Mail Archives: djgpp/1995/08/02/10:44:25
> )I'm writing a database program. The source is about 35000 lines so I think
> )it is big enough for comparison. I was using Borland C 3.00. It generated
> )code of about 230K. Now I have switched to DJGPP 1.12maint4 (not 2.0 yet)
> )and it generates code of about 230K. So in my case the code size is about
> )the same (note that you have to strip the symbols from the executable
> )before comparing and I also used -O3 for the best optimization).
>
> This part is reasonable and useful.
>
> )Anyway, the code size was about the same but the speed gain with DJGPP was
> )enourmous.
>
> This part is not. "Enourmous [sic]" is not an objective measure.
Sorry about this, but I don't have any real objective numbers. I was
just trying to give a general feeling, and what I wanted to say was that
I was impressed by the speed gain. But of course the comparison is not
really fair for Borland C 3.0 since this compiler only compiles to
16-bit code.
Speed is always difficult to quantize, therefore I tend to use subjective
measures.
> Second, perhaps DJGPP is not suitable for all applications. Perhaps it
> is better for some, and not so good for others.
I think this is true for ALL programs. But I also think DJGPP is a very good
general purpose C compiler. Watcom is probably also a very good general
purpose compiler (I wouldn't know, I've never used it) but Watcom is expensive
in comparison with DJGPP.
> Mike
Greetings,
==============================================================================
Jorrit DOT Tyberghein AT uz DOT kuleuven DOT ac DOT be, University Hospitals KU Leuven BELGIUM
Tourist, Rincewind decided, meant "idiot".
-- (Terry Pratchett, The Colour of Magic)
==============================================================================
- Raw text -