------- Comment #20 from hubicka at gcc dot gnu dot org 2008-09-26 21:12
-------
The testcase now compiles at -O2 in resonable time.
Checking enabled compiler:
CFG verifier : 15.02 ( 5%) usr 0.01 ( 0%) sys 15.13 ( 5%) wall
0 kB ( 0%) ggc
df reaching defs : 7.57 ( 2%) usr 0.27 ( 7%) sys 7.81 ( 2%) wall
0 kB ( 0%) ggc
df live regs : 13.13 ( 4%) usr 0.00 ( 0%) sys 13.16 ( 4%) wall
0 kB ( 0%) ggc
df live&initialized regs: 3.62 ( 1%) usr 0.00 ( 0%) sys 3.60 ( 1%) wall
0 kB ( 0%) ggc
df use-def / def-use chains: 3.83 ( 1%) usr 0.03 ( 1%) sys 3.85 ( 1%)
wall 0 kB ( 0%) ggc
df reg dead/unused notes: 8.56 ( 3%) usr 0.01 ( 0%) sys 8.53 ( 3%) wall
23492 kB ( 3%) ggc
tree SSA verifier : 22.27 ( 7%) usr 0.04 ( 1%) sys 22.39 ( 7%) wall
0 kB ( 0%) ggc
tree STMT verifier : 27.32 ( 9%) usr 0.42 (11%) sys 27.65 ( 9%) wall
0 kB ( 0%) ggc
expand : 22.57 ( 7%) usr 0.10 ( 3%) sys 22.74 ( 7%) wall
142836 kB (16%) ggc
integrated RA : 72.71 (24%) usr 0.00 ( 0%) sys 74.09 (23%) wall
5404 kB ( 1%) ggc
reload : 17.30 ( 6%) usr 0.00 ( 0%) sys 20.38 ( 6%) wall
26179 kB ( 3%) ggc
TOTAL : 306.65 3.67 316.44
908200 kB
So we do resonable job here. I wonder where those 7% of expansion time goes
to.
We still need over 2GB RAM because of dataflow info. I am going to test how
much difference Kenny's patch make.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37448
