https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65076

--- Comment #63 from Markus Trippelsdorf <trippels at gcc dot gnu.org> ---
I think we can close this bug. gcc-6 now compiles faster than gcc-5
and the resulting binary also runs slightly faster.


CXXFLAGS="-std=c++14 -w -Ofast" for all compilers 

gcc-4.9: 24.199 compile time

gcc-5 :  27.311 compile time

 Performance counter stats for './a.out --cartvis 1.0 0.0 --rhomin 1e-8 -n 20'
(5 runs):

       3200.800294      task-clock (msec)         #    1.000 CPUs utilized     
      ( +-  0.05% )  (99.99%)
                 7      context-switches          #    0.002 K/sec             
      ( +- 25.56% )  (99.99%)
                 0      cpu-migrations            #    0.000 K/sec             
      (99.99%)
            17,310      page-faults               #    0.005 M/sec             
      ( +-  0.01% )  (99.99%)
    10,252,877,203      cycles                    #    3.203 GHz               
      ( +-  0.05% )  (66.63%)
       185,288,587      stalled-cycles-frontend   #    1.81% frontend cycles
idle     ( +-  0.51% )  (66.65%)
     6,775,673,735      stalled-cycles-backend    #   66.09% backend  cycles
idle     ( +-  0.07% )  (66.68%)
     8,661,344,185      instructions              #    0.84  insns per cycle    
                                                  #    0.78  stalled cycles per
insn  ( +-  0.10% )  (66.72%)
       660,295,960      branches                  #  206.291 M/sec             
      ( +-  0.45% )  (66.75%)
        10,126,045      branch-misses             #    1.53% of all branches   
      ( +-  0.35% )  (66.68%)

       3.201031138 seconds time elapsed                                        
 ( +-  0.05% )

gcc-6 :  26.258 compile time

 Performance counter stats for './a.out --cartvis 1.0 0.0 --rhomin 1e-8 -n 20'
(5 runs):

       3134.574551      task-clock (msec)         #    1.000 CPUs utilized     
      ( +-  0.13% )  (99.99%)
               111      context-switches          #    0.035 K/sec             
      ( +- 58.99% )  (99.99%)
                 0      cpu-migrations            #    0.000 K/sec             
      (99.99%)
            17,308      page-faults               #    0.006 M/sec             
      ( +-  0.00% )  (99.99%)
    10,035,141,160      cycles                    #    3.201 GHz               
      ( +-  0.12% )  (66.65%)
       173,085,246      stalled-cycles-frontend   #    1.72% frontend cycles
idle     ( +-  2.71% )  (66.65%)
     6,752,556,601      stalled-cycles-backend    #   67.29% backend  cycles
idle     ( +-  0.15% )  (66.66%)
     8,195,763,733      instructions              #    0.82  insns per cycle    
                                                  #    0.82  stalled cycles per
insn  ( +-  0.19% )  (66.74%)
       635,509,854      branches                  #  202.742 M/sec             
      ( +-  0.46% )  (66.75%)
         9,900,182      branch-misses             #    1.56% of all branches   
      ( +-  0.69% )  (66.67%)

       3.135833601 seconds time elapsed                                        
 ( +-  0.15% )

clang:   39.317 compile time

 Performance counter stats for './a.out --cartvis 1.0 0.0 --rhomin 1e-8 -n 20'
(5 runs):

       6026.232154      task-clock (msec)         #    1.000 CPUs utilized     
      ( +-  0.03% )  (99.99%)
               129      context-switches          #    0.021 K/sec             
      ( +- 91.87% )  (99.99%)
                 0      cpu-migrations            #    0.000 K/sec             
      ( +-100.00% )  (99.99%)
            17,314      page-faults               #    0.003 M/sec             
      ( +-  0.00% )  (99.99%)
    19,304,680,254      cycles                    #    3.203 GHz               
      ( +-  0.04% )  (66.64%)
       190,735,168      stalled-cycles-frontend   #    0.99% frontend cycles
idle     ( +-  3.62% )  (66.64%)
    10,488,375,287      stalled-cycles-backend    #   54.33% backend  cycles
idle     ( +-  0.08% )  (66.69%)
    19,895,781,820      instructions              #    1.03  insns per cycle    
                                                  #    0.53  stalled cycles per
insn  ( +-  0.08% )  (66.70%)
       945,624,800      branches                  #  156.918 M/sec             
      ( +-  0.37% )  (66.69%)
        13,241,690      branch-misses             #    1.40% of all branches   
      ( +-  0.42% )  (66.67%)

       6.027538343 seconds time elapsed                                        
 ( +-  0.04% )

Reply via email to