Reducing power requirements of instruction scheduling through dynamic allocation of multiple datapath resources

ABSTRACT

The "one-size-fits-all" philosophy used for permanently allocating datapath resources in today's superscalar CPUs to maximize performance across a wide range of applications results in the overcommitment of resources in general. To reduce power dissipation in the datapath, the resource allocations can be dynamically adjusted based on the demands of applications. We propose a mechanism to dynamically, simultaneously and independently adjust the sizes of the issue queue (IQ), the reorder buffer (ROB) and the load/store queue (LSQ) based on the periodic sampling of their occupancies to achieve significant power savings with minimal impact on performance. Resource upsizing is done more aggressively (compared to downsizing) using the relative rate of blocked dispatches to limit the performance penalty. Our results are validated by the execution of SPEC 95 benchmark suite on a substantially modified version of Simplescalar simulator, where the IQ, the ROB, the LSQ and the register files are implemented as separate structures, as is the case with most practical implementations. For the SPEC 95 benchmarks, the use of our technique in a 4-way superscalar processor results in a power savings in excess of 70% within individual components and an average power savings of 53% for the IQ, LSQ and ROB combined for the entire benchmark suite with an average performance penalty of only 5%.

REFERENCES

• 1: Advanced Configuration and Power Interface Specification (Intel,Microsoft, Toshiba), 1999.
• 2: David H. Albonesi, Selective cache ways: on-demand cache resource allocation, Proceedings of the 32nd annual ACM/IEEE international symposium on Microarchitecture, p.248-259, November 16-18, 1999, Haifa, Israel
• 3: David Brooks , Vivek Tiwari , Margaret Martonosi, Wattch: a framework for architectural-level power analysis and optimizations, Proceedings of the 27th annual international symposium on Computer architecture, p.83-94, June 2000, Vancouver, British Columbia, Canada
• 4: David Brooks , Margaret Martonosi, Dynamic Thermal Management for High-Performance Microprocessors, Proceedings of the 7th International Symposium on High-Performance Computer Architecture, p.171, January 20-24, 2001
• 5: R. Iris Bahar , Srilatha Manne, Power and energy reduction via pipeline balancing, Proceedings of the 28th annual international symposium on Computer architecture, p.218-229, June 30-July 04, 2001, Göteborg, Sweden
• 6: Rajeev Balasubramonian , David Albonesi , Alper Buyuktosunoglu , Sandhya Dwarkadas, Memory hierarchy reconfiguration for energy and performance in general-purpose processor architectures, Proceedings of the 33rd annual ACM/IEEE international symposium on Microarchitecture, p.245-257, December 2000, Monterey, California, United States [doi>10.1145/360128.360153]
• 7: Dileep P. Bhandarkar, Alpha implementations and architecture: complete reference and guide, Digital Press, Newton, MA, 1996
• 8: Burger, D., and Austin, T. M., "The SimpleScalar tool set: Version 2.0", Tech. Report, Dept. of CS, Univ. of Wisconsin-Madison, June 1997 and documentation for all SimpleScalarreleases (through version 3.0).
• 9: Alper Buyuktosunoglu , David Albonesi , Stanley Schuster , David Brooks , Pradip Bose , Peter Cook, A circuit level implementation of an adaptive issue queue for power-aware microprocessors, Proceedings of the 11th Great Lakes symposium on VLSI, p.73-78, March 2001, West Lafayette, Indiana, United States [doi>10.1145/368122.368807]
• 10: Alper Buyuktosunoglu , Stanley Schuster , David Brooks , Pradip Bose , Peter W. Cook , David H. Albonesi, An Adaptive Issue Queue for Reduced Power at High Performance, Proceedings of the First International Workshop on Power-Aware Computer Systems-Revised Papers, p.25-39, November 12, 2000
• 11: Cai, G.,"ArchitecturalLevelPower/PerformanceOptimization and Dynamic Power Estimation", in Proc. of the Cool-Chips tutorial. An Industrial Perspective on Low Power Processor Design in conjunction with MICRO-32, 1999.
• 12: Daniele Folegnani , Antonio González, Energy-effective issue logic, Proceedings of the 28th annual international symposium on Computer architecture, p.230-239, June 30-July 04, 2001, Göteborg, Sweden
• 13: Ghiasi, S., Casmira, J.,and Grunwald, D.,"UsingIPCvariation in workloads with externally specified rates to reduce power consumption", in Proc. of the Workshop on Complexity-Effective Design, June 2000.
• 14: Michael Huang , Jose Renau , Seung-Moon Yoo , Josep Torrellas, A framework for dynamic energy efficiency and temperature management, Proceedings of the 33rd annual ACM/IEEE international symposium on Microarchitecture, p.202-213, December 2000, Monterey, California, United States [doi>10.1145/360128.360149]
• 15: Iyer, A. and Marculescu, D., "Run-time Scaling of MicroarchitectureResources in a Processor for Energy Savings", in Proc. of Kool Chips Workshop, held in conjunction with MICRO-33, December 2000.
• 16: Stefanos Kaxiras , Zhigang Hu , Margaret Martonosi, Cache decay: exploiting generational behavior to reduce cache leakage power, Proceedings of the 28th annual international symposium on Computer architecture, p.240-251, June 30-July 04, 2001, Göteborg, Sweden
• 17: Gurhan Kucuk , Kanad Ghose , Dimitry V. Ponomarev , Peter M. Kogge, Energy: efficient instruction dispatch buffer design for superscalar processors, Proceedings of the 2001 international symposium on Low power electronics and design, p.237-242, August 2001, Huntington Beach, California, United States [doi>10.1145/383082.383144]
• 18: Microprocessor Report, various issues, 1996-1999.
• 19: Palacharla, S., Jouppi, N. P. and Smith, J. E., "Quantifying the complexity of superscalar processors", Technical report CS-TR-96-1308, Dept. of CS, Univ. of Wisconsin, 1996.
• 20: Ponomarev, D., Kucuk, G., Ghose, K., "Dynamic Allocation of Datapath Resources for Low Power", in Proc. of Workshop on Complexity-Effective Design, held in conjunction with ISCA-28, June 2001.
• 21: Sherwood, T. and Calder, B., "Time Varying Behavior of Programs", Tech. Report No. CS99-630, Dept. of Computer Science and Engg., UCSD, August 1999.
• 22: David W. Wall, Limits of instruction-level parallelism, Proceedings of the fourth international conference on Architectural support for programming languages and operating systems, p.176-188, April 08-11, 1991, Santa Clara, California, United States
• 23: Wilcox, K., Manne, S., "Alpha processors: A History of Power Issuesand a Look to the Future",in Cool-Chips Tutorial, November 1999.
• 24: Se-Hyun Yang , Babak Falsafi , Michael D. Powell , Kaushik Roy , T. N. Vijaykumar, An Integrated Circuit/Architecture Approach to Reducing Leakage in Deep-Submicron High-Performance I-Caches, Proceedings of the 7th International Symposium on High-Performance Computer Architecture, p.147, January 20-24, 2001
• 25: V. Zyuban , P. Kogge, Optimization of high-performance superscalar architectures for energy efficiency, Proceedings of the 2000 international symposium on Low power electronics and design, p.84-89, July 25-27, 2000, Rapallo, Italy [doi>10.1145/344166.344522]

INDEX TERMS

Keywords:
dynamic instruction scheduling, energy-efficient datapath, power reduction, superscalar processor