[Bug middle-end/39326] Segmentation fault with -O1, out of memory with -O2

["rguenther at suse dot de" <gcc-bugzilla at gcc dot gnu dot org>]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=39326

--- Comment #28 from rguenther at suse dot de <rguenther at suse dot de> 2013-03-07 08:44:28 UTC ---
On Wed, 6 Mar 2013, steven at gcc dot gnu.org wrote:

> 
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=39326
> 
> Steven Bosscher <steven at gcc dot gnu.org> changed:
> 
>            What    |Removed                     |Added
> ----------------------------------------------------------------------------
>                  CC|                            |steven at gcc dot gnu.org
> 
> --- Comment #24 from Steven Bosscher <steven at gcc dot gnu.org> 2013-03-06 23:39:27 UTC ---
> (In reply to comment #22)
> > 4.8.0  -O2  (terminated after 9 minutes waiting, LIM being the offender, I
> > suspect domwalk ...)  >2.5GB
> > 
> > applying domwalk fix ...
> 
> It is LIM, for sure. I've been watching in GDB for a while at some
> back traces, and it's spent minutes already in this DOM walk:
> 
> #5  0x0000000000b841e1 in walk_dominator_tree (walk_data=0x7fffffffdd60,
> bb=0x7fffef033958) at ../../trunk/gcc/domwalk.c:187
> #6  0x0000000000c02d73 in determine_invariantness () at
> ../../trunk/gcc/tree-ssa-loop-im.c:1189
> #7  tree_ssa_lim () at ../../trunk/gcc/tree-ssa-loop-im.c:2632
> #8  0x000000000075bcd7 in execute_one_pass (pass=0x12323e0 <pass_lim>) at
> ../../trunk/gcc/passes.c:2330
> #9  0x000000000075c0f5 in execute_pass_list (pass=0x12323e0 <pass_lim>) at
> ../../trunk/gcc/passes.c:2378
> 
> This is supposed to be cheap. Is this a known bottle-neck?
> 
> Pathetic...

Yes, it's known - and there are several known (to me ...) ways to
make constant factor compile-time and memory-usage improvements...

(I _think_ we have a bug for LIMs slowness, if you can't find it
quickly you can create one and assign me - I have some TLC patches
locally queued for 4.9, but they don't help the slowness very much)