On Thu, Oct 22, 2009 at 11:06:32PM +0200, Thomas Bächler wrote:

Sascha Siegel schrieb:

...

Hi,

can someone tell my whats the reason for building the arch-kernel with "# CONFIG_CC_OPTIMIZE_FOR_SIZE is not set"?

Thank you!

Optimizing for size sacrifices performance. Read the gcc documentation about the -O{1,2,3,s} options.

I don't know if it is as simple as that. I recall reading somewhere that under certain circumstances a binary optimized with -Os is faster than a binary optimized with -O2. The reason for this is that a smaller binary may load faster than a big one and cause less page faults.

On Thu, Oct 22, 2009 at 06:15:45PM -0500, Aaron Griffin wrote:

On Thu, Oct 22, 2009 at 5:54 PM, André Ramaciotti da Silva wrote:

...

On Thu, Oct 22, 2009 at 11:06:32PM +0200, Thomas Bächler wrote:

...

Sascha Siegel schrieb:

...

Hi,

can someone tell my whats the reason for building the arch-kernel with "# CONFIG_CC_OPTIMIZE_FOR_SIZE is not set"?

Thank you!

Optimizing for size sacrifices performance. Read the gcc documentation about the -O{1,2,3,s} options.

I don't know if it is as simple as that. I recall reading somewhere that under certain circumstances a binary optimized with -Os is faster than a binary optimized with -O2.

The reason for this is that a smaller binary may load faster than a big one and cause less page faults.

I really doubt the kernel is even close to the boundary for something like this

Agreed, especially as the kernel is loaded only once. I just went a little bit off-topic while still on-topic. :)

Sascha Siegel schrieb:

...

Hi,

can someone tell my whats the reason for building the arch-kernel with "# CONFIG_CC_OPTIMIZE_FOR_SIZE is not set"?

Thank you!

Optimizing for size sacrifices performance. Read the gcc documentation about the -O{1,2,3,s} options.

I remember a few kernel devs recommending size. It was to reduce the number of cache misses and improve CPU cache utilisation, avoiding RAM and L2 cache. Fedora/RHEL appears to optimize for size. That was a while ago when I was doing some kernel stuff.... so I could be wrong.