http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50557
--- Comment #7 from William J. Schmidt wschmidt at gcc dot gnu.org 2011-10-10
12:40:01 UTC ---
I don't have anything too helpful to add. This code as it stands is balanced
on a knife's edge for register usage for the particular target, so it's
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50557
--- Comment #6 from Igor Zamyatin izamyatin at gmail dot com 2011-10-07
10:33:33 UTC ---
Indeed, overall register pressure is not increased. Even before IRA dumps show
that register pressure is actually kept on the same level.
Looks like it is
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50557
--- Comment #5 from William J. Schmidt wschmidt at gcc dot gnu.org 2011-09-30
14:30:56 UTC ---
Reassociation isn't doing anything untoward here that raises register pressure.
The problem must be occurring downstream. Likely the scheduler is
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50557
--- Comment #3 from Igor Zamyatin izamyatin at gmail dot com 2011-09-29
08:34:45 UTC ---
William, thanks for quick response!
With -funroll-loops regression is still present.
Do you want me to attach some dumps?
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50557
--- Comment #4 from William J. Schmidt wschmidt at gcc dot gnu.org 2011-09-29
12:16:46 UTC ---
No, that's OK. I should be able to reproduce this on a pool machine.
It may be difficult to come up with a good heuristic here given that
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50557
Richard Guenther rguenth at gcc dot gnu.org changed:
What|Removed |Added
CC||wschmidt at
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50557
--- Comment #1 from Igor Zamyatin izamyatin at gmail dot com 2011-09-28
11:52:18 UTC ---
Created attachment 25373
-- http://gcc.gnu.org/bugzilla/attachment.cgi?id=25373
testcase
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50557
--- Comment #2 from William J. Schmidt wschmidt at gcc dot gnu.org 2011-09-28
12:13:50 UTC ---
The fix for 49749 is intended to remove dependencies between loop iterations.
One possibility would be to condition the changes on the presence of
