MODEL DOWNLOAD AND SETUP¶

Prerequisites¶

There are a few software requirements your computer needs in order to execute FUNWAVE–TVD and to visualize the results. Please ensure you have installed the necessary software prior to downloading the source code.

You will need:

Access to the Internet (Ethernet or WiFi)
SSH Secure Shell Client (e.g., Putty or Cygwin on Windows)
Working suite of compilers and Message Passing Interface (MPI) wrappers to build a parallel version of FUNWAVE-TVD
- If you are using a Mac OS X machine
  
  We recommend you install GNU compilers (e.g., gcc, g++, gfortran) via Homebrew. If you don’t have Homebrew installed, type the following in your terminal:
```
cd ~
mkdir homebrew
curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh
tar xz --strip 1 -C homebrew
```
  If you didn’t previously have these compilers on your system (say via Xcode), then the gcc, g++, gfortran compilers would be placed under /usr/local/bin/ directory; otherwise they could be placed under /usr/local/Cellar/gcc/9.1.0/bin/. You will then need to install the GNU suite of compilers using brew, where the current versions are 9.1.0.
  
  Important: you want to use all suite components [gcc, g++, gfortran] of the same version to avoid possible issues/compiler conflicts later on:
```
cd ~
brew install gcc
brew install gfortran
```
  Lastly, you will need to install MPI compiler wrappers:
```
cd ~
brew install mpich
```
  Note: if you had a trace of previous compilers, you may need to explicitly tell Homebrew during installation to point to/use the newly installed compiler via FULL PATH in your terminal prior to executing the MPI installation command described above. You can do this by:
```
export HOMEBREW_FC = /usr/local/Cellar/gcc/9.1.0/bin/gfortran
export HOMEBREW_CC = /usr/local/Cellar/gcc/9.1.0/bin/gcc
export HOMEBREW_CCX = /usr/local/Cellar/gcc/9.1.0/bin/g++
```
- If you are using a Linux machine
  
  You can use either your system-wide installation commands in your terminal/command line (if you have admin/sudo privileges). In your Ubuntu, Debian, Mint, or other Debian-based distribution, you can check if you have a working suite of compilers via:
```
which gcc
which g++
which gfortran
```
  Otherwise, if you don’t have the suite of compilers already, type the following in your terminal to download and install:
```
sudo apt install gcc
sudo apt install gfortran
sudo apt install mpich
```
  For Red Hat, Fedora, or CentOS Linux kernels, you would replace apt in the three commands above with yum.
  
  Alternatively, you can also use Homebrew to install the required packages. See here for more details: https://docs.brew.sh/Homebrew-on-Linux. It should be noted that Homebrew only does local installation (i.e., no system-wide sudo commands for installation are allowed).
- If you are using a Windows 10 machine
  
  Instructions on how to install FUNWAVE-TVD on Windows 10 can be found here.
- If you are using a Windows OS lower than Windows 10
  
  We recommend that you download the latest version of Cygwin (https://www.cygwin.com). Cygwin is a bash shell/unix emulation program, and contains many of the tools such as tar, gzip/gunzip, and cpp, which will be useful for installation and compilation of FUNWAVE-TVD. In addtion, if you don’t have access to a High Performance Computing (HPC) machine at your home location, you can use the Amazon AWS Cloud Computing service. For instructions on how to use AWS with FUNWAVE, visit MINI-FUNWAVE-TVD Workshop 2019 and review pages 35-41 of the tutorial document (pdf1).
A post-processing toolbox (e.g., MATLAB and Python)

Both MATLAB and Python post-processing scripts are provided with most practice examples. If you do not have access to a MATLAB license, we recommend that you install an Anaconda Python package, usable on any platform.
- Downloading and installing a Python Package (Anaconda)
  
  Arguably one of the best and most comprehensive FREE packages for the Python language, along with most tools and modules (e.g., NumPy, Matplotlib, etc.) is distributed by the Continuum Analytics under the Anaconda package. It is available for Linux, Mac OS X, and Windows machines. You do NOT need administrator privileges to install the Anaconda package; you can do so as a standard user on all three platforms.
  
  For the Anaconda package, go to: https://www.anaconda.com/distribution/.
  
  Choose the appropriate platform (Linux, Mac OS X, Windows) by selecting the correct tab and getting the Anaconda distribution that comes with Python 3.7 (not 2.7). You can either download the Graphical Installer (recommended), or if you are comfortable with the terminal in the Linux/Mac OS X environment, you can also download it through the command line.
  
  For the complete list of packages/modules included in the Anaconda Python distribution, see: https://docs.continuum.io/anaconda/packages/pkg-docs.

Download source code¶

Version beta: not fully tested. click here to download/clone from GitHub

Version 3.4 Released July 12 2019: click here to download from GitHub

Version 3.3 Released July 19 2018: click here to download from GitHub

Version 3.2 Released Jan 27 2018: click here to download from GitHub

Version 3.1 Released Sep 2017: Used for FUNWAVE-TVD Workshop 2017 click here to download

Version 3.0: Released Apr 2017 download here

Versions 1.0, 1.1, 2.0, 2.1: Please contact fyshi@udel.edu

Compile and setup¶

Uncompress the code from the package downloaded

You have two options to compile the program

Use a more automated Makefile. If you can compile the code successfully, go to 3.

Use the following traditional method.

Modify “Makefile” as needed. An example “Makefile” is shown below. The primary variables you will need to change/check are the EXEC in the BEGIN MAKEFILE section, and the flags in the following section. The EXEC variable defines the name of the executable that will be generated. This should be descriptive of which flags (i.e., modules) are active in the “Makefile” for good book-keeping when you have multiple executables.

There are several flags in “Makefile” that specify different compiling structures and turn on/off modules depending on your simulation. To turn on/off a flag, simply comment the line by adding “#” before the command (e.g., # FLAG_8 = -DVESSEL turns the shipwake module “off” in the new executable).

-DDOUBLE_PRECISION: use double precision, default is single precision
-DPARALLEL: use parallel mode, default is serial mode
-DCARTESIAN: Cartesian version, otherwise Spherical version
-DINTEL: if INTEL compiler is used, this option can make use of FPORT for the RAND() function
-DCRAY: for CRAY RAND() and system commands
-DCOUPLING: nesting mode
-DSPHERICAL_IJ_STATION: in spherical mode, if you want your station locations defined by grid point (I,J). Otherwise, station locations should be defined by (lat lon)
-DVESSEL: include shipwake module
-DSEDIMENT: include sediment and morphological module
-DWIND: include wind effect
-DMETEO: include meteo tsunami module
-DMANNING: use Manning formula for bottom friction
-DCHECK_MASS_CONSERVATION: correct mass conservation problem caused by wetting/drying
-DTRACKING: include Lagrangian tracking module
CPP: path to CPP directory
FC: Fortran compiler

This sample “Makefile” shows a case where the shipwakes module (e.g., -DVESSEL) is active, and the model will be executed in parallel (-DPARALLEL) with a Cartesian coordinate system (-DCARTESIAN) using an MPI F90 Compiler (FC = mpif90).

#-----------BEGIN MAKEFILE-------------------------------------------------
    SHELL         = /bin/sh
    DEF_FLAGS     = -P -traditional
    EXEC          = funwave_vessel
#==========================================================================
#--------------------------------------------------------------------------
#        PRECISION          DEFAULT PRECISION: SINGLE
#                           UNCOMMENT TO SELECT DOUBLE PRECISION
#--------------------------------------------------------------------------

    FLAG_1 = -DDOUBLE_PRECISION
    FLAG_2 = -DPARALLEL
#             FLAG_3 = -DSAMPLES
    FLAG_4 = -DCARTESIAN
#             FLAG_6 = -DINTEL
#             FLAG_7 = -DMIXING
#             FLAG_8 = -DCOUPLING
#             FLAG_9 = -DZALPHA
#             FLAG_10 = -DMANNING
#             FLAG_11 = -DSPHERICAL_IJ_STATION
    FLAG_12 = -DVESSEL
#             FLAG_13 = -DVIS_KENNEDY
#             FLAG_14 = -DVESSEL_PANEL_SOURCE
#             FLAG_15 = -DREALISTIC_VESSEL_BODY
#             FLAG_16 = -DMETEO
#             FLAG_17 = -DWIND
#--------------------------------------------------------------------------
#  mpi defs
#--------------------------------------------------------------------------
 CPP      = /usr/bin/cpp
 CPPFLAGS = $(DEF_FLAGS)
 FC       = mpif90
 DEBFLGS  =
 OPT      =
 CLIB     =
#==========================================================================

 FFLAGS = $(DEBFLGS) $(OPT)
 MDEPFLAGS = --cpp --fext=f90 --file=-
 RANLIB = ranlib
#--------------------------------------------------------------------------
#  CAT Preprocessing Flags
#--------------------------------------------------------------------------
   CPPARGS = $(CPPFLAGS) $(DEF_FLAGS) $(FLAG_1) $(FLAG_2) \
             $(FLAG_3) $(FLAG_4) $(FLAG_5) $(FLAG_6) \
             $(FLAG_7) $(FLAG_8) $(FLAG_9) $(FLAG_10)  \
             $(FLAG_11) $(FLAG_12) $(FLAG_13) $(FLAG_14) \
             $(FLAG_15) $(FLAG_16) $(FLAG_17) $(FLAG_18) \
             $(FLAG_19) $(FLAG_20) $(FLAG_21) $(FLAG_22) \
             $(FLAG_23) $(FLAG_24)
#--------------------------------------------------------------------------
#  Libraries
#--------------------------------------------------------------------------

#            LIBS  = $(PV3LIB) $(CLIB)  $(PARLIB) $(IOLIBS) $(MPILIB) $(GOTMLIB)
#            INCS  = $(IOINCS) $(GOTMINCS)


#--------------------------------------------------------------------------
#  Preprocessing and Compilation Directives
#--------------------------------------------------------------------------
.SUFFIXES: .o .f90 .F .F90

.F.o:
        $(CPP) $(CPPARGS) $*.F > $*.f90
        $(FC)  -c $(FFLAGS) $(INCS) $*.f90
        /bin/rm $*.f90
#--------------------------------------------------------------------------
#  FUNWAVE-TVD Source Code.
#--------------------------------------------------------------------------

MODS  = mod_param.F mod_global.F mod_input.F mod_vessel.F mod_bathy_correction.F \
        mod_meteo.F mod_parallel_field_io.F

MAIN  = main.F bc.F fluxes.F init.F io.F tridiagnal.F       \
        breaker.F derivatives.F dispersion.F etauv_solver.F \
        sponge.F sources.F masks.F parallel.F statistics.F \
        wavemaker.F mixing.F nesting.F misc.F samples.F\

SRCS = $(MODS)  $(MAIN)

OBJS = $(SRCS:.F=.o)

#--------------------------------------------------------------------------
#  Linking Directives
#--------------------------------------------------------------------------

$(EXEC):        $(OBJS)
                $(FC) $(FFLAGS) $(LDFLAGS) -o $(EXEC) $(OBJS) $(LIBS)
#               mv $(EXEC) ../bin/
#--------------------------------------------------------------------------

#--------------------------------------------------------------------------
#  Tar Up Code
#--------------------------------------------------------------------------

tarfile:
        tar cvf funwave_tvd.tar *.F  Makefile

#--------------------------------------------------------------------------
#  Cleaning targets.
#--------------------------------------------------------------------------

clean:
                /bin/rm -f *.o *.mod

clobber:        clean
                /bin/rm -f *.f90 *.o $(EXEC)

Compile the code

The command make, by default, will look for a file named “Makefile” in the current directory. It will read the contents of the file to find the target program or project that needs to be built (i.e., compiled). Generally, once a program is compiled, an executable file is generated. Any changes made to the source file (e.g., “Makefile”) will need to be re-compiled, and a new executable will need to be generated. To clean, or remove, the files generated by the makefile and create new ones, the make clean command is used prior to compiling with make.

In your terminal, navigate to the directory /src/ containing the file “Makefile”, and type the following:
```
make clean

make
```
The executable file such as funwave_vessel or mytvd (specified in “Makefile”, EXEC = executable_name) will be generated in the /src/ folder. Note: always use make clean after modifying the “Makefile”.

If you want to use a specific “Makefile” under a different name (e.g., “makefile-svg”), the command make -f "makefile-svg" is used.
Run the model

Modify “input.txt” as needed. An example “input.txt” file is presented on the Definitions of Parameters page where you can review the available input options. Review the required variables for specific modules (e.g., shipwakes) on the Examples page.

To run FUNWAVE locally, navigate to the directory containing the input files needed for your simulation and type the following command in your terminal:
```
mpirun -np 4 /src/funwave_vessel
```
Here, the parallel processing MPI is established and the number of processors, -np is set to 4. This number will change depending on the capacity of your local machine. If your computer uses an Intel processor, visit this webpage to assist with identifying the number of processors available to you. Next, the path to the FUNWAVE exectuable funwave_vessel is called upon, and the model begins the simulation.

If you are running FUNWAVE in an HPC environment, you will need to review the requirements for submitting jobs in that respective environment (e.g., preparing a PBS script).

Download simple examples¶

Simple examples are included in the package of Version 3.1 and higher (click here to download from GitHub) . They are located in the directory /simple_cases/. The simple examples serve as baseline cases for testing your system. You can also choose a simple case similar to your modeling scenario to set up your case. A brief list of examples are listed below; the full list of simple cases can be found here. These simple examples are also used during the FUNWAVE training workshop.

Download benchmark tests¶

Benchmark tests are validation and verification (V&V) cases with model comparisons with lab or field experiment data. Available benchmarks: click here to download from GitHub