Re: [Boost-bugs] [Boost C++ Libraries] #5735: proto should force functions to be inline

Subject: Re: [Boost-bugs] [Boost C++ Libraries] #5735: proto should force functions to be inline
From: Boost C++ Libraries (noreply_at_[hidden])
Date: 2011-07-26 21:57:35

#5735: proto should force functions to be inline
-------------------------------+--------------------------------------------
  Reporter:  mgaunard          |       Owner:  eric_niebler
      Type:  Bugs              |      Status:  closed      
 Milestone:  To Be Determined  |   Component:  proto       
   Version:  Boost 1.47.0      |    Severity:  Problem     
Resolution:  invalid           |    Keywords:              
-------------------------------+--------------------------------------------

Comment (by eric_niebler):

 Wow, the paper is from this year, and they think that Blitz++ and uBLAS
 represent the bleeding edge of ET technology. That's laughable.

 But the perf measurements are interesting. I can't see where they say what
 compiler they're using for their perf benchmarks, can you? If they're
 using that buggy GCC compler version, then yes, it will grossly screw up
 the results. They talk about cache issues, but only consider data size and
 not code size.

 Joel Falcou is the one closest to this issue. He's filed specific bugs
 against Proto (function X should be changed in way Y for better inlining),
 which I have addressed promptly. He hasn't noticed any problems that would
 require any across-the-board change like the one you're suggesting. Or if
 he has, he hasn't said anything about it. If you can show me a specific
 case where a (non-buggy, mainstream) compiler generates slow Proto code
 because of failed inlining, I'd be more easily convinced.

 But you should bring this up on the Proto list and make sure Joel sees it.
 I'd be curious to hear his thoughts.

-- 
Ticket URL: <https://svn.boost.org/trac/boost/ticket/5735#comment:3>
Boost C++ Libraries <http://www.boost.org/>
Boost provides free peer-reviewed portable C++ source libraries.