[Bug tree-optimization/72712] [7 Regression] Tenfold compile time regression

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=72712

--- Comment #5 from Richard Biener <rguenth at gcc dot gnu.org> ---
Now that we have another (early) pass things should be even worse.  Current
trunk at -O2 with -fno-checking improved though (I guess by means of a cost
model or so):

 phase opt and generate  :  69.39 (99%) usr   0.98 (98%) sys  72.38 (99%) wall 
289872 kB (97%) ggc
 tree SSA incremental    :  16.94 (24%) usr   0.00 ( 0%) sys  17.36 (24%) wall 
 13582 kB ( 5%) ggc
 backwards jump threading:   5.04 ( 7%) usr   0.00 ( 0%) sys   5.20 ( 7%) wall 
115843 kB (39%) ggc
 PRE                     :  18.37 (26%) usr   0.25 (25%) sys  19.13 (26%) wall 
   270 kB ( 0%) ggc