--- Comment #18 from steven at gcc dot gnu dot org 2010-01-28 23:22 ---
Function that seems to cost the most time is add_functions(), which is one big
basic block of ~7500 insns (~500 of them call insns).
List scheduling is quadratic in the number of insns per basic block. I don't
know
--- Comment #17 from steven at gcc dot gnu dot org 2010-01-28 23:03 ---
Introduced by http://gcc.gnu.org/ml/gcc-cvs/2009-09/msg00094.html.
Needs 1.3GB and 68s to compile on an 800MHz AMD64 x86_64 X
armv5tejl-unknown-linux-gnueabi compiler. The compiler spends 87% of its time
in the sche
--- Comment #16 from steven at gcc dot gnu dot org 2010-01-28 21:55 ---
Re. comment #9 -- had anyone actually tried ARM code size with sched1 enabled
and the -fsched-pressure flag added?
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=41399
--- Comment #15 from ramana at gcc dot gnu dot org 2009-12-11 09:53 ---
*** Bug 42351 has been marked as a duplicate of this bug. ***
--
ramana at gcc dot gnu dot org changed:
What|Removed |Added
--- Comment #14 from rguenth at gcc dot gnu dot org 2009-12-08 11:54
---
It needs 60MB on i?86-linux with -O2 -g -fschedule-insns. Thus this is a
target specific problem. Is the arm automaton particularly large?
--
rguenth at gcc dot gnu dot org changed:
What|Remov
