M. Haungs, P. Sallee, and M. Farrens. Branch transition rate: A new metric for improved branch classi - cation analysis. In Proc. HPCA, pages 241-250, 2000. Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
Branch Classication For Smt Fetch Gating - Knijnenburg, A.Ramirez, Larriba, .. (2002) (Correct)
....con dent that the branch is predicted correctly. Otherwise, we stall the thread until the branch is resolved. There are at least two ways to ensure that a prediction will probably be correct. First, we can use a branch classi er to classify branches as easy or hard to predict. It has been shown [4, 5] that many branches are stable during the execution of a program, that is, they have the same direction many times in a row and are therefore easy to predict. The branch classi er records whether a branch is stable. In this case, it is classi ed as easy to predict, and as hard otherwise. We ....
M. Haungs, P. Sallee, and M. Farrens. Branch transition rate: A new metric for improved branch classi - cation analysis. In Proc. HPCA, pages 241-250, 2000.
Branch Classification to Control Instruction Fetch .. - Knijnenburg.. (2002) (Correct)
....is heavily biased: they are either taken or not taken for over 95 of the times they are encountered. This observation has been exploited previously to enhance branch prediction accuracy [1, 12] Also, branches can have a low transition rate, that is, they do not change direction often. In [5] it has been shown that over 60 of the dynamic branches have a transition rate of less than 5 . Hence, branches tend to be stable during the execution of a program and are therefore easy to predict. In order to decide between branch classification and confidence estimation as the mechanism to ....
M. Haungs, P. Sallee, and M. Farrens. Branch transition rate: A new metric for improved branch classification analysis. In Proc. HPCA, pages 241--250, 2000.
Branch Classification for SMT Fetch Gating - Knijnenburg, Ramirez, Larriba.. (Correct)
....is heavily biased: they are either taken or not taken for over 95 of the times they are encountered. This observation has been exploited previously to enhance branch prediction accuracy [1, 12] Also, branches can have a low transition rate, that is, they do not change direction often. In [5] it has been shown that over 60 of the dynamic branches have a transition rate of less than 5 . Hence, many branches are stable during the execution of a program and are therefore easy to predict. The branch classi er records whether branches are stable, that is, they have had the same ....
M. Haungs, P. Sallee, and M. Farrens. Branch transition rate: A new metric for improved branch classi cation analysis. In Proc. HPCA, pages 241-250, 2000.
Branch Classification to Control Instruction Fetch - In Simultaneous Multithreaded (Correct)
....is heavily biased: they are either taken or not taken for over 95 of the times they are encountered. This observation has been exploited previously to enhance branch prediction accuracy [1, 12] Also, branches can have a low transition rate, that is, they do not change direction often. In [5] it has been shown that over 60 of the dynamic branches have a transition rate of less than 5 . Hence, branches tend to be stable during the execution of a program and are therefore easy to predict. In order to decide between branch classification and confidence estimation as the mechanism to ....
M. Haungs, P. Sallee, and M. Farrens. Branch transition rate: A new metric for improved branch classification analysis. In Proc. HPCA, pages 241--250, 2000.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC