[asterisk-bugs] [Zaptel 0011283]: [patch] Blackfin optimizations in arith.h

noreply at bugs.digium.com noreply at bugs.digium.com
Mon Nov 19 01:30:19 CST 2007


A NOTE has been added to this issue. 
====================================================================== 
http://bugs.digium.com/view.php?id=11283 
====================================================================== 
Reported By:                otterson
Assigned To:                
====================================================================== 
Project:                    Zaptel
Issue ID:                   11283
Category:                   General
Reproducibility:            always
Severity:                   tweak
Priority:                   normal
Status:                     new
Zaptel Version:              SVN 
SVN Branch (only for SVN checkouts, not tarball releases): N/A  
SVN Revision (number only!): 3239 
Disclaimer on File?:        N/A 
Request Review:              
====================================================================== 
Date Submitted:             11-18-2007 10:39 CST
Last Modified:              11-19-2007 01:30 CST
====================================================================== 
Summary:                    [patch] Blackfin optimizations in arith.h
Description: 
This is an optimization tweak for zaptel's arith.h.  The purpose of the
tweak is to improve the performance of zaptel conference mixer on the
Blackfin/uclinux platform.  This tweak should have no effect on any other
platforms.

The tweak replaces the ACSS and SCSS saturating add and subtract code with
a blackfin GCC builtin function that uses the Blackfin's internal
saturating add and subtract instructions.  This eliminates one or two
conditional branches for every sample added to the mix.
====================================================================== 

---------------------------------------------------------------------- 
 tzafrir - 11-19-07 01:30  
---------------------------------------------------------------------- 
Where exactly is BFIN defined?

Maybe it would be safer to use CONFIG_BLACKFIN (assuming that this is
kernel code)? 

Issue History 
Date Modified   Username       Field                    Change               
====================================================================== 
11-19-07 01:30  tzafrir        Note Added: 0073893                          
======================================================================




More information about the asterisk-bugs mailing list