[Liboil] yet another copy_u8

[Adam D. Moss]

I know I shouldn't be wasting time on the little C implementations,
but I was wondering if Duff's Device still held any merit - seems
it does (up to 30% improvement for non-tiny n here over
copy_u8_llints, which is itself up to around 30% faster than
copy_u8_ints - hope I didn't add an egregious bug).

static void
copy_u8_llints_duff (uint8_t *dest, uint8_t *src, int n)
{
   int i;
   switch(n&7) {
   case 7: *dest++ = *src++;
   case 6: *dest++ = *src++;
   case 5: *dest++ = *src++;
   case 4: *dest++ = *src++;
   case 3: *dest++ = *src++;
   case 2: *dest++ = *src++;
   case 1: *dest++ = *src++;
   default: ;
   }
   n >>= 3;
   while (n) {
     switch (n & 15) {
     default:
     case 0: *(uint64_t *)dest = *(uint64_t *)src; dest += 8; src += 8;
     case 15: *(uint64_t *)dest = *(uint64_t *)src; dest += 8; src += 8;
     case 14: *(uint64_t *)dest = *(uint64_t *)src; dest += 8; src += 8;
     case 13: *(uint64_t *)dest = *(uint64_t *)src; dest += 8; src += 8;
     case 12: *(uint64_t *)dest = *(uint64_t *)src; dest += 8; src += 8;
     case 11: *(uint64_t *)dest = *(uint64_t *)src; dest += 8; src += 8;
     case 10: *(uint64_t *)dest = *(uint64_t *)src; dest += 8; src += 8;
     case 9: *(uint64_t *)dest = *(uint64_t *)src; dest += 8; src += 8;
     case 8: *(uint64_t *)dest = *(uint64_t *)src; dest += 8; src += 8;
     case 7: *(uint64_t *)dest = *(uint64_t *)src; dest += 8; src += 8;
     case 6: *(uint64_t *)dest = *(uint64_t *)src; dest += 8; src += 8;
     case 5: *(uint64_t *)dest = *(uint64_t *)src; dest += 8; src += 8;
     case 4: *(uint64_t *)dest = *(uint64_t *)src; dest += 8; src += 8;
     case 3: *(uint64_t *)dest = *(uint64_t *)src; dest += 8; src += 8;
     case 2: *(uint64_t *)dest = *(uint64_t *)src; dest += 8; src += 8;
     case 1: *(uint64_t *)dest = *(uint64_t *)src; dest += 8; src += 8;
     }
     n = (n - 1) & ~(int)15;
   }
}
OIL_DEFINE_IMPL (copy_u8_llints_duff, copy_u8);


--adam