This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Extreme PRE compile times
- From: Daniel Berlin <dberlin at dberlin dot org>
- To: law at redhat dot com
- Cc: gcc at gcc dot gnu dot org
- Date: Thu, 13 Nov 2003 11:41:01 -0500 (EST)
- Subject: Re: Extreme PRE compile times
- References: <200311131628.hADGSbnu019737@speedy.slc.redhat.com>
On Thu, 13 Nov 2003 law@redhat.com wrote:
> In message <Pine.LNX.4.56.0311131112230.29117@dberlin.org>, Daniel Berlin write
> s:
> >> Clearly PRE is still doing something that's not terribly intelligent given
> >> that PRE + the increase the GC time accounts for 90% of the total
> >> compilation time for this file.
> >
> >As I said, it's being worked on.
> >
> >This also looks like it was done with checking on.
> >Please turn checking off and try again.
> Err, that's not a solution.
There is no way i can control the amount of garbage collection that
occurs with checking on.
If other passes garbage collected, as they should (but don't), they would
have a similar problem in this case.
I can only speed up PRE, not the GC.
If you really want me to reduce the GC times, i'll simply
make it collect after every 5th or 10th expression.
The amount of garbage it generates for 16000 blocks isn't very high
right now (~5 meg), it's just large enough to make collection happen with
the checking heuristics (expansion of >4096k).
> This needs to work in a reasonable manner
> even with checking enabled.
> It's not acceptable for PRE to blow up
> compile times like that. If it wasn't for PRE that file would compile
> in 2 minutes, with PRE it takes 15.
I can reduce the PRE times, as I said.
I can't change the GC times without collecting less (there is no other
option here).
>
> Jeff
>
>