[Bug tree-optimization/119389] [15 Regression] tree FRE very slow when dealing with big switch statements

[pheeck at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119389

--- Comment #1 from Filip Kastl <pheeck at gcc dot gnu.org> ---
This is how it looks on r15-6120-g56946c801a7cf3a831a11870b7e11ba08bf9bd87

> gcc bugreport.c -O1 -c -ftime-report
Time variable                                  wall           GGC
 phase setup                        :   0.00 (  0%)  1898k (  2%)
 phase parsing                      :   0.09 (  2%)    13M ( 15%)
 phase opt and generate             :   4.07 ( 98%)    72M ( 83%)
 callgraph construction             :   0.01 (  0%)  4096k (  5%)
 callgraph functions expansion      :   3.15 ( 76%)    60M ( 69%)
 callgraph ipa passes               :   0.88 ( 21%)  2275k (  3%)
 ipa function summary               :   0.01 (  0%)  6712  (  0%)
 ipa modref                         :   0.13 (  3%)   400  (  0%)
 cfg construction                   :   0.01 (  0%)  3204k (  4%)
 cfg cleanup                        :   0.12 (  3%)  2554k (  3%)
 trivially dead code                :   0.02 (  0%)     0  (  0%)
 df scan insns                      :   0.01 (  0%)    48  (  0%)
 df live regs                       :   0.11 (  3%)     0  (  0%)
 df live&initialized regs           :   0.04 (  1%)     0  (  0%)
 df reg dead/unused notes           :   0.08 (  2%)  1347k (  2%)
 register information               :   0.01 (  0%)     0  (  0%)
 alias analysis                     :   0.04 (  1%)  1281k (  1%)
 rebuild jump labels                :   0.01 (  0%)     0  (  0%)
 preprocessing                      :   0.02 (  1%)  4774k (  5%)
 lexical analysis                   :   0.01 (  0%)     0  (  0%)
 parser function body               :   0.05 (  1%)  8457k (  9%)
 inline parameters                  :   0.03 (  1%)  8912  (  0%)
 tree gimplify                      :   0.01 (  0%)  2318k (  3%)
 tree CFG construction              :   0.01 (  0%)  3836k (  4%)
 tree CFG cleanup                   :   0.03 (  1%)     0  (  0%)
 tree copy propagation              :   0.01 (  0%)     0  (  0%)
 tree PTA                           :   0.02 (  0%)   128  (  0%)
 tree SSA rewrite                   :   0.01 (  0%)  1469k (  2%)
 tree operand scan                  :   0.01 (  0%)   949k (  1%)
 dominator optimization             :   1.05 ( 25%)  3067k (  3%)
 tree CCP                           :   0.05 (  1%)    56  (  0%)
 tree FRE                           :   1.01 ( 24%)   496  (  0%)
 tree forward propagate             :   0.02 (  0%)     0  (  0%)
 tree SSA uncprop                   :   0.01 (  0%)     0  (  0%)
 tree switch lowering               :   0.01 (  0%)  7244k (  8%)
 tree modref                        :   0.22 (  5%)   480  (  0%)
 dominance computation              :   0.11 (  3%)     0  (  0%)
 out of ssa                         :   0.01 (  0%)   360  (  0%)
 expand                             :   0.02 (  1%)    14M ( 17%)
 post expand cleanups               :   0.01 (  0%)   480  (  0%)
 forward prop                       :   0.10 (  2%)     0  (  0%)
 CSE                                :   0.04 (  1%)     0  (  0%)
 dead code elimination              :   0.02 (  1%)     0  (  0%)
 dead store elim1                   :   0.01 (  0%)   559k (  1%)
 dead store elim2                   :   0.02 (  0%)   468k (  1%)
 loop init                          :   0.02 (  1%)  4552  (  0%)
 branch prediction                  :   0.01 (  0%)   104  (  0%)
 combiner                           :   0.03 (  1%)   325k (  0%)
 if-conversion                      :   0.09 (  2%)  8771k ( 10%)
 integrated RA                      :   0.10 (  2%)  7146k (  8%)
 LRA non-specific                   :   0.05 (  1%)    56  (  0%)
 LRA virtuals elimination           :   0.01 (  0%)     0  (  0%)
 LRA reload inheritance             :   0.01 (  0%)     0  (  0%)
 LRA create live ranges             :   0.02 (  0%)     0  (  0%)
 reload CSE regs                    :   0.03 (  1%)   883k (  1%)
 thread pro- & epilogue             :   0.01 (  0%)   880  (  0%)
 if-conversion 2                    :   0.01 (  0%)     0  (  0%)
 combine stack adjustments          :   0.01 (  0%)     0  (  0%)
 hard reg cprop                     :   0.03 (  1%)     0  (  0%)
 reorder blocks                     :   0.03 (  1%)  1082k (  1%)
 shorten branches                   :   0.01 (  0%)     0  (  0%)
 final                              :   0.03 (  1%)  5189k (  6%)
 straight-line strength reduction   :   0.01 (  0%)     0  (  0%)
 initialize rtl                     :   0.01 (  0%)    12k (  0%)
 access analysis                    :   0.01 (  0%)     0  (  0%)
 rest of compilation                :   0.06 (  1%)  1367k (  2%)
 remove unused locals               :   0.01 (  0%)     0  (  0%)
 TOTAL                              :   4.16           87M

I'll try to bisect.