🚀 go-pugleaf

Hallo granville!

Friday June 30 1995 22:22, granville wrote to All:

 g>  The correct way to state this is you can commonly produce code that runs
 g> faster than typical assembler. In code completion / bug rates there is
 g> no comparison at all.....!
 g> In any reasonably complex system, the algorithm is more important than
 g> the code detail, and that is where HLL's have it all over assembler. I
 g> have seen systems get a 4x speed boost when coded in Modula-2, because
 g> the programmer could
 g>  a) work closer to the problem
 g>  b) code, measure and optimise multiple approaches quickly

I have seen the opposite far more often. When programming in assembler, you
have to understand what you are doing, if you don't, you lock up your system.

I don't plea for pure assembler, but a lot of things can be done better in assembler.

Some things can't be done easily in Modula-2 (or TurboPascal or Basic)

An example : try mirroring a BITSET of an arbritary size.

        11001 becomes 10011.
       110101 becomes 101011. etc.

I REALLY like to see a modula routine for this (the length of the bitset is a parameter of the procedure, CARDINAL) that uses shifting. Other concepts( with
a for loop) are less optimal in both assembler and modula2, (except when
compiling 386+, then the FOR algoritm might be more favourable, because the IN and INCL/EXCL command are only one instruction large).

My assembly routine is, parameter handling inclusive, 72 bytes.

Marco  (M.W.T.J.v.d.Voort@stud.tue.nl or MarcoV@stack.urc.tue.nl)