Re: Interested about number crunching in Ada

[Gautier <gautier@fakeaddress.nil>]

anon:

> From GNAT Documentation:  The GIGI procedure is defined as Gnat-To-Gnu.

Bravo, *that's* the point: GNAT-to-GNU, not GNAT-to-C !
Thank you for finding yourself your mistakes ;-).
You are mistaken by the fact that in its early days, GCC was the "GNU C 
Compiler", but for a long time now it is the "GNU Compiler Collection".

> And if you dump each step of the GNAT to object as I stated (see 
> previous post) you can use a gcc or other gnu compilers to continue 
> compiling the code.  

You can also take any other compiler's sources and dump its tables, internal 
representations for optimizing; and usually commercial compilers (even if you 
don't have their sources) also offer to dump the assembler code beside the 
corresponding object code. The fact that the object code is spit by the 
compiler executable and not by as.exe makes absolutely no difference for the 
code itself; in compilation time it takes maybe a fraction of second more, 
that's all.
The steps: Source -> Internal representation (tables, tree or whatever) -> 
Machine code (eventually with 1-to-1 equivalent asm inbetween) are the same on 
every compiler. The fact that GNAT does it openly and via several executables 
gives you the impression that it is more complicated; it is just more 
convenient to do so in the context of the GNU toolchain. Conversely, if the 
same stuff was packed into a single .exe and there was no documentation about 
how it works, you would believe that it is simpler because it is all stuffed 
in a black box. It's a bit naive...

>> As for HP Ada, I don't care about politics, only about performance :-).
> 
> Well if HP's Ada performance is like TI's PASCAL.  It depends on how 
> the system is configured.  [...]

Well, that's the same for all systems. I happend some years ago to compare DEC 
Fortran and DEC Ada on the same AXP machine with some matrix crunching code. 
Both timings were very near, which was a compliment for the Ada compiler, but 
less a surprise when I heard that the compilers back-ends were the same (like 
for the GNU Compiler Collection ;-) ). That's it. And all this dramatic story 
about TI Pascal, politics, jobs and money seem very thrilling but (I'm afraid) 
a bit obscure to me...
______________________________________________________________
Gautier         -- http://www.mysunrise.ch/users/gdm/index.htm
Ada programming -- http://www.mysunrise.ch/users/gdm/gsoft.htm

NB: For a direct answer, e-mail address on the Web site!