Message-ID: <211@looking.UUCP>
Date: Sat, 10-Nov-84 00:00:00 EST
Article-I.D.: looking.211
Posted: Sat Nov 10 00:00:00 1984
Date-Received: Mon, 12-Nov-84 09:42:47 EST
References: <3105@utah-cs.UUCP>
Organization: Looking Glass Software, Waterloo, Ont
Lines: 38

I was asked for a good programming problem to see what processor has the
tightest code.   We held such a test around here once, and I'll describe
it and see who on the net wants to give it a try for their favourite
microprocessor.


PROBLEM:
	Write an assembler subroutine that prints small roman numerals.

	To be specific, the routine must take a roman numeral from 0-255
	in the accumulator, and print it.  Zero should result in the
	null string.

	Assume you have a routine called PRINT you can do a proper subroutine
	call to which prints the ascii character in the accumulator and
	alters no registers or flags.

	Make no assumptions about the code location or the location of
	internal data tables within the code - ie. don't try and say your
	code runs in zero-page or other such tricks.

	Execution time is not really important, but it shouldn't be grossly
	out of line.  If you want to include the worst case execution time
	for your routine, you are encouraged to do so.

	Output may be in either case, but must all be in the same case.

	Your routine must enter and exit like a normal subroutine (ie. use
	the RTS or equivalent instruction)

Mostly we want to see the results for small chips like 6502, z-80, 6800,
6809 and others.  We wouldn't mind looking at results from larger chips
like 8086, 68000, z-8000 or vax, but they would be a seperate contest.

Next week, I will announce the size (I already know) of the best routine
we have come up with so far.  After that, I'll post the code.
-- 
Brad Templeton, Looking Glass Software Ltd. - Waterloo, Ontario 519/884-7473