Array append performance

[Lionello Lunesu]

The problem is that the timing for the small arrays cannot be trusted as 
there's more overhead than actual code being tested. I've changed your 
code by doing each test 100_000_000/i times. I've also added 'cpuid' for 
synchronization, as Don suggested:

import std.c.string;

long timer()
{
     asm
     {   naked                   ;
         push ECX;
         push EBX;
         mov EAX, 0              ;
         cpuid                   ;
         pop EBX;
         pop ECX;
         rdtsc                   ;

         ret                     ;
     }
}

void test1(byte[] a, byte[] b)
{
     a[] = b[];
}

void test2(byte[] a, byte[] b)
{
     memcpy(a.ptr, b.ptr, a.length);
}

void main()
{
     for (int i = 4; i < 100_000_000; i *= 2)
     {
         auto b = new byte[i];
         auto a = new byte[i];

         test1(a,b);//pre-cache

         auto count = 100_000_000 / i;
         auto start = timer();
         while (count--)
             test1(a, b);
         auto end = timer();
         auto r1 = end - start;

         count = 100_000_000 / i;
         start = timer();
         while (count--)
             test2(a, b);
         end = timer();
         auto r2 = end - start;

         printf("i: %8d,\t[]=[]: %8lld,\tmemcpy: %8lld\n", i, r1, r2);
     }
}

Results on my Core2 Duo:

i:        4,    []=[]: 2752539660,      memcpy: 429537312
i:        8,    []=[]: 1004458383,      memcpy: 229984623
i:       16,    []=[]: 492054948,       memcpy: 128590560
i:       32,    []=[]: 309960657,       memcpy: 135892809
i:       64,    []=[]: 204971832,       memcpy: 79507359
i:      128,    []=[]: 131925789,       memcpy: 52222626
i:      256,    []=[]: 73768896,        memcpy: 40184559
i:      512,    []=[]: 43827948,        memcpy: 28040859
i:     1024,    []=[]: 29030985,        memcpy: 21078432
i:     2048,    []=[]: 21645027,        memcpy: 18953199
i:     4096,    []=[]: 18099945,        memcpy: 16022259
i:     8192,    []=[]: 15905007,        memcpy: 15399036
i:    16384,    []=[]: 15567399,        memcpy: 15795747
i:    32768,    []=[]: 32244507,        memcpy: 30654063
i:    65536,    []=[]: 30824595,        memcpy: 31286475
i:   131072,    []=[]: 32248179,        memcpy: 30950757
i:   262144,    []=[]: 34880868,        memcpy: 31166667
i:   524288,    []=[]: 31220802,        memcpy: 35813277
i:  1048576,    []=[]: 42463341,        memcpy: 42121386
i:  2097152,    []=[]: 122218776,       memcpy: 118839150
i:  4194304,    []=[]: 111864294,       memcpy: 110864988
i:  8388608,    []=[]: 110600226,       memcpy: 107638083
i: 16777216,    []=[]: 97014645,        memcpy: 95789574
i: 33554432,    []=[]: 79344468,        memcpy: 78277374
i: 67108864,    []=[]: 77437629,        memcpy: 78478767

Conclusion: memcpy wins big time for i <= 8192.