Input parameters are generally specified in a Python script
$ make .data
*.data files that the Fortran code requires.
Most example directories contain a
Makefile that offers several
$ make help
for a list. Often:
$ make .plots
is all you need to type to create the data files, compile the code, run it, and produce plots as png and html files.
Or, if you just want to run the code and produce output without making all the plots (and then do the plotting interactively, for example):
$ make .output
Note: There is a dot before
output in the above
The directory where output and plots are stored is specified in the Makefile.
The Makefile in most directories includes a common Makefile found at $CLAW/clawutil/src/Makefile.common that does most of the work. If you get the error message:
Makefile: /clawutil/src/Makefile.common: No such file or directory
then the environment variable CLAW is not set properly. See Set environment variables.
The “make .output” command runs the code and stores the name of the output directory in the file
.outputand it is the modification time of this file that is checked relative to the dependencies. (Note: the unix command ls generally does not display files that start with a dot so this file may be invisible unless you use “ls -a”.)
If you want to re-run the code and encounter:
$ make .output $ make: `.output' is up to date.
you can remove the file
.outputto force the code to be run again.
Similarly, remove the file
.plotsto force the plots to be recreated.
If you change the compiler flags FFLAGS in the Makefile or as an environment variable, then you should make sure that all files used are recompiled with the new flags. The Makefiles as written do not catch this dependency and will not recompile all the .o files when the Makefile changes. To force recompilation, use:
$ make new
See Fortran Compilers for more about compiler flags.