This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: 3.0 high priority: optimization/3083: C++ frontend consumesinacceptable amounts of CPU with -O3
Mark Mitchell wrote:
>
> >>>>> "Gerald" == Gerald Pfeifer <pfeifer@dbai.tuwien.ac.at> writes:
>
> >> the -O3 option was broken?? (i.e. -O3 compilations were faster
> >> than they really would have been if -O3 was really working...)
>
> I consider -O3 to be essentially worthless in its current state. I
> never use it my own code.
>
> I know there are programs that will run faster with -O3 -- but there
> are lots of problems that will compile slower and then run slower. We
> should really improve the -O3 inlining heuristics; our current
> approach is too naive.
Mark, FYI, while you were on holiday this popped up, and I wrote
http://gcc.gnu.org/ml/gcc/2001-05/msg01651.html
IMO, it's far too late to go poking around in that heuristic to get
-O3 sensible for 3.0
nathan
--
Dr Nathan Sidwell :: http://www.codesourcery.com :: CodeSourcery LLC
'But that's a lie.' - 'Yes it is. What's your point?'
nathan@codesourcery.com : http://www.cs.bris.ac.uk/~nathan/ : nathan@acm.org