Title
Adaptive Information Processing: An Effective Way To Improve Perceptron Branch Predictors
Abstract
Perceptron branch predictors achieve high prediction accuracy by capturing correlation from very long histories. The required hardware, however, limits the history length to be explored practically. In this paper, an important observation is made that the perceptron weights can be used to estimate the strength of branch correlation. Based such an estimate, adaptive schemes are proposed to preprocess history information so that the input vector to a perceptron predictor contains only those history bits with the strongest correlation. In this way, a much larger history-information set can be explored effectively without increasing the size of perceptron predictors. For the distributed Championship Branch Prediction (CBP-1) traces, our proposed scheme achieves a 47% improvement over a g-share predictor of the same size. For SPEC2000 benchmarks, our proposed scheme outperforms the g-share predictor by 35% on average.
Publication Date
4-1-2005
Publication Title
Journal of Instruction-Level Parallelism
Volume
7
Document Type
Article
Personal Identifier
scopus
Copyright Status
Unknown
Socpus ID
24144492336 (Scopus)
Source API URL
https://api.elsevier.com/content/abstract/scopus_id/24144492336
STARS Citation
Gao, Hongliang and Zhou, Huiyang, "Adaptive Information Processing: An Effective Way To Improve Perceptron Branch Predictors" (2005). Scopus Export 2000s. 4029.
https://stars.library.ucf.edu/scopus2000/4029