allocate issue

[milenko76]

I have problems with my code.As a result of calculation I got rank2 l(21,1) array.Due to the THE ITERATIVE PROCESS I will get the same array in every iteration.How to allocate values and use them later.Should I create new array
like lbig(iter,nmax,1)=l(nmax,1)?

 

[FJacq]

First of all a question : why do you get such strange matrix of rankb2 l(21,1), i.e. a single column ? Why not simply a vector l(21) ?

If you want to store all iterative results, then you have several possibilities :
- writing the successive results in a file,
- storing the results in a large matrix like lbig(nmax,nitermax). I prefer to use the last index for the iteration but this is mainly a question of taste... You don't need to use 3 dimensions for the array lbig.
=> lbig,iter)=l,1)
- using a linked list to allocate only the needed memory when a new iteration is performed.

François Jacq

 

[milenko76]

Well I have done this but:
lmat=p+q
l=reshape(lmat,shape=shape(l))
lbig,iter)=l
write(*,*)l
ircg.for(156): error #6366: The shapes of the array expressions do not conform. [LBIG]
lbig,iter)=l
-------^
real*4 lbig(nmax,itermax) and l(21)
lbig,iter) is which rank 2 or 1?

 

[FJacq]

lbig,iter) is a vector (rank 1)

Notice that I wrote, in my example : lbig,iter)=l,1) and not lbig,iter)=l

François Jacq

 

[milenko76]

No,doesn't work either.
ircg.for(156): error #6351: The number of subscripts is incorrect. [L]
lbig,iter)=l,1)
--------------------^
Well I have reshaped l with l=reshape(lmat,shape=shape(l)) and l(21).

 

[FJacq]

Show us the actual declarations of all arrays please (especially lbig and l). Without them, we cannot explain the error.

François Jacq

 

[milenko76]

integer*4 n1,i,j,icd,ivarprd,k,iper,is,fa,fb,iter,x
integer*4 iprd,iprd0,isite,ico
character fexpdat*200,fmodel*200,fout*200
real*4 prunit,period
real:: g,g1,e,e1
real:: zxy_re_e,zxy_im_e,zxy_re_h,zxy_im_h
real:: sens_zxy_ree,sens_zxy_ime,sens_zxy_reh,sens_zxy_imh
real:: ro_e,ro_h,phase_e,phase_h
real:: sens_ro_e,sens_phase_e,sens_ro_h,sens_phase_h
real:: ro_ed, phase_ed,ro_hd,phase_hd
real:: dmro_e,dmphase_e,dmro_h,dmphase_h
real*4 norm20,norm21
real:: zag21,zag22,koefnumero,dmfnorm
real,dimension(21) :: l
real, dimension(80,21) :: sens,wfm
real, dimension(21,80) :: ist,ftw
integer,dimension(80,80) :: wd,wd2
real,dimension(80,1):: d,zag1
real,dimension(1,21):: trlt
real,dimension(21,1):: l0,p,q,lmat,dlm
integer,dimension(21,21) :: wm,wm2,wm3
lbig is declared in module :
module mt2ddat
c ==============
c Further parameters of the 2D model related to the output data of the
c modeling
use params
c
implicit none
c
c Actual number of periods, number of sites on the surface we are
c interested in
integer*4 np,nsites
c
c Periods of the EM field
real*4 per(npmax)
c
c Positions of the sites of interest in terms of the horizontal
c mesh index (1 is the left margin, N is the right margin);
c IRCPR is a flag indicating that the reciprocity approach will (1)
c or will not (0) used in the sensitivity evaluations
integer*4 isit(nmax),ircpr
c
c MT transfer functions for one single period at all surface mesh
c nodes and MT transfer functions aggregated for all periods computed
real*4 zmd1p(nmax,nfmax),zmd(npmax,nmax,nfmax)
real*4 zmd2p(nmax,nfmax),zmd2(npmax,nmax,nfmax)
real*4 aprph1d(nmax,nfmax)
real*4 sens1p(nmax,nfmax)
real*4 data(nmax,nfmax)
real*4 dmf(nmax,nfmax)
real*4 alpha

c all calculation arrays for inversion
real*4 lbig(nmax,itermax)

end module
c

 

[milenko76]

Parameters in this module:
module params
c =============
c Maximum values of parameters of the 2D model
implicit none
c
c Maximum number of mesh lines (including the margins) in the
c horizontal and vertical direction
integer*4 nmax,mmax
parameter(nmax=41,mmax=35)
c
c Maximum number of periods and of various MT transfer functions
integer*4 npmax,nfmax
parameter(npmax=20,nfmax=4)
c
c Maximum number of uniform conductivity domains in the 2D model
integer*4 ncmax
parameter(ncmax=(nmax-1)*(mmax-1))
c Maximum number of iterations
integer*4 itermax
parameter(itermax=50)
end module

 

[milenko76]

It works now I have mdae changes ito parameters file.

 

[FJacq]

You have a problem because of lbig(nmax,itermax) and l(21) with nmax=41

You should declare either l(nmax) or lbig(21,itermax).

Of course, in setting nmax=21, you solve also your trouble but your code would be a little bit cleaner in declaring the same way all related arrays.

François Jacq

 

[milenko76]

Well I have introduced new variable nnmax=21,you are right the second version should be cleaner.