GCC Optimization

Randy McMurchy lfs-user at mcmurchy.com
Sat Feb 10 11:23:45 PST 2007


Athena P wrote these words on 02/10/07 13:18 CST:

> Surely specifying the CPU architecture is still worth while. For example
> using this string must give speed improvements.
> 
>  "-O3 -march=prescott -march=prescott -mtune=prescott -mmmx -msse -
> msse2 -msse3 -m3dnow -pipe -mfpmath=sse" -fomit-fram-pointer"
> 
> Or am I missing something?

[Please don't top-post your replies]

I've never done optimizing, so I cannot fairly comment. I trust the
developers, and use their defaults.

How much "speed" to you expect to gain in percent? Let's say you
get 5% (seems *extremely* high to me that you could achieve that
much). So a process takes 5 minutes (300 seconds), you would save
15 seconds. Substantial. But then 3 packages down the road, you
have subtle build issues, or strange problems running some
application, is the 5% worth it?

If the answer is "yes" for you, then drive on. To me, it is a risk
not worth taking on my production systems.

-- 
Randy

rmlscsi: [bogomips 1003.26] [GNU ld version 2.16.1] [gcc (GCC) 4.0.3]
[GNU C Library stable release version 2.3.6] [Linux 2.6.14.3 i686]
13:23:00 up 31 days, 13:37, 1 user, load average: 0.43, 0.27, 0.20



More information about the lfs-support mailing list