A.5.261 PSWAPD: Swap Packed Data
PSWAPD mm1,mm2/m64 ; 0F 0F /r BB [PENT,3DNOW]
PSWAPD swaps the packed doublewords in the source operand, and stores
the result in the destination operand.
In the K6-2 and K6-III processors, this opcode uses the mnemonic PSWAPW,
and it swaps the order of words when copying from the source to the
destination.
The operation in the K6-2 and K6-III processors is
dst[0-15] = src[48-63];
dst[16-31] = src[32-47];
dst[32-47] = src[16-31];
dst[48-63] = src[0-15].
The operation in the K6-x+, ATHLON and later processors is:
dst[0-31] = src[32-63];
dst[32-63] = src[0-31].