This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: criteria.html open issues


> I do not understand the philosophy behind only enabling strict aliasing
> at -O3. Generally we often find -O3 to be a disadvantage because of the
> excessive inlining which can cause icache penalties. But strict aliasing
> seems like something that should come along with -O2.

Think about what the -O* levels are supposed to mean.
As the numbers go up:

1. the output gets better
2. the compile takes longer
3. non-standard code breaks

If -O3 is a disadvantage, then -O3 is broken. The excessive inlining
is a bug. Why have -O3 if it generally doesn't work?

Strict aliasing should not come along with -O2 because it breaks
lots of existing code. It looks like something that would make
compiles take longer too.

So -O3 should be changed to add one thing: strict aliasing.
The excessive inlining should be removed.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]