perfect compile options

atark at thepipeline.net atark at thepipeline.net
Tue Jan 9 18:04:41 PST 2001


Getting rid of the optimizations is probably a good idea.  Does is
really matter that ls runs 5ms faster? IMO, optimizations really only
matter most in certain cases.  When you have many repetitive actions ie.
stuff like distributed.net clients, graphics rendering, or gzip/bzip. 
In those case using certain optimizations or even pgcc can make a big
difference. Search around the net and read about it, then optimize
specific programs or libraries if you have a need. 

Look at it this way, why would a programmer optimize an GUI or
something?  They go after high usage/number crunching areas of the
program, not a user interface.

Of course, I have not done any benchmarks, so take it with a grain of
salt if you like, but I am a programmer. :-)

--Andy


Gerard Beekmans wrote:
> 
> On January  9, 2001 12:06 pm, Jan Stifter wrote:
> > does anybody know, what compile options are perfect for a pentium-pro
> > running at 200 Mhz?
> >
> > i want to compile the things only for this system and want to speed
> > the tools up as much as i can.
> >
> > the book mentions
> >
> > CFLAGS="-O3 -march=xxx"
> > CXXFLAGS=$CFLAGS
> >
> > this seems not very optimized for me.
> > many thanks
> 
> And it isn't no. Actually the next book release doens't include these
> optimizations anymore. It often causes more grief than worth, and even with
> optimization I never found any performance increase, so I stopped using it
> (then again others may notice an increase, it highly depends on the used
> hardware).
> 
> You'd have to read the gcc info page to find out about all the options you
> can pass to GCC and what they mean. Based on those descriptions you decide
> whether you want to use it or not.
> 
> --
> Gerard Beekmans
> www.linuxfromscratch.org
> 
> -*- If Linux doesn't have the solution, you have the wrong problem -*-
> 
> --
> Unsubscribe: send email to lfs-discuss-request at linuxfromscratch.org
> and put unsubscribe in the subject header of the message

-- 
Unsubscribe: send email to lfs-discuss-request at linuxfromscratch.org
and put unsubscribe in the subject header of the message




More information about the lfs-dev mailing list