------- Comment #58 from sebastian dot pop at cri dot ensmp dot fr 2005-11-03 19:31 ------- Subject: Re: [4.0/4.1 Regression] IV-OPTS is O(N^3)
Here are again the numbers for mainline with no other patch: time ./gcc/cc1 -O2 ~/ex/pr18595_10.c real 0m0.164s user 0m0.116s sys 0m0.018s time ./gcc/cc1 -O2 ~/ex/pr18595_100.c tree PRE : 0.33 ( 3%) usr 0.01 ( 3%) sys 0.34 ( 3%) wall 7742 kB ( 2%) ggc tree loop bounds : 0.77 ( 7%) usr 0.03 (10%) sys 0.80 ( 7%) wall 53463 kB (17%) ggc tree iv optimization : 7.70 (74%) usr 0.19 (66%) sys 7.97 (73%) wall 220347 kB (71%) ggc real 0m10.912s user 0m10.448s sys 0m0.317s For 200 nested loops it begins to swap even with -fno-tree-pre, so we'd need a patch for limiting this. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18595
