>> 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.
not page faults but the issue is about the CPU L1/L2 cache strain.
BTW,
the help entry says
| CONFIG_CC_OPTIMIZE_FOR_SIZE:
| Enabling this option will pass "-Os" instead of "-O2" to gcc
| resulting in a smaller kernel.
| If unsure, say Y.
alsi Ingo Molnar says here http://lkml.org/lkml/2007/9/13/88 :
| also, if you want to maximize performance, it usually makes more sense
| to build with these flipped around:
|
| # CONFIG_CC_OPTIMIZE_FOR_SIZE is not set
| CONFIG_FORCED_INLINING=y
|
| i.e.:
|
| CONFIG_CC_OPTIMIZE_FOR_SIZE=y
| # CONFIG_FORCED_INLINING is not set
|
| because especially on modern x86 CPUs, smaller x86 code is faster. (and
| it also takes up less I-cache size)
--
damjan
