Abstract
The practice of speculation in resolving data dependences has been recently studied as a means of extracting more instruction level parallelism (ILP). Each instruction’s outcome is predicted by value predictors. The instruction and its dependent instructions can be executed in parallel, thereby exploiting ILP aggressively. One of the serious hurdles for realizing data speculation is the huge hardware budget required by the predictors. In this paper, we propose techniques that exploit frequent value locality, resulting in a significant budget reduction. Based on these proposals, we evaluate two value predictors, named the zero-value predictor and the 0/1-value predictor. The zero-value predictor generates only value 0. Similarly, the 0/1-value predictor generates only values 0 and 1. Simulation results show that the proposed predictors have greater performance than does the last-value predictor which requires a hardware budget twice as large as that of the predictors. Therefore, the zero- and the 0/1-value predictors are promising candidates for cost-effective and practical value predictors which can be implemented in real microprocessors.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Brooks D., Martonosi M.: Dynamically exploiting narrow width operands to improve processor power and performance. 5th Int. Symp. on High Performance Computer Architecture (1999)
Burger D., Austin T.M.: The SimpleScalar tool set, version 2.0. ACM SIGARCH Computer Architecture News, 25(3) (1997)
Burtscher M., Zorn B.G.: Hybridizing and coalescing load value predictors. Int. Conf. on Computer Design (2000)
Calder B., Reinman G., Tullsen D.M.: Selective value prediction. 26th Int. Symp. on Computer Architecture (1999)
Del Pino S., Pinuel L., Moreno R.A., Tirado F.: Value prediction as a cost-effective solution to improve embedded processor performance. 3rd Int. Meeting on Vector and Parallel Processing (2000)
Fagin B.: Partial resolution in branch target buffers. IEEE Trans. Comput., 46(10) (1997)
Fu C-y., Jennings M.D., Larin S.Y., Conte T.M.: Software-only value speculation scheduling. Tech. Rep., Dept of Electr. and Comp. Eng., North Carolina State University (1998)
Gabbay F.: Speculative execution based on value prediction. Tech. Rep., Dept of Electr. Eng., Technion (1996)
Lee C-C., Chen I-C.K., Mudge T.N.: The bi-mode branch predictor. 30th Int. Symp. on Microarchitecture (1997)
Levy M. NEC processor goes out of order. Microprocessor Report, 15(9) (2001)
Lipasti M.H., Wilkerson C.B., Shen J.P.: Value locality and load value prediction. Int. Conf. on Architectural Support for Programming Languages and Operation Systems VII (1996)
Morancho E., Llaberia J.M., Olive A.: Split last-address predictor. Int. Conf. on Parallel Architectures and Compilation Techniques (1998)
Rychlik B., Faistl J.W., Krug B.P., Kurland A.Y., Sung J.J., Velev M.N., Shen J.P.: Efficient and accurate value prediction using dynamic classification. Tech. Rep., Dept of Electr. Comp. Eng., Carnegie Mellon University (1998)
Sato T., Arita I.: Table size reduction for data value predictors by exploiting narrow width values. 14th Int. Conf. on Supercomputing (2000)
Sato T., Arita I.: Partial resolution in data value predictors. 29th Int. Conf. on Parallel Processing (2000)
Sato T., Arita I.: Reducing hardware budget of data value predictors by exploiting frequent value locality. IEICE Tech. Rep. CPSY2000-62 (2000)
Sato T.: Evaluating the impact of reissued instructions on data speculative processor performance. Microprocessors and Microsystems, 25(9) (2002)
Sazeides Y., Smith J.E.: Implementations of context based value predictors. Tech. Rep., Dept of Electr. Comp. Eng., University of Wisconsin-Madison (1997)
Sohi G.S.: Instruction issue logic for high-performance, interruptible, multiple functional unit, pipelined computers. IEEE Trans. Comput., 39(3) (1990)
Tullsen D.M., Seng J.S.: Storageless value prediction using prior register values. 26th Int. Symp. on Computer Architecture (1999)
Wang K., Franklin M.: Highly accurate data value prediction using hybrid predictors. 30th Int. Symp. on Microarchitecture (1997)
Zhang Y., Yang J., Gupta R.: Frequent value locality and value-centric data cache design. Int. Conf. on Architectural Support for Programming Languages and Operation Systems IX (2000)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2002 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Sato, T., Arita, I. (2002). Low-Cost Value Predictors Using Frequent Value Locality. In: Zima, H.P., Joe, K., Sato, M., Seo, Y., Shimasaki, M. (eds) High Performance Computing. ISHPC 2002. Lecture Notes in Computer Science, vol 2327. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-47847-7_11
Download citation
DOI: https://doi.org/10.1007/3-540-47847-7_11
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-43674-4
Online ISBN: 978-3-540-47847-8
eBook Packages: Springer Book Archive