C++ programming (was Newbie Alert: Beginning parallel program ming with Scyld)

Toon Moene toon at moene.indiv.nluug.nl
Wed Oct 16 15:11:34 PDT 2002

Robert G. Brown wrote:

> Different compilers are suitable for different kinds of tasks, and even
> now it could easily be that fortran, written a certain way, e.g.
> vectorizes better than C.  For one thing, C programmers (really speaking
> only for myself, not necessarily ALL C programmers:-) tend to use
> pointers a lot because we tend to use dynamic allocation of memory a
> lot, and use this dynamic memory in lots of "interesting" ways to build
> data structures (such as linked lists, trees, arrays of structs) that
> defeat certain kinds of optimization, which like to be able to presume
> that an array starts on a given index, is a given size, and can be
> unrolled in a certain way to give optimal performance.

Yep, I can clearly see that Fortran wouldn't buy you much then.  You can 
do all of the above with Fortran 95, but the programs would be more 
verbose than the equivalent C programs, and I would be surprised if the 
Fortran compilers were actually capable of taking advantage of the more 
strict aliasing syntax (you have to indicate each item in your program 
that can be the target of a pointer as TARGET, so any that aren't can be 
assumed alias-free).

What I think the strength of Fortran is, is the ease to use multi-rank 
arrays - which is the reason Fortran is such a natural language for, 
e.g. weather forecasting code.

In Fortran, you can write:

       READ*,NX,NY,NZ   ! Read dimensions of simulation
       ... allocation of PS, T, and Q ...
       DIMENSION PS(NX,NY)    ! Surface pressure
       DIMENSION T(NX,NY,NZ)  ! Temperature
       DIMENSION Q(NX,NY,NZ)  ! Specific humidity
       DIMENSION TD(NX,NY,NZ) ! Dew point temperature

where TD is an array local to subroutine `COMPUTE' that has the same 
dimensions and layout as T and Q.  Although C99 will give you the 
equivalent of automatic arrays (like TD), an array in C is still 
basically a pointer augmented with some funny syntax and it will not 
allow such clear declarations as the above.

It gets more interesting if the problem at hand has a most `natural' 
translation using variable upper *and lower* bounds, like

       DIMENSION A(N-M: N+M), B(-M: M), C(N, M)

and mutatis mutandis for higher-than-1-rank.

Perhaps C++ solves that problem - I do not know enough of the language 
to tell.

Toon Moene - mailto:toon at moene.indiv.nluug.nl - phoneto: +31 346 214290
Saturnushof 14, 3738 XG  Maartensdijk, The Netherlands
Maintainer, GNU Fortran 77: http://gcc.gnu.org/onlinedocs/g77_news.html
Join GNU Fortran 95: http://g95.sourceforge.net/ (under construction)

More information about the Beowulf mailing list