Architectural contesting: exposing and exploiting temperamental behavior
Pages 28 - 35
Abstract
Previous studies have proposed techniques to dynamically change the architecture of a processor to better suit the characteristics of the workload at hand. However, all such approaches are prone to a fundamental trade-off between the architectural diversity they can provide and the latency of architectural change, their fixed-configuration performance and the complexity of finding the best architectural configuration for the workload at hand. In this study we argue that the full potential of dynamic architectural customization can only be achieved by diminishing the effect of the degree of available architectural diversity on the aforementioned performance factors.
The performance of a statically designed processing core in a heterogeneous multi-core system is independent of the architectural diversity available. In addition, it is apparent that concurrent execution of code on differently architected cores automatically reveals which architecture is more suitable for the characteristics of a particular workload.
We therefore propose architectural contesting; the redundant execution of code on a number of differently architected processors (each customized for a different set of workload characteristics) in a leader follower arrangement, such that the leader and follower cores continuously shift roles as one core or the other becomes more favorable for new code phases. In this manner effective execution is naturally transferred from one static architecture to the other with little latency.
In this study, we show that the contesting of only processor width can yield an average speedup of 7.5% and up to 12.5% in integer SPEC benchmarks.
References
[1]
S. Palacharla, N. P. Jouppi, and J. E. Smith, "Complexity-effective superscalar processors", In the 24th Inter. Symp. on Comp. Arch. (ISCA), June 2--4, 1997.
[2]
R. Kumar, K. I. Farkas, N. P. Jouppi, P. Ranganathan, and D. M. Tullsen. "Single-ISA Heterogeneous Multi-core Architectures: The Potential for Processor Power Reduction", In Inter. Symp. on Micro (MICRO), Dec. 2003
[3]
J. E. Smith. "Instruction-level distributed processing", IEEE Computer, 34(4):59.65, April 2001.
[4]
S. Balakrishnan, R. Rajwar, M. Upton, K. Lai, "The Impact of Performance Asymmetry in Emerging Multicore Architectures" 32nd Inter Symp on Comp Arch., June 2005.
[5]
A. Dhodapkar and J. Smith. "Managing Multi-Configuration Hardware via Dynamic Working Set Analysis", In Inter. Symp. on Comp. Arch. (ISCA), pp 233--244, 2002.
[6]
M. Huang, J. Renau, and J. Torrellas. "Profile-based energy reduction for high-performance processors". In 4th Workshop on Feedback-Directed and Dynamic Optimization (FDDO4), Dec. 2001.
[7]
T. Sherwood, S. Sair, and B. Calder, "Phase tracking and prediction", In 30th Inter. Symp. on Comp. Arch., 2003
[8]
H. Akkary and M. A. Driscoll, "A Dynamic Multithreading Processor", in Proc. 31st. Ann. Int. Symp. on Micro, 1998.
[9]
K. Sundaramoorthy, Z. Purser, and E. Rotenberg. "Slipstream Processors: Improving both Performance and Fault Tolerance" ASPLOS-IX, Nov. 2000.
[10]
D. Albonesi. "Dynamic IPC/clock rate optimization", the 25th Inter. Symp. on Comp. Arch. (ISCA), pp. 282--292, June 1998.
[11]
D. Ponomarev, G. Kucuk, O. Ergin, K. Ghose, P. Kogge, "Energy-Efficient Issue Queue Design", IEEE Transactions on Very Large Scale Integration Systems, 2003.
[12]
D. Folegnani and A. Gonzlez. "Energy-Effective Issue Logic", Inter. Symp. on Comp. Arch. (ISCA) 2001.
[13]
S.-H. Yang, M. D. Powell, B. Falsafi, K. Roy, and T. Vijaykumar, "An integrated circuit/architecture approach to reducing leakage in deep-submicron high-performance caches" In Inter. Symp. on High-Perf. Comp. Arch., pp 147--157, 2001.
[14]
R. Balasubramonian, D. Albonesi, A. Buyuktosunoglu, and S. Dwarkadas. "Memory Hierarchy Reconfiguration for Energy and Performance in General-Purpose Processor Architectures" In Inter. Symp. on Micro., pp 245--257, December 2000.
[15]
D. H. Albonesi et al. "Dynamically Tuning Processor Resources with Adaptive Processing". In IEEE Computer, December 2003.
[16]
R. Balasubramonian, D. Albonesi, A. Buyuktosunoglu, S. Dwarkadas, "Memory Hierarchy Reconfiguration for Energy and Performance in General-Purpose Processor Architectures", Inter. Symp. on Micro., 2000.
[17]
M. Huang, J. Renau, J. Torrellas, "Positional Adaptation of Processors: Application to Energy Reduction", Inter. Symp. on Comp. Arch., 2003.
[18]
E. Hao, P. Chang, and Y. Patt, "The Effect of Speculatively Updating Branch History on Branch Prediction Accuracy, Revisited", 27th Inter. Symp. on Micro. (MICRO), pp. 228--232, 1994.
[19]
K. Compton and S. Hauck. "Reconfigurable computing: A survey of systems and software", ACM Computing Surveys, 34(2), 2002.
[20]
S. G. Dropsho, G. Semeraro, D. H. Albonesi, G. Magklis, M. L. Scott, "Dynamically Trading Frequency for Complexity in a GALS Microprocessor", Inter. Symp. on Microarch. (MICRO), pp. 157--168, 2004.
[21]
D. Burger, S. Kaxiras, and J. R. Goodman, "DataScalar Architectures", 24th Inter. Symp. on Comp. Arch. (ISCA), June, 1997.
[22]
L. Chen, D. H. Albonesi, and S. Dropsho, "Dynamically Matching ILP Characteristics Via a Heterogeneous Clustered Microarchitecture", IBM Watson Conf. on the Intera. Between Arch., Circuits, and Compilers, pp. 136--143, Oct. 2004.
[23]
R. Kumar, D. M. Tullsen, N. P. Jouppi "Core architecture optimization for heterogeneous chip multiprocessors", Parallel Arch. and Comp. Tech. (PACT), pp. 23--32, 2006.
[24]
M. Cintra, J. F. Martínez, and J. Torrellas. Architectural support for scalable speculative parallelization in shared-memory multiprocessors. In Inter. Symp. on Comp. Arch. (ISCA), pp. 13--24, 2000
[25]
M. Franklin, "Multiscalar Processors", Kluwer Academic Publishers, 2002.
[26]
D. Tarjan, S. Thoziyoor; N. P. Jouppi, "CACTI 4.0", Technical report HPL-2006-86, 2006.
Index Terms
- Architectural contesting: exposing and exploiting temperamental behavior
Recommendations
Architectural Considerations for Application-Specific Counterflow Pipelines
ARVLSI '99: Proceedings of the 20th Anniversary Conference on Advanced Research in VLSIApplication-specific processor design is a promising approach for meeting the performance and cost goals of a system. Application- specific processors are especially promising for embedded systems (e.g., digital cameras, cellular phones, etc.) where a ...
Bottlenecks in Multimedia Processing with SIMD Style Extensions and Architectural Enhancements
Multimedia SIMD extensions such as MMX and AltiVec speed up media processing; however, our characterization shows that the attributes of current general-purpose processors enhanced with SIMD extensions do not match very well with the access patterns and ...
Comments
Information & Contributors
Information
Published In
Copyright © 2007 Authors.
Publisher
Association for Computing Machinery
New York, NY, United States
Publication History
Published: 01 June 2007
Published in SIGARCH Volume 35, Issue 3
Check for updates
Qualifiers
- Article
Contributors
Other Metrics
Bibliometrics & Citations
Bibliometrics
Article Metrics
- 0Total Citations
- 238Total Downloads
- Downloads (Last 12 months)1
- Downloads (Last 6 weeks)0
Reflects downloads up to 25 Dec 2024
Other Metrics
Citations
View Options
Login options
Check if you have access through your login credentials or your institution to get full access on this article.
Sign in