$ g++ -O3 a.cpp; sudo perf stat -r 10 ./a.out
cast:      24 ms
memcpy:    24 ms
memmove:   24 ms
std::copy: 47 ms
(counter:  1787289600)

cast:      24 ms
memcpy:    24 ms
memmove:   24 ms
std::copy: 47 ms
(counter:  1787289600)

cast:      24 ms
memcpy:    24 ms
memmove:   24 ms
std::copy: 47 ms
(counter:  1787289600)

cast:      24 ms
memcpy:    24 ms
memmove:   24 ms
std::copy: 47 ms
(counter:  1787289600)

cast:      24 ms
memcpy:    24 ms
memmove:   24 ms
std::copy: 47 ms
(counter:  1787289600)

cast:      25 ms
memcpy:    24 ms
memmove:   24 ms
std::copy: 47 ms
(counter:  1787289600)

cast:      24 ms
memcpy:    24 ms
memmove:   24 ms
std::copy: 47 ms
(counter:  1787289600)

cast:      24 ms
memcpy:    24 ms
memmove:   24 ms
std::copy: 47 ms
(counter:  1787289600)

cast:      29 ms
memcpy:    24 ms
memmove:   24 ms
std::copy: 48 ms
(counter:  1787289600)

cast:      23 ms
memcpy:    24 ms
memmove:   24 ms
std::copy: 47 ms
(counter:  1787289600)


 Performance counter stats for './a.out' (10 runs):

            120.20 msec task-clock                #    0.996 CPUs utilized            ( +-  0.46% )
                 9      context-switches          #    0.079 K/sec                    ( +- 65.80% )
                 0      cpu-migrations            #    0.000 K/sec
               146      page-faults               #    0.001 M/sec                    ( +-  0.18% )
       512,609,946      cycles                    #    4.265 GHz                      ( +-  0.14% )
     1,605,215,244      instructions              #    3.13  insn per cycle           ( +-  0.01% )
       400,813,958      branches                  # 3334.604 M/sec                    ( +-  0.01% )
           105,629      branch-misses             #    0.03% of all branches          ( +-  2.05% )

          0.120648 +- 0.000563 seconds time elapsed  ( +-  0.47% )

$ g++ -Ofast a.cpp; sudo perf stat -r 10 ./a.out
cast:      24 ms
memcpy:    47 ms
memmove:   23 ms
std::copy: 24 ms
(counter:  1787289600)

cast:      24 ms
memcpy:    47 ms
memmove:   24 ms
std::copy: 24 ms
(counter:  1787289600)

cast:      24 ms
memcpy:    47 ms
memmove:   24 ms
std::copy: 24 ms
(counter:  1787289600)

cast:      24 ms
memcpy:    47 ms
memmove:   24 ms
std::copy: 24 ms
(counter:  1787289600)

cast:      24 ms
memcpy:    47 ms
memmove:   24 ms
std::copy: 24 ms
(counter:  1787289600)

cast:      24 ms
memcpy:    47 ms
memmove:   24 ms
std::copy: 24 ms
(counter:  1787289600)

cast:      24 ms
memcpy:    47 ms
memmove:   24 ms
std::copy: 24 ms
(counter:  1787289600)

cast:      24 ms
memcpy:    47 ms
memmove:   25 ms
std::copy: 24 ms
(counter:  1787289600)

cast:      24 ms
memcpy:    47 ms
memmove:   24 ms
std::copy: 24 ms
(counter:  1787289600)

cast:      23 ms
memcpy:    47 ms
memmove:   24 ms
std::copy: 24 ms
(counter:  1787289600)


 Performance counter stats for './a.out' (10 runs):

            119.10 msec task-clock                #    0.998 CPUs utilized            ( +-  0.19% )
                 0      context-switches          #    0.003 K/sec                    ( +- 71.15% )
                 0      cpu-migrations            #    0.000 K/sec
               145      page-faults               #    0.001 M/sec                    ( +-  0.23% )
       510,528,708      cycles                    #    4.287 GHz                      ( +-  0.18% )
     1,605,033,817      instructions              #    3.14  insn per cycle           ( +-  0.00% )
       400,769,184      branches                  # 3365.053 M/sec                    ( +-  0.00% )
            98,303      branch-misses             #    0.02% of all branches          ( +-  0.20% )

          0.119285 +- 0.000223 seconds time elapsed  ( +-  0.19% )