Performance of runtime optimization on BLAST

Abhinav Das; Jiwei Lu; Howard Chen; Jinpyo Kim; Pen Chung Yew; Wei Chung Hsu; Dong Yuan Chen

doi:10.1109/CGO.2005.25

Performance of runtime optimization on BLAST

Abhinav Das, Jiwei Lu, Howard Chen, Jinpyo Kim, Pen Chung Yew, Wei Chung Hsu, Dong Yuan Chen

Computer Science and Engineering

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

6 Scopus citations

Abstract

Optimization of a real world application BLAST is used to demonstrate the limitations of static and profile-guided optimizations and to highlight the potential of runtime optimization systems. We analyze the performance profile of this application to determine performance bottlenecks and evaluate the effect of aggressive compiler optimizations on BLAST. We find that applying common optimizations (e.g. O3) can degrade performance. Profile guided optimizations do not show much improvement across the board, as current implementations do not address critical performance bottlenecks in BLAST. In some cases, these optimizations lower performance significantly due to unexpected secondary effects of aggressive optimizations. We also apply runtime optimization to BLAST using the ADORE framework. ADORE is able to detect performance bottlenecks and deploy optimizations resulting in performance gains up to 58% on some queries using data cache prefetching.

Original language	English (US)
Title of host publication	Proceedings of the 2005 International Symposium on Code Generation and Optimization, CGO 2005
Pages	86-95
Number of pages	10
DOIs	https://doi.org/10.1109/CGO.2005.25
State	Published - 2005

Publication series

Name	Proceedings of the 2005 International Symposium on Code Generation and Optimization, CGO 2005
Volume	2005

Access

10.1109/CGO.2005.25

OpenUrl availability

Full text

Cite this

Das, A., Lu, J., Chen, H., Kim, J., Yew, P. C., Hsu, W. C., & Chen, D. Y. (2005). Performance of runtime optimization on BLAST. In Proceedings of the 2005 International Symposium on Code Generation and Optimization, CGO 2005 (pp. 86-95). Article 1402079 (Proceedings of the 2005 International Symposium on Code Generation and Optimization, CGO 2005; Vol. 2005). https://doi.org/10.1109/CGO.2005.25

Performance of runtime optimization on BLAST. / Das, Abhinav; Lu, Jiwei; Chen, Howard et al.
Proceedings of the 2005 International Symposium on Code Generation and Optimization, CGO 2005. 2005. p. 86-95 1402079 (Proceedings of the 2005 International Symposium on Code Generation and Optimization, CGO 2005; Vol. 2005).

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

Das, A, Lu, J, Chen, H, Kim, J, Yew, PC, Hsu, WC & Chen, DY 2005, Performance of runtime optimization on BLAST. in Proceedings of the 2005 International Symposium on Code Generation and Optimization, CGO 2005., 1402079, Proceedings of the 2005 International Symposium on Code Generation and Optimization, CGO 2005, vol. 2005, pp. 86-95. https://doi.org/10.1109/CGO.2005.25

@inproceedings{1f2e499f4874405282c88e48139ff75f,

title = "Performance of runtime optimization on BLAST",

abstract = "Optimization of a real world application BLAST is used to demonstrate the limitations of static and profile-guided optimizations and to highlight the potential of runtime optimization systems. We analyze the performance profile of this application to determine performance bottlenecks and evaluate the effect of aggressive compiler optimizations on BLAST. We find that applying common optimizations (e.g. O3) can degrade performance. Profile guided optimizations do not show much improvement across the board, as current implementations do not address critical performance bottlenecks in BLAST. In some cases, these optimizations lower performance significantly due to unexpected secondary effects of aggressive optimizations. We also apply runtime optimization to BLAST using the ADORE framework. ADORE is able to detect performance bottlenecks and deploy optimizations resulting in performance gains up to 58% on some queries using data cache prefetching.",

author = "Abhinav Das and Jiwei Lu and Howard Chen and Jinpyo Kim and Yew, {Pen Chung} and Hsu, {Wei Chung} and Chen, {Dong Yuan}",

year = "2005",

doi = "10.1109/CGO.2005.25",

language = "English (US)",

isbn = "076952298X",

series = "Proceedings of the 2005 International Symposium on Code Generation and Optimization, CGO 2005",

pages = "86--95",

booktitle = "Proceedings of the 2005 International Symposium on Code Generation and Optimization, CGO 2005",

}

TY - GEN

T1 - Performance of runtime optimization on BLAST

AU - Das, Abhinav

AU - Lu, Jiwei

AU - Chen, Howard

AU - Kim, Jinpyo

AU - Yew, Pen Chung

AU - Hsu, Wei Chung

AU - Chen, Dong Yuan

PY - 2005

Y1 - 2005

N2 - Optimization of a real world application BLAST is used to demonstrate the limitations of static and profile-guided optimizations and to highlight the potential of runtime optimization systems. We analyze the performance profile of this application to determine performance bottlenecks and evaluate the effect of aggressive compiler optimizations on BLAST. We find that applying common optimizations (e.g. O3) can degrade performance. Profile guided optimizations do not show much improvement across the board, as current implementations do not address critical performance bottlenecks in BLAST. In some cases, these optimizations lower performance significantly due to unexpected secondary effects of aggressive optimizations. We also apply runtime optimization to BLAST using the ADORE framework. ADORE is able to detect performance bottlenecks and deploy optimizations resulting in performance gains up to 58% on some queries using data cache prefetching.

AB - Optimization of a real world application BLAST is used to demonstrate the limitations of static and profile-guided optimizations and to highlight the potential of runtime optimization systems. We analyze the performance profile of this application to determine performance bottlenecks and evaluate the effect of aggressive compiler optimizations on BLAST. We find that applying common optimizations (e.g. O3) can degrade performance. Profile guided optimizations do not show much improvement across the board, as current implementations do not address critical performance bottlenecks in BLAST. In some cases, these optimizations lower performance significantly due to unexpected secondary effects of aggressive optimizations. We also apply runtime optimization to BLAST using the ADORE framework. ADORE is able to detect performance bottlenecks and deploy optimizations resulting in performance gains up to 58% on some queries using data cache prefetching.

UR - http://www.scopus.com/inward/record.url?scp=33646846455&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33646846455&partnerID=8YFLogxK

U2 - 10.1109/CGO.2005.25

DO - 10.1109/CGO.2005.25

M3 - Conference contribution

AN - SCOPUS:33646846455

SN - 076952298X

SN - 9780769522982

T3 - Proceedings of the 2005 International Symposium on Code Generation and Optimization, CGO 2005

SP - 86

EP - 95

BT - Proceedings of the 2005 International Symposium on Code Generation and Optimization, CGO 2005

ER -

Performance of runtime optimization on BLAST

Abstract

Publication series

Access

OpenUrl availability

Other files and links

Cite this