dist_fn Module

A type to record information about connections between iglo points Store the iglo and iproc indices for connections to the left and right of a given point. Initialised to no connections, indicated by negative values

Components

Used to represent the input configuration of dist_fn

Components

Type-Bound Procedures

Used to represent the input configuration of source

Components

Type-Bound Procedures

Used to represent the input configuration of dist_fn_species. There should be one of this namelist for each species simulated.

Components

Type-Bound Procedures

Calculates the value of the magnetic drifts (curvature + grad-B) corresponding to the theta grid location given by ig for the iglo index of interest.

Arguments

Return Value real

Calculates the value of the Coriolis drift at the location corresponding to the passed indices.

Arguments

Return Value real

Compute the spacing in kx indices between connected kx domains (cells) for the lowest non-zonal ky.

Arguments

Return Value integer

Helper function for finding the leftmost it of supercell

Arguments

Return Value integer

Helper function for finding the rightmost it of supercell

Arguments

Return Value integer

Returns true if we're using linked boundary conditions and false otherwise

Arguments

Return Value logical

Gets the default name for this namelist

Arguments

Return Value character(len=CONFIG_MAX_NAME_LEN)

Gets the default requires index for this namelist

Arguments

Return Value logical

Get the module level config instance

Arguments

Return Value type(dist_fn_config_type)

Gets the default name for this namelist

Arguments

Return Value character(len=CONFIG_MAX_NAME_LEN)

Gets the default requires index for this namelist

Arguments

Return Value logical

Get the module level config instance

Arguments

Return Value type(source_config_type)

Gets the default name for this namelist

Arguments

Return Value character(len=CONFIG_MAX_NAME_LEN)

Gets the default requires index for this namelist

Arguments

Return Value logical

Get the array of module level config instances. If it isn't allocated, then return a zero-length array

Arguments

Return Value type(dist_fn_species_config_type), allocatable, dimension(:)

FIXME : Add documentation

Arguments

FIXME : Add documentation

Arguments

FIXME : Add documentation

Arguments

FIXME : Add documentation

Arguments

FIXME : Add documentation

Arguments

FIXME : Add documentation

Arguments

FIXME : Add documentation

Arguments

FIXME : Add documentation

Arguments

FIXME : Add documentation

Arguments

FIXME : Add documentation

Arguments

FIXME : Add documentation

Arguments

Arguments

Arguments

FIXME : Add documentation

Arguments

FIXME : Add documentation

Arguments

FIXME : Add documentation

Arguments

FIXME : Add documentation

Arguments

FIXME : Add documentation
CMR, 19/10/10: Override itor_over_B, if "dist_fn_knobs" parameter btor_slab ne 0 Not ideal to set geometry quantity here, but its historical!

Arguments

FIXME : Add documentation

Arguments

FIXME : Add documentation

Arguments

FIXME : Add documentation

Arguments

Initialised the on grid v|| (vpa) and grid-centred v|| (vpac) arrays

Arguments

FIXME : Add documentation

Arguments

FIXME : Add documentation

Arguments

Compute and store the Bessel functions required for future usage, aj0 and aj1.

Arguments

FIXME : Add documentation

Arguments

Calculates the homogeneous solution g_h used as a part of the trapped particle and linked boundary condition solve in invert_rhs. This solution only depends on r so can be calculated during initialisation. Note that r depends on physics terms that contain dt so we must recalculate the solution g_h if the timestep changes. !!!!!!!!!!!!!!!!!!!!!!!!! time advance vpar < 0 ! !!!!!!!!!!!!!!!!!!!!!!!!! time advance vpar < 0 homogeneous part: g_h !!!!!!!!!!!!!!!!!!!!!!!!! time advance vpar > 0 ! !!!!!!!!!!!!!!!!!!!!!!!!! First set BCs for trapped particles at lower bounce point

Arguments

FIXME : Add documentation

Arguments

FIXME : Add documentation

Arguments

Extracts the field like data corresponding to the supercell with the {it, ik} wavenumber pair as a member and constructs the ballooning space form. By default includes duplicate theta points but can drop these on request.

Arguments

Assigns a label to each {it, ik} point denoting which unique supercell the point belongs to.

Arguments

For the passed jshift0 value determine the it (kx) indices to the left and right of each {it, ik} index pair. Used to construct the connected boundary conditions.

Arguments

Look up and store the iglo index and responsible processor for connections to the left and right of each local iglo index. Note this is only interested in passing particles and the non-trapped wfb. Trapped particles are considered to have no connections.

Arguments

Set the save_h flag based on connections data

Arguments

Count how many links are to the left/right of each {it,ik} point

Arguments

Count the links to the left and right of each {it, ik} point and work out how many values are required to set the bc.

Arguments

Count the number of unique supercell sizes and setup the fields implicit arrays M_class and N_class.

Arguments

Populates the redistribute types describing the communication pattern required to deal with linked boundary conditions. Only deals with passing particles, including the wfb if passing_wfb is true. DD>Note the communications setup here are often equivalent to an all-to-all type

Arguments

Populates the redistribute types describing the communication pattern required to deal with linked boundary conditions for the mixed wfb. This is used when passing_wfb = trapped_wfb = false. This is the default.

Arguments

Setup the redistribute associated with passing the incoming boundary values to the previous connected cell, storing at its connected boundary. This is required in order to allow for non-zero incoming boundary conditions on internal cells.

Arguments

Coordinates the calculation of all data required for later application of linked boundary conditions. Deals with identifying the communication pattern and creating the associated redistributes used to perform the required communication.

Arguments

Copy of CMR's init_pass_right (revision 2085) but in customisable direction to aid code reuse and with optional range in theta (i.e. so can pass more than just boundary points) An example use is: call init_pass_ends(pass_right,'r',1,'c',3) call fill(pass_right,gnew,gnew) This will pass gnew(ntgrid-2:ntgrid,1,iglo) to gnew(-ntgrid-2:-ntgrid,1,iglo_linked)

Arguments

FIXME : Add documentation

Arguments

FIXME : Add documentation

Arguments

FIXME : Add documentation

Arguments

This function calculates the distribution function at the next timestep. It calculates both the inhomogeneous part, gnew, due to the sources (principly the drive terms and the nonlinear term) and the homogeneous part, g1. The actual evolution of the dist func is done in the subroutine invert_rhs.

Arguments

Advance collision term one timestep by implicit Euler.

Arguments

FIXME : Add documentation

Arguments

Subroutine to setup a redistribute object to be used in enforcing parity

Arguments

Return the iglo corresponding to the part of the domain given by iglo reflected in theta=theta0

Arguments

Enforce requested parity

Arguments

FIXME : Add documentation !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Default choice: solve self-consistent equations !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Solve self-consistent terms + include external i omega_d * F_0 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Do matrix multiplications !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! special source term for totally trapped particles !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! CMR, 13/10/2014: Upper limit of following loops setting source changed from ntgrid to ntgrid-1 Source is allocated as: source(-ntgrid:ntgrid-1), so ntgrid is out of bounds.

Arguments

This is a version of get_source_term which does both sign (sigma) together and uses precalculated constant terms. Leads to more memory usage than original version but can be significantly faster (~50%)

Arguments

Calculates the regular source term on the entire local domain and stores in the output source.

Arguments

Calculates the "new" distribution function for the passed source calculated using the passed fields and current distribution in g.

Arguments

Apply one of the "end of the grid" parallel boundary conditions

Arguments

Sets sum of homogeneous and inhomogeneous solutions to give a result gnew(ntgrid,2) = gnew(-ntgrid,2) gnew(ntgrid,1) = gnew(-ntgrid,1) ie periodic bcs at the ends of the flux tube.

Arguments

A shorthand for apply_linked_boundary_conditions passing the regular module level instances of each of the required arguments. Primarily intended for testing to avoid having to expose additional data arrays.

Arguments

Applies linked (twist-shift) boundary conditions to passed distribution function DD>Note these fill routines are often equivalent to an all-to-all type

Arguments

FIXME : Add documentation

Arguments

Ensure that linked boundary values of passed complex field are single valued (e.g. kperp2(ntgrid,ikx,iky) is equal to kperp2(-ntgrid,ikx_link,iky) as these correspond to the same location).

Arguments

Ensure that linked boundary values of passed real field are single valued (e.g. kperp2(ntgrid,ikx,iky) is equal to kperp2(-ntgrid,ikx_link,iky) as these correspond to the same location).

Arguments

Compute velocity space integrals over $h_s$: $$\mathrm{antot} = n_s q_s \int d^3 v f_0 J_0(Z_s) h_s$$ $$\mathrm{antota} = 2 \beta n_s q_s \sqrt{\tfrac{T_s}{m_s}} \int d^3 v f_0 J_0(Z_s) h_s$$ $$\mathrm{antotp} = n_s T_s \int d^3 v f_0 \frac{v_\perp^2 J_1(Z_s)}{Z_s} h_s$$

Arguments

Compute velocity space integrals over $h_s$: $$\mathrm{antot} = n_s q_s \int d^3 v f_0 J_0(Z_s) h_s$$ $$\mathrm{antota} = 2 \beta n_s q_s \sqrt{\tfrac{T_s}{m_s}} \int d^3 v f_0 J_0(Z_s) h_s$$ $$\mathrm{antotp} = n_s T_s \int d^3 v f_0 \frac{v_\perp^2 J_1(Z_s)}{Z_s} h_s$$

Arguments

Check if the passed distribution function shaped array has the same value at the duplicate theta points arising from the linked boundary conditions.

Arguments

FIXME : Add documentation
DD>Make sure gamtots are single valued

Arguments

FIXME : Add documentation

Arguments

FIXME : Add documentation

Arguments

FIXME : Add documentation

Arguments

FIXME : Add documentation

Arguments

FIXME : Add documentation

Arguments

Getan_nogath has been substantially changed to enable gf_lo field calculations, it can now operate in 2 modes; using the standard integrate_species (integrate_species_sub in le_grids) when called from fields_local and gf_lo_integrate is false (which is teh default, or doing the integrate in place when called from fields_gf_local or from fields_local when gf_lo_integrate is true. Note, if this is called from fields_local and gf_lo_integrate is true then the calculation will be done locally but there is a function called after this has finished in fields_local that sends the data back (from gf_lo to g_lo layouts). When call with gf_lo_integrate = .true. this routine does a gather that converts the gnew array from g_lo data distribution to gf_lo data distribution and stores the result in gfarray. With gf_lo_integrate = .false. g_in is used instead and this is in g_lo data distribution.

Arguments

Calls getan_nogath_from_dfn to calculate antot arrays for current gnew value

Arguments

Calculates the potentials consistent with the passed nonadiabatic distribution function, f_in. Note this is not what GS2 usually evolves as g/gnew but the adjusted version.

Arguments

Precomputes constants used in the inversion of the field equations: gamtot * phi - gamtot1 * bpar = antot kperp2 * apar = antota beta/2 * gamtot1 * phi + (beta * gamtot2 + 1) * bpar = - beta * antotp

Arguments

Inverts the field equations: gamtot * phi - gamtot1 * bpar = antot kperp2 * apar = antota beta/2 * gamtot1 * phi + (beta * gamtot2 + 1) * bpar = - beta * antotp

Arguments

Calls get_fields_direct_from_dfn to solve for fields consistent with current gnew

Arguments

Calculates the potentials directly from the passed distribution function and compares the result to the passed potentials. Returns the maximum relative error for each potential.

Arguments

Routine to dump the current source term used in invert_rhs.

Arguments

Routine to dump the homogeneous solution

Arguments

FIXME : Add documentation

Arguments

FIXME : Add documentation

Arguments

Calculate the flux surface average term for the adiabatic response.

Arguments

Set the module level config types Will abort if the module has already been initialised to avoid inconsistencies.

Arguments

Reads in the dist_fn_knobs namelist and populates the member variables

Arguments

Writes out a namelist representing the current state of the config object

Arguments

Resets the config object to the initial empty state

Arguments

Broadcasts all config parameters so object is populated identically on all processors

Arguments

Reads in the source_knobs namelist and populates the member variables

Arguments

Writes out a namelist representing the current state of the config object

Arguments

Resets the config object to the initial empty state

Arguments

Broadcasts all config parameters so object is populated identically on all processors

Arguments

Reads in the dist_fn_species_knobs namelist and populates the member variables

Arguments

Writes out a namelist representing the current state of the config object

Arguments

Resets the config object to the initial empty state

Arguments

Broadcasts all config parameters so object is populated identically on all processors

Arguments