Re: Machine code parsers (entropy of machine code)

Derek M Jones <derek@knosof.co.uk> wrote
| I once did a little experiment. I measured how well gzip compressed
| executable programs for a vareity of machines. There did seem to
| be some correlation between compression ratios for the same programs,
| compiled for different cpu's.


andrey@ix.netcom.com "Andrey I. Savov" writes:
> Well, you see, the gzip compression of executable code cannot give you
> a good idea about the entropy of the machine code, because gzip
> compresses everything as raw data.


Several people pointed this out in private e-mail. Some also pointed
out that data values would be very common. I had made the assumption
that because I was using large programs such effects would be small.
But I did not check this assumption :-(


> As for different CPUs I'd expect RISC code to have smallest entropy
> (compresses best).


I would agree with this.




| Do you think optmised code will have a higher or lower entropy?


> It's hard to say, I'd expect higher than non-optimized, because
> usually some redundancies are removed during the optimization.


Unoptimised code might have a higher compression ration because
it contains longer sequences of duplicate instructions. Optimisations
tend to mangle code -> fewer common sequences.


I guess we need some experimental results.


derek


--
Derek M Jones tel: +44 (0) 1252 520 667
Knowledge Software Ltd email: derek@knosof.co.uk
Applications Standards Conformance Testing http://www.knosof.co.uk
--

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