistical process variation is an area of growing concern. For highly replicated circuits such as SRAMs and flip flops, a rare statistical event for one circuit may induce a not-so-rare system failure. The authors of [1] proposed Statistical Blockade as a Monte Carlo technique that allows us to efficiently filter-to block-unwanted samples insufficiently rare in the tail distributions we seek. However, there are significant practical problems with the technique. In this work, we show common scenarios in SRAM design where these problems render Statistical Blockade ineffective. We then propose significant extensions to make Statistical Blockade practically usable in these common scenarios. We show speedups of 102+ over standard Statistical Blockade and 104+ over standard Monte Carlo, for an SRAM cell in an industrial 90nm technology.

UR - http://www.scopus.com/inward/record.url?scp=47649084200&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=47649084200&partnerID=8YFLogxK

U2 - 10.1109/VLSI.2008.54

DO - 10.1109/VLSI.2008.54

M3 - Conference contribution

AN - SCOPUS:47649084200

SN - 0769530834

SN - 9780769530833

T3 - Proceedings of the IEEE International Frequency Control Symposium and Exposition

SP - 131

EP - 136

BT - Proceedings - 21st International Conference on VLSI Design, VLSI DESIGN 2008

Y2 - 4 January 2008 through 8 January 2008

ER -

Recursive statistical blockade: An enhanced technique for rare event simulation with application to SRAM circuit design

Abstract

Publication series

Other

ASJC Scopus subject areas

Online availability

Library availability

Related links

Fingerprint

Cite this