SPINS User Guide

Welcome to the SPINS case setup page.

Getting SPINS and compiling

(Chris: Is it a good idea to put it on GitHub?)

SPINS consists of a bunch of C++ source files and a bunch of case files, and it requires four libraries. UMFPack, AMD and Blitz++ are supplied with SPINS, and it uses the system-installed FFTW.

Directory structure:

• cpp/src - SPINS source files
• cpp/src/cases - A few dozen example case files
• cpp/AMD - AMD library
• cpp/UMFPack - UMFPack library
• cpp/UFconfig - Helper for compiling AMD/UMFPack

How To: Install SPINS on a new machine 1. Get the code from git - Create a directory in which to store the code. - In that directory type "git clone http://belize.math.uwaterloo.ca/~csubich/spins". - This will create a directory called spins. 2. Make the code - Go to the spins directory. - Type "./make_deps.sh". - Go to the systems directory. - Type "./makemake.sh". This will build the code based on the system architecture. Note that there are several files that include instructions for several systems e.g. gpc.sh. makemake.sh will try to guess which instructions to use if you do not include options in your makemake.sh call from the command line.

How To: Compile - Enter the src directory. - Type "make cases/your_case.x". - This requires a file called your_case.cpp in the cases directory. There are several cases included with the code so you may one to start with one of those. - After successful compilation, an executable called your_case.x is created.

How To: Run the code - Please be careful not to run in your home directory on machines like belize or winisk. - The code can be ran using mpriun e.g "mpirun -np 4 ./your_case.x" - Please note that some cases may require command line options or a configure file called spins.conf.

The basics

The SPINS model is a Navier-Stokes solver that gets parameters and initial/boundary conditions from calls to user-provided routines. The user-provided routines are encapsulated in class derived from BaseCase (see BaseCase.hpp).

Creating your own custom configuration involves supplying the user-provided routines in a derived class based on BaseCase. The case file cases/doc_minimal.cpp shows the structure of a case file. It usually makes sense to start with a similar case file and customise it.

Calling sequence

Chris: can you verify this?

To help understand where/when to do certain operations, it's good to know the order that the user routines are called. The order is as follows:

• The constructor
• The type_* routines
• The size_* and and length_* routines
• The tracer number routines (numActive, numPassive, numtracers)
• is_mapped and do_mapping
• init_time
• init_vels
• init_tracer (single tracer) or init_tracers (multiple tracers)

At each time step, the following routines are called

• forcing:
  • If no active tracers, only passive_forcing is called
  • If active tracers, first vel_forcing, then tracer_forcing
• analysis

Examples of common operations

You can find examples of how to do various operations by digging through the case files. Some of the common operations are reproduced here.

Grid

Generating the grid vectors

At the top of your file include the global tensor indices:

// Tensor variables for indexing
blitz::firstIndex ii;
blitz::secondIndex jj;
blitz::thirdIndex kk;

Define arrays for the grid vectors:

// Arrays to hold the x, y, and z points
Array<double,1> xx, yy, zz;

You can initialise them in the constructor. For Periodic/free-slip the grid spacing is even:

// The constructor
casename() : xx(split_range(Nx)), yy(Ny), zz(Nz) {
    // Assumes periodic or free-slip (cosine)
    xx = Lx*(ii+0.5)/Nx; // x-coordinate: periodic
    yy = Ly*(ii+0.5)/Ny; // y-coordinate: periodic
    zz = Lz*(ii+0.5)/Nz; // z-coordinate: free-slip
}

For no-slip (Chebyshev) the grid spacing is not even:

// The constructor
casename() : xx(split_range(Nx)), yy(Ny), zz(Nz) {
    // Assumes no-clip (Chebychev) in all dimensions
    xx = -Lx*cos(ii*M_PI/(Nx-1))/2; // x-coordinate: Chebyshev
    yy = -Ly*cos(ii*M_PI/(Ny-1))/2; // y-coordinate: Chebyshev
    zz = -Lz*cos(ii*M_PI/(Nz-1))/2; // z-coordinate: Chebyshev
}

Outputting the grid

Create array tmp, then generate each grid and write the array and the grid reader out:

// Compute the grids and write them out
DTArray *gridtmp = alloc_array(Nx,Ny,Nz);

*gridtmp = xx(ii) + 0*jj + 0*kk;
write_array(*gridtmp,"xgrid");
write_reader(*gridtmp,"xgrid",false);

*gridtmp = 0*ii + yy(jj) + 0*kk;
write_array(*gridtmp,"ygrid");
write_reader(*gridtmp,"ygrid",false);

*gridtmp = 0*ii + 0*jj + zz(kk);
write_array(*gridtmp,"zgrid");
write_reader(*gridtmp,"zgrid",false);

delete(gridtmp);

Mapped grid

To Do: write this!

Initialisation

Velocities

To Do: write this!

Tracers

To Do: write this!

Forcing

To Do: write this!

Velocities

To Do: write this!

Tracers

To Do: write this!

Analysis

Energy Diagnostic

If you're using a periodic grid, use this for computing kinetic energy diagnostic

double ke = 0.5*rho_0*pssum(sum(u*u+v*v+w*w))/(Nx*Ny*Nz)*Lx*Ly*Lz; // KE
if (master()) {
  FILE * en_record = fopen("energy_record.txt","a");
  assert(en_record);
  fprintf(en_record,"%.8f %.14g\n",time,ke);
  fclose(en_record);
}

If you're on a Chebyshev grid, you can use this for the KE computation

double ke = pssum(sum((*get_quad_x())(ii)*(*get_quad_y())(jj)*(*get_quad_z())(kk)*
                  (pow(u(ii,jj,kk),2)+pow(v(ii,jj,kk),2)+pow(w(ii,jj,kk),2))));

and you will need to compute the quadrature weights in the constructor

// Compute the quadrature weights
compute_quadweights(size_x(),size_y(),size_z(),
                    length_x(),length_y(),length_z(),
                    type_x(),type_y(),type_z());

If you're on a mapped grid, To Do: write this!

Saving snapshots

To Do: write this!

Computing vorticity

To Do: write this!