laplace example: inputs

[Stephen Siegel]

The laplace example is looking good. But instead of initializing the values of the function to 0.0 on the interior and 10.0 on the boundary, why not make all of the initial values an input to the program. That would be much more general. That means adding an input variable like

input real[ny][nx] initialValues;

Do this in both sequential and parallel versions. Use these values to set the initial values.

Also, instead of copying all the values between old, new, and tmp, instead use a single 3-d array and a mod-2 counter to accomplish the same thing.

[zirkel]

The laplace example is not quite working yet, but this part is done.