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: Bootstrap failures


> >Hi,
> >looks like your patch further increase memory consumption.  Overall
> >footprint of -O1 insn-attrtab compilation is now 15% up compared to day
> >before yesterday.
> 
> This is due to having more vops at store and load sites (basically one
> more per load and store and call).
> This is because all insn-attrtab does is access global memory, and
> call getter funcs so basically every load and store and call is a
> touch of the nonlocal vop.
> 
> Sadly, *this* vop is  hard to avoid, because we were getting incorrect
> results before.  The type-specific patch usually avoided them by TBAA,
> but we aren't doing that anymore.  This is just a case where we have
> to use an extra vop to be correct in all cases, even though it doesn't
> matter in 99% of cases right now.

Yep, I am aware of that and to my shame, I noticed the problem some time
ago while playing around with the code but managed to convince myself
that somehow this has to be solved elsewhere...

I am attaching the full memory dumps before and after your patches in
case something obvious can be noticed in them, but there is probably not
much - as expected the number of SSA names and PHI nodes roughly double.

The scheduling issue is probably worth to look at...

Honza

Attachment: today
Description: Text document

Attachment: yesterday
Description: Text document

Attachment: day_before_yesterday
Description: Text document


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