关于_mm_shuffle_epi32的使用

貌似这个函数很强大：

    __m128i src    = _mm_set_epi32( 0x3333, 0x2222, 0x1111, 0x0000);
    __m128i dest  = _mm_shuffle_epi32(aas, _MM_SHUFFLE(1, 0, 3, 2));

结果值：
    dest = {0x00001111 00000000 00003333 00002222}

 用途:不超过3个指令完成许多普通数据移位操作，例如广播，交换，反转等。

"No more than 3 instructions, using PSHUFLW/PSHUFHW/PSHUFD, are required to
implement many common data shuffling operations. Broadcast, Swap, and Reverse"

  具体参考:

  1. http://msdn.microsoft.com/en-us/library/4d3eabky(vs.71).aspx

  2. http://www.intel.com/design/processor/manuals/248966.pdf