Re: variable length instructions, was Chicken or the Egg?

On 15 Oct 2000, VBDis wrote:


> michael@moria.de schreibt:
> >The reason for starting with short instructions is that you usually get a
> >smaller executable


Smaller executable, but not neccesarily faster. Sometimes speed comes
at the expense of size, and often it is not the code that makes an
executable large.


[snip]


> Assume that you have two (or more) instructions, which refer to a
> target at an offset, which would just fit into a single byte. Then two
> stable states are possible, with regards to code size, where 2
> instructions with byte-size arguments will fit, or where 2 instruction
> with word-size arguments will move the target away 2 more bytes, so
> that word-size arguments really are required.


In the M601 chip the two least significant bits of an instruction offset
are not recorded, eliminating the effect of size on PC relative branches.
This assumes that all instructions must be aligned to four-byte boundries.


-Dietrich Epp

Return to the comp.compilers page.
Search the comp.compilers archives again.