Running the Program and Continuing Old Runs

Running the program simply involves running the executable that you created when you compiled it. There are no command line parameters, so you can just type ``./latticeeasy'' and the program will run and create all of its output files in the current directory.

Sometimes, however, you might want to continue a run that you've already done. Suppose, for example, that you ran the program to $t=100$ and then decided you wanted to see what happens at later times. This is possible if you have set scheckpoint to save a grid image for the first run. (See below.) Simply edit parameters.h so that $tf=200$ (or whatever you want it to be), set continue_run to be one or two, and rerun. If $continue\_run=1$ then the program will append the new data to the output files from the previous run. If $continue\_run=2$ then new files will be created with a different extension. The default extension for output files is ``_0.dat," but if you choose to create new output files for a run continuation they will have the extension ``_1.dat" for the first continuation, ``_2.dat" for the next one, and so on.

If you set the program to continue a run and there is no grid image file in the current directory it will simply begin a new run. If the program detects a grid image file in the directory with a stored time at or after the value of tf it will simply give you a warning message and exit.

Note that if you do a run to $t=100$ and then continue it to $t=200$ the resulting output won't necessarily be exactly the same as if you had run it directly to $t=200$. Assuming you left the parameter noutput_times at $1000$ for all of the runs the program would generate means and variances at roughly $1000$ times from $0$ to $100$ and then another $1000$ from $100$ to $200$, so doing the run in two stages will produce twice as many output times as doing it in one step. You can change this behavior by adjusting the number of output times per run. The actual field evolution should be the same either way.