Creating 2D plots in C/Fortran

• /home1/00478/tg457444/502/Pgm2/Fortran
• /home1/00478/tg457444/502/Pgm2/C

You should be able to cd to your directory, and then

       cp either-Fortran-or-C-directory-above/* .

and this would copy all the files from that directory to your account.
OK, so then what?

Each directory has a couple source files and a README file.
README-Plotting-Fortran/or/C.txt how to call the routines

• contr.f90 or contr.c - the 2D contour plotting routine
  
• sfc.f90 or sfc.c - the Surface routine

particularly for Fortran, do please replace your old sfc.f90 with this one.

• in C, you will have to add the contr() prototype information to the other prototype declarations near the top of your main pgm2.c routine.
  
• In Fortran, you will need to add the INTERFACE settings to those already present near the top of your main pgm2.f90 routine. Instructions on how to do that are in the contr.f90 or contr.c files.

The last step to prepare your code to do the contour plots is to change the Makefile so everything compiles OK. You would edit "Makefile" the way you edit anything else - it is just a text file. Here are the changes you would want to make there:

• Where it says PROGRAM =, change that to PROGRAM = pgm2.
  
• (yes, please change your main program to be pgm2.f90 or pgm2.c, not pgm1)
• Change all the places where pgm1 appears in the Makefile to be "pgm2"
• Adding contr:
• in the line that says OBJECTS =, add "contr.o" to that list.
  
• in the line that says SOURCE =, add contr.f90 or contr.c to that list.

That should be all the changes you need. Then add a call to the contr routine in your main program (pgm2.f90 or pgm2.c) after you call your ic() routine, so you can plot contours of your 2D s1() array or u() array or v() array. But do know that there is a trick to not plot the ghost points of the s1 array, which is described in the README file in each directory (Fortran or C) that you are copying.

• in Fortran, we can use array subscripting to tell the program to plot only the interior (physical, not-ghost) grid points. For example, as shown in the README, call contr( s1(1:nx,1:ny), ... This tells Fortran to only pass the interior values to the contr() routine. when plotting u or v, you either need to plot all the points - in which case you tell contr that the dimensions are nx+1,ny (for U) or nx,ny+1 (for V), or average the points to the center s1() locations and plot that. I do the latter in my plots. One way to do the latter - averaging velocity points to the s1 locations - is to plot (for U): 0.5*( u( 1:nx,1:ny)+u(2:nx+1,1:ny) ) which averages everything in X ... and gives you a nx,ny result.
  
  
• in C, I set up an array just for plotting. I called it splot() and it has dimensions [NX][NY] -- no ghost points. (see the README for an example). So you can use this array for plotting s1 or s2 or u or v, as long as you interpolate everything to the (nx,ny) dimensions of splot. Once you have done that, you call contr in the usual way. contr(NX,NY,splot, ... note that in copying s1 to splot() we subtract off I1 or J1 from the array indices (see the README file for an example) to "remove" the ghost point aspects of the indexing, as contr() doesn't understand ghost points. See the README file for an example of how to do this. for plotting u or v, either plot all the points, in which case you tell contr that the dimension are NX+1,NY (for U) or NX,NY+1 (for V), or average the U or V values to the center s1() locations i.e. unstaggering them. For u, you would have splot(i,j) = 0.5*( u(i,j)+u(i+1,j) ) to unstagger in the X direction for U - and then do contr( NX,NY,splot... as before.

With those changes to your code and the Makefile, you can hopefully then just type "make pgm2" to build the program.