$x = $y * 16;
$x = $y << 4; #Much faster
First I must note that I love these types of micro-optimizations. Durring the winter I work on a robotics competition where we code in pBasic, and we have to constantly worry about effiency. That said:
There are multiple layers of abstraction between the perl code you write and the machine code produced. When we use a high level language (like perl), we should write what comes natural ($x = $y * 16;), and trust that the interpreter will turn it into the most efficient machine representation. Also, when someone comes along and has to maintain the code, an asembly programmer may very easily understand your optimizations, but odds are a typical perl programmer may not.
The 15 year old, freshman programmer,
Stephen Rawls