http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43396
Sebastian Pop changed:
What|Removed |Added
Status|ASSIGNED|RESOLVED
Resolution|---
--- Comment #4 from spop at gcc dot gnu dot org 2010-03-16 21:38 ---
I think it is not due to the number of parameters: as 3 is pretty low.
It might be the code that we generate for all the cases.
I will give it a look.
--
spop at gcc dot gnu dot org changed:
What|Rem
--- Comment #3 from zsojka at seznam dot cz 2010-03-16 21:32 ---
With various values of --param graphite-max-nb-scop-params=n, compile time
changes as follows:
n=1..3 "zero" time (<0.1s)
n=4..10 ~8.2s
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43396
--- Comment #2 from spop at gcc dot gnu dot org 2010-03-16 21:23 ---
You may want to lower the --param graphite-max-nb-scop-params=5
the current value is at 10. Does this help?
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43396
--- Comment #1 from zsojka at seznam dot cz 2010-03-16 20:39 ---
Created an attachment (id=20119)
--> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=20119&action=view)
reduced testcase (from testsuite/gcc.c-torture/execute/builtins/memcpy-chk.c)
gcc -O1 -fgraphite-identity -c pr43396.c
