Mail Archives: djgpp/2004/03/23/09:30:42
Andrew Cottrell <andspam at swiftdsl dot com dot au> wrote in message news:<tqnv50pv1c4eh2de8774a6ok8vrc5kjpau AT 4ax DOT com>...
> >It's interesting to note the differences in compiling speed when i
> >updated each to allegro 4.03. The make session was very long under
> >the p75 (but the compiled examples ran very smoothly). Works fine
> >though. I do have an oldish win2000 laptop but that runs mingw, and
> >linux of course has gnu tools anyway.
> What you are seeing is that as more RAM is used up in the compilation
> process then the time goes up exponential (in most cases). As the
> Allegro C++ code becomes more complex and later versions of GCC use
> more RAM to compile the app the more the swap file is used. This
> occurs with other compilers and will continue as time goes on.
Indeed, though i'd be a little surprised if DJGPP was filling up the
full 16mb compiling allegro i guess it could have been. I haven't
updated it's version of GCC since i downloaded the package some time
back, so i think it's using 2.95something. When that P75 was new it
has 8mb ram and Windows 95a installed. It ran ok too, I remember a
friend who had a 486-dx2-66 with the same ram thinking it was very
fast!
> I
> remember when I first used a 386 PC it could compile C code very fast,
> but the compiler was not very smart and as such the resulting code was
> not optimized compared to the current compilers.
>
> >no big point here, just a nod of appreciation to the scalable and
> >useful DJGPP package, and wondering if anyone has it on an even more
> >modest machine - and what you've been writing with it recently.
> As from other responses there are allot of old PC's still being used
> with DJGPP which are not high enough spec to run Windows.
That's good to hear, it's good to see older machines in useful roles.
- Raw text -