Hi everyone:
I realize this topic has come up in the past but I haven't seen a resolution to my particular problem. I am trying to use Seagrid for the first time and simply trying to follow the example in the tutorial.
I am using a Linux 64-bit machine with Matlab 7.1.0.183 (R14).
I realized (after a few headaches) that I had to specify the integer values (8-bit) for pointers and such in the .F files. Making the appropriate changes, I can compile and successfully test mexinside and mexrect. Unfortunately I run into problems with mexsepeli. While the program runs, the computed grid error is not always zero. This problem originates here:
c Load the data into Fortran matrices.
c
c x
p = mxGetPr(prhs(1))
size_m = mxGetM ( prhs(1) )
size_n = mxGetN ( prhs(1) )
size_mn = size_m*size_n
call mxCopyPtrToReal8(p,tempx,size_mn)
k = 0
do 10 j = 0, size_n - 1
do 5 i = 0, size_m - 1
x(i,j) = tempx(k,0)
k = k + 1
5 continue
10 continue
The 'boundary' of the matrix x is specified correctly but the interior of the matrix remains to be zero. So then, the error in the middle of the matrix is obviously greater than zero. The same occurs with y.
I believe I've specified all the integers correctly but can't seem to fix this problem.
Any thoughts?
Thanks so much
Joe