>
>I just tested this. Using shifting vs. multiplication is faster with about 3/10000000 seconds.
>
The multiply is actually faster than the bitrshift. Probably due to the function call. I found that an empty UDF is 300% slower than a bitrshift and if you include an lparameters statement than the UDF is 700% slower.

>
>A good C/C++ compiler should optimize internally code that multiplies by 256 using a right shift, IMHO.
>
Your HO isn't far from Microsofts HO. Here is the assembly of a simple C++ app.

8:        y = x * 256;
0040103C   movsx       eax,word ptr [x]
00401040   shl         eax,8
00401043   mov         word ptr [y],ax

Peter