This is the mail archive of the 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: [www] Document C++ -O0 improvements

On Fri, 29 Oct 2004, Giovanni Bajo wrote:
<li>When compiling without optimizations (-O0), GCC will be faster
than any previous version.  This is especially true for the C++
front-end, for which independent testers have measured speed-ups up
to 25% in real-world production code, compared to the 3.4 family
(which was <strong>already</strong> the fastest version to date).
<emph>...</emph> would probably be more appropriate in this context.

Sure. Gerald, OK with that change?


Paolo: notice that 2.95 is still faster than 3.4 for most C code (see the usual
OpenBSD rants), so I would like to get this paragraph in independently. When
you come back with numbers for C, we can decide what to do.

...though, given the quote above, isn't the statement a bit too strong ("any previous version")?

FWIW, I did some benchmarking of my own yesterday, which is way it took
me some time to respond, and found that 2.95 and 4.0 are about the same
speed for the full application behind PR8361 (C++, at -O0).


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