I've been hearing that LFS is not very well suited for 64-bit architectures and the CLFS is superior in that aspect. I'm designing a system that is to be as fast as possible (it will be doing CGI rendering, so I want to give it as much computing power as possible). I've completed BLFS successfully, but unoptimized, which impacts performance quite a lot.

Should I switch to CLFS (I am building this on each computer individually, so there should be no worries about cross-compatibility)? Or should I stick with ol' LFS (too bad I compiled an ALFS build with the wrong -march option...).

Thanks!

Hi,

LFS is 32 bit, CLFS is either true 64 or a multilib system. BLFS is a 32 bit build and there is, too my knowledge, no BCLFS. (see post #4)

About setting the -march= flag: If you have gcc 4.2 or better it is best to set this flag to native (-march=native) instead of the specific cpu (I.e: CFLAGS="-march=prescott -O2 -pipe -fomit-frame-pointer").

When using -march=native, gcc 4.2 and above automatically detects the optimum settings to use based on your processor. -march=native also detects and applies additional settings beyond -march specific to your CPU.

BTW: Not all the packages that you are building like it when you set CFLAGS / CXXFLAGS. Some are mention specifically. Be careful and do check to see if all went as planned.

Hope this helps.

So I should not use LFS for a system that needs to be as fast as possible? If there is no BCLFS, then I can't really put X on my system, can I...

Thanks!

Hi,

After a little digging I found this: CBLFS Wiki

Guess there is a CBLFS, need to adjust my knowledge