This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

RE: branch probabilities on multiway branches


The calculate branch probabilities algorithm (1) in the Wu Larus paper
also evenly distributes branch probabilities when number of outgoing
edges is > 2, e.g. switch cases implemented as jump tables.

Are they any known heuristics to generate better branch probabilities
in this case?


-----Original Message-----
From: Jan Hubicka [mailto:hubicka@ucw.cz] 
Sent: 14 April 2010 00:44
To: Rahul Kharche
Cc: gcc@gcc.gnu.org; sdkteam-gnu
Subject: Re: branch probabilities on multiway branches

> Hi All,
> 
> The following bit of code in predict.c implies branch probabilities
> are strictly evenly distributed for multiway branches at present. The
> comment suggests it is possible to generate better estimates for more
> generic cases, apart from being involved. Could anyone point me to
> the reference and/or if an implementation exists already.

There is Wu & Larus paper cited in the comment at the begging of file.

Honza


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]