Sciweavers

Share
JILP
2000

Speculative Updates of Local and Global Branch History: A Quantitative Analysis

8 years 5 months ago
Speculative Updates of Local and Global Branch History: A Quantitative Analysis
In today's wide-issue processors, even small branch-misprediction rates introduce substantial performance penalties. Worse yet, inadequate branch prediction creates a bottleneck at the fetch stage, restricting other opportunities for improving performance. The choice of how to predict conditional-branch outcomes is the primary lever on prediction accuracy. But the choice of when to update the predictor with branch outcomes is a second powerful lever, and the subject of this paper. In history-based predictors like gshare, many mispredictions result from commit-time update of the history: typical pipelined processors predict branches in the fetch stage, but update the predictor in the commit stage, making the predictor's state temporarily outof-date. As pipelines grow longer--in particular, when branches can spend many cycles in the instruction window waiting to issue--this problem becomes worse. Prior work on this subject has discussed the need for speculative update in a glo...
Kevin Skadron, Margaret Martonosi, Douglas W. Clar
Added 19 Dec 2010
Updated 19 Dec 2010
Type Journal
Year 2000
Where JILP
Authors Kevin Skadron, Margaret Martonosi, Douglas W. Clark
Comments (0)
books