Hi,
I remember reading a book or website (or even a long article) on fast embedded programming tricks. Along the lines of "divide by 10, shift left three times and add it back to itself". Of course that one isn't correct but there were a whole lot of fast algorithms in one place -- I just cannot remember where I saw it.
The specific one I am looking for is a fast "shift large buffer right by n bits", and I seem to remember there was a quick way of doing it in a HLL.
Anybody remember a reference like this,