LG

Exécution

perf stat ./program

 Performance counter stats for './program':

            856,75 msec task-clock:u              #    0,999 CPUs utilized          
                 0      context-switches:u        #    0,000 K/sec                  
                 0      cpu-migrations:u          #    0,000 K/sec                  
               110      page-faults:u             #    0,128 K/sec                  
     2 995 631 212      cycles:u                  #    3,497 GHz                    
     1 434 273 421      stalled-cycles-frontend:u #   47,88% frontend cycles idle   
     4 802 352 286      instructions:u            #    1,60  insn per cycle         
                                                  #    0,30  stalled cycles per insn
     1 200 437 383      branches:u                # 1401,161 M/sec                  
            32 390      branch-misses:u           #    0,00% of all branches        

       0,857405569 seconds time elapsed

       0,857183000 seconds user
       0,000000000 seconds sys

branch-misses : c'est les échecs des prédictions de branches. Measuring Branch Prediction Accuracy Using Linux Perf Tools Archive de 2013 le 20/02/2020 Why is processing a sorted array faster than processing an unsorted array? Archive du 27/06/2012 le 20/02/2020
stalled-cycles-frontend : le cache est mal utilisé.
stalled-cycles-backend : le CPU est en attente d'une ressource (mémoire, disque, …).