RE: [Patch] Fix istream::ignore + minor things

["Damour, James A" <James dot Damour at dfa dot state dot ny dot us>]

I'm quite inexperienced when it come to the GNU build chain in general
and gcc optimizations in particular, and I'm subscribed to this list as
a learning experience, so please forgive me if I have belabored the
obvious.  

If I understand your position correctly, you recommend against the use
of the __builtin_expect function except when performance data shows a
clear improvement.  Put a different way, you prefer brevity and clarity
in the source over some unproven performance improvement.  Thus, I
believe that your answer to my question is, "unless the performance
improvement is demonstrated to be substantial, the more simple construct
has greater benefit".

Thank you for your clarification, and I apologize for not searching the
mail archives for previous discussions about this built-in function.

P.S. Paolo, I got your email too, but I believe that you were saying
approximately the same thing as Gaby (i.e. "don't use if unless you can
prove you need it").

-----Original Message-----
From: gdr@integrable-solutions.net [mailto:gdr@integrable-solutions.net]

Sent: Tuesday, May 25, 2004 8:29 AM
To: Damour, James A
Cc: Paolo Carlini; libstdc++
Subject: Re: [Patch] Fix istream::ignore + minor things

"Damour, James A" <James.Damour@dfa.state.ny.us> writes:

| Would your ignore code benefit from using __builtin_expect (see 
| example below)?
| 
| -          if (__n != numeric_limits<streamsize>::max())
| +          if (__builtin_expect(__n !=
| numeric_limits<streamsize>::max(),1))

I've asked in the past for data, in supporting the __builtin_expect
cluttering in the codes (not that I'm firmly opposed, but I would need
to see data).  It was claimed that it would be faster with a different
compiler, but the actual experiment with GCC wa unconclusive.  So far.

-- Gaby