Abstract
This work investigates the potential of direction-correlations to improve branch prediction. There are two types of direction- correlation: affectors and affectees. This work considers for the first time their implications at a basic level. These correlations are determined based on dataflow graph information and are used to select the subset of global branch history bits used for prediction. If this subset is small then affectors and affectees can be useful to cut down learning time, and reduce aliasing in prediction tables. This paper extends previous work explaining why and how correlation-based predictors work by analyzing the properties of direction-correlations. It also shows that branch history selected using oracle knowledge of direction-correlations improves the accuracy of the limit and realistic conditional branch predictors, that won at the recent branch prediction contest, by up to 30% and 17% respectively. The findings in this paper call for the investigation of predictors that can learn efficiently correlations from long branch history that may be non-consecutive with holes between them.
This is a preview of subscription content, log in via an institution to check access.
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
WIlkerson, C., Stark, J.: Introduction to JILP’s Special Edition for Finalists of the Championship Branch Prediction (CBP1) Competition. Journal of Instruction-Level Parallelism 7 (2005)
Jiménez, D.A.: The Second Championship Branch Prediction Competition. Journal of Instruction-Level Parallelism 9 (2007)
Seznec, A.: The Idealistic GTL Predictor. Journal of Instruction-Level Parallelism 9 (2007)
Evers, M., Patel, S.J., Chappel, R.S., Patt, Y.N.: An Analysis of Correlation and Predictability: What Makes Two-Level Branch Predictors Work. In: 25th International Symposium on Computer Architecture (June 1998)
McFarling, S.: Combining Branch Predictors. Technical Report DEC WRL TN-36, Digital Western Research Laboratory (June 1993)
Thomas, R., Franklin, M., Wilkerson, C., Stark, J.: Improving Branch Prediction by Dynamic Dataflow-based Identification of Correlated Branches from a Large Global History. In: 30th International Symposium on Computer Architecture, pp. 314–323 (June 2003)
Jimenez, D.A., Lin, C.: Dynamic Branch Prediction with Perceptrons. In: 7th International Symposium on High Performance Computer Architecture (February 2001)
Burger, D., Austin, T.M., Bennett, S.: Evaluating Future Microprocessors: The SimpleScalar Tool Set. Technical Report CS-TR-96-1308, University of Wisconsin-Madison (July 1996)
Perelman, E., Hamerly, G., Biesbrouck, M.V., Sherwood, T., Calder, B.: Using SimPoint for Accurate and Efficient Simulation. In: International Conference on Measurement and Modeling of Computer Systems (2003)
Seznec, A.: The L-TAGE Branch Predictor. Journal of Instruction-Level Parallelism 9 (2007)
Seznec, A.: Analysis of the O-GEometric History Length branch predictor. In: 32nd International Symposium on Computer Architecture (2005)
Smith, J.E.: A Study of Branch Prediction Strategies. In: 8th International Symposium on Computer Architecture, pp. 135–148 (May 1981)
Yeh, T.Y., Patt, Y.N.: Two-Level Adaptive Branch Prediction. In: 24th International Symposium on Microarchitecture, pp. 51–61 (November 1991)
Juan, T., Sanjeevan, S., Navarro, J.J.: Dynamic History-Length Fitting: A third level of adaptivity for branch prediction. In: 25th International Symposium on Computer Architecture, pp. 155–166 (June 1998)
Michaud, P.: A PPM-like, Tag-based Predictor. Journal of Instruction-Level Parallelism 7 (2005)
Farcy, A., Temam, O., Espasa, R., Juan, T.: Dataflow analysis of branch mispredictions and its application to early resolution of branch outcomes. In: 31st International Symposium on Microarchitecture, pp. 59–68 (December 1998)
Thomas, R., Franklin, M.: Using Dataflow Based Context for Accurate Value Prediction. In: 2001 International Conference on Parallel Architectures and Compilation Techniques, pp. 107–117 (September 2001)
Sazeides, Y.: Dependence Based Value Prediction. Technical Report CS-TR-02-00, University of Cyprus (February 2002)
Constantinidis, K., Sazeides, Y.: A Hardware Based Method for Dynamically Detecting Instruction Isomorphism and its Application to Branch Prediction. In: 2nd Value Prediction Workshop (2004)
Chen, L., Dropsho, S., Albonesi, D.H.: Dynamic Data Dependence Tracking and its Application to Branch Prediction. In: 9th International Symposium on High Performance Computer Architecture, pp. 65–76 (February 2003)
Mahlke, S., Natarajan, B.: Compiler Synthesized Dynamic Branch Prediction. In: 29th International Symposium on Microarchitecture, pp. 153–164 (December 1996)
Jacobson, Q., Rottenberg, E., Smith, J.E.: Path-Based Next Trace Prediction. In: 30th International Symposium on Microarchitecture, pp. 14–23 (December 1997)
Gao, F., Sair, S.: Exploiting Intra-function Correlation with the Global History. In: Hämäläinen, T.D., Pimentel, A.D., Takala, J., Vassiliadis, S. (eds.) SAMOS 2005. LNCS, vol. 3553, Springer, Heidelberg (2005)
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 2008 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Sazeides, Y., Moustakas, A., Constantinides, K., Kleanthous, M. (2008). The Significance of Affectors and Affectees Correlations for Branch Prediction. In: Stenström, P., Dubois, M., Katevenis, M., Gupta, R., Ungerer, T. (eds) High Performance Embedded Architectures and Compilers. HiPEAC 2008. Lecture Notes in Computer Science, vol 4917. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-77560-7_17
Download citation
DOI: https://doi.org/10.1007/978-3-540-77560-7_17
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-77559-1
Online ISBN: 978-3-540-77560-7
eBook Packages: Computer ScienceComputer Science (R0)