Loading...
Loading...

Go to the content (press return)

Confidence estimation using data value correlation for branch prediction reversal

Author
Aragon, J.; Gonzalez, A.
Type of activity
Report
Date
2001-02
Code
UPC-DAC-2001-4
Abstract
Branch misprediction penalties are one of the major hurdles in the roadmap towards higher clock frequencies, deeper pipelines and bigger instruction windows in superscalar processors. Numerous proposals attempt to reduce misprediction rates mainly by means of reducing the effects of destructive aliasing in the prediction tables. Recently, branch prediction reversal has been proved to be an effective alternative approach to dropping misprediction rates by means of adding a Confidence Estimator to...
Group of research
ARCO - Microarchitecture and Compilers

Participants