This is the mail archive of the gcc-patches@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]
Other format: [Raw text]

Re: [PATCH][C++] Improve PR15855 (compile-time/memory hog)


On Fri, 2005-09-23 at 16:12 -0400, Daniel Berlin wrote:
> >> I sincerely doubt we actually have 5 million SFT's.  I'd be *incredibly*
> >> surprised.
> > We have 5 million _mappings_ for SFT variables during an incremental
> > SSA update.  There's something like 20k SFT variables IIRC.
> 
> >
> > If you don't believe me, then take a looksie :-)  Maybe I'm counting
> > wrong :-)
> >
> 
> No, you aren't, but actually, it shouldn't matter.
> 
> > The first big incremental update for
> > static_initialization_and_destruction:
> >
> >
> > Number of virtual NEW -> OLD mappings:   43504
> > Number of real NEW -> OLD mappings:          0
> > Number of total NEW -> OLD mappings:     43504
> >
> > Number of virtual symbols: 21950
> 
> Which seems fine.
Probably within reason.

> > The second big incremental update has:
> >
> > Number of virtual NEW -> OLD mappings: 5443467
> > Number of real NEW -> OLD mappings:          0
> > Number of total NEW -> OLD mappings:   5443467
> >
> > Number of virtual symbols: 21950
> >
> 
> Uh, why did we need 5 million mappings here?
I haven't a clue.  Whatever's going on is certainly excessive,
that's about all I know.  I suspect whomever figures this out
will get a nice little speedup on 15855 :-)


Jeff


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