Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!swrinde!elroy.jpl.nasa.gov!ncar!bierstat.scd.ucar.edu!morreale
From: morreale@bierstadt.scd.ucar.edu (Peter Morreale)
Newsgroups: comp.lang.fortran
Subject: Re: vectorization question
Message-ID: <10836@ncar.ucar.edu>
Date: 29 Mar 91 19:32:47 GMT
References: <1991Mar29.141313.7418@ariel.unm.edu>
Sender: news@ncar.ucar.edu
Reply-To: morreale@bierstadt.scd.ucar.edu (Peter Morreale)
Organization: Scientific Computing Division / NCAR, Boulder Co.
Lines: 44

In article <1991Mar29.141313.7418@ariel.unm.edu>, prentice@triton.unm.edu (John Prentice) writes:
> Consider the following loop:
> 
>       do 30 k=1,kmax
> 	  do 20 j=1,jmax
> 	      do 10 i=1,imax
> 		  a(i,j,k)=...
>    10         continue
>    20     continue
>    30 continue
> 
> On the Cray, only the inner most loop will vectorize.  Does anyone
> have a suggestion for how to collapse this loop while still using
> the three dimensional array?  In other words, I need a vectorized
> equivalent to:
> 

  Whether or not you can collaspe the loop structure depends entirely on
  whether i, j, or k is used on the right hand side of the assignment. 

  If the loop indices are not used, you could collaspe the loop as
  follows:

	DO 30 K = 1, IMAX*JMAX*KMAX
	   A(K,1,1) = <whatever>
     30 CONTINUE


  In any event, the Cray Fortran preprocessor (fpp) can help
  significantly in determining whether or not the loop can collaspe. 
  (since it *will* collaspe the loop if it can)  You can still trick
  fpp, but it's a terrific start.

  I'd suggest you run the routine through fpp and see what it produces;

	   % fpp sub.f > sub.fpp

-PWM
------------------------------------------------------------------
Peter W. Morreale                  email:  morreale@ncar.ucar.edu
Nat'l Center for Atmos Research    voice:  (303) 497-1293
Scientific Computing Division     
Consulting Office
------------------------------------------------------------------