adam_flume_fnl_euler_kernels

ADAM, FLUME FNL device kernels of the Euler model.

The model-agnostic kernel bodies (adam_flume_fnl_{face,aux}_kernels_agnostic.INC) instantiated on the Euler physics: the private arrays are sized by the Euler constants NV_K = NV_EULER, NV_AUX_K = NV_AUX (issue #41, section 4), plus the Euler kernels whose body depends on the variables (conservation integrals, signal speed). Same kernel rules as adam_flume_fnl_kernels (issue #35, D-11/D-12).

Source: src/app/flume/fnl/adam_flume_fnl_euler_kernels.F90

Dependencies

Contents

• compute_face_fluxes_dev
• add_gravity_edge_correction_dev
• compute_q_aux_dev
• count_nonfinite_dev
• compute_positivity_factors_dev
• blend_positivity_fluxes_dev
• compute_seam_positivity_factors_host
• compute_backbone_fluxes_host
• positivity_cell_factor
• blend_inadmissible_ghosts_dev
• compute_conservation_dev
• compute_lambda_max_dev
• compute_lambda_max_dissipative_dev
• face_split_fluxes
• backbone_face_flux
• cell_sources
• internal_energy

Subroutines

compute_face_fluxes_dev

Compute the WENO face fluxes of direction d: face (i,j,k) lies between cells (i,j,k) and (i+di,j+dj,k+dk), so the face array starts at index 0 along d.

Per face (device twin of the CPU compute_face_fluxes): gather the stencil m = 1-S ... S into constant-bound privates, project and split it, reconstruct every field (the private v is the packed stencil of the WENO primitive), back-project.

subroutine compute_face_fluxes_dev(d, di, dj, dk, ni, nj, nk, ngc, blocks_number, S, gamma, ch, is_characteristic, weno_a_gpu, weno_p_gpu, weno_d_gpu, weno_zeps, weno_sigma, q_gpu, q_aux_gpu, fl_gpu)

Arguments

Name	Type	Intent	Attributes	Description
d	integer(kind=I4P)	in		Direction, 1=x, 2=y, 3=z.
di	integer(kind=I4P)	in		Unit step along d.
dj	integer(kind=I4P)	in		Unit step along d.
dk	integer(kind=I4P)	in		Unit step along d.
ni	integer(kind=I4P)	in		Grid dimensions.
nj	integer(kind=I4P)	in		Grid dimensions.
nk	integer(kind=I4P)	in		Grid dimensions.
ngc	integer(kind=I4P)	in		Grid dimensions.
blocks_number	integer(kind=I4P)	in		Actual blocks number.
S	integer(kind=I4P)	in		WENO stencil half-width.
gamma	real(kind=R8P)	in		Specific heats ratio.
ch	real(kind=R8P)	in		GLM cleaning speed (models with GLM only).
is_characteristic	logical	in		Characteristic (or conservative) variables.
weno_a_gpu	real(kind=R8P)	in		WENO optimal weights.
weno_p_gpu	real(kind=R8P)	in		WENO polynomials coefficients.
weno_d_gpu	real(kind=R8P)	in		WENO smoothness indicators coefficients.
weno_zeps	real(kind=R8P)	in		WENO parameter avoiding division by zero.
weno_sigma	real(kind=R8P)	in		WENO descaler switch (0 Jiang-Shu, 1 SI).
q_gpu	real(kind=R8P)	in		Conservative variables.
q_aux_gpu	real(kind=R8P)	in		Auxiliary variables.
fl_gpu	real(kind=R8P)	inout		Face fluxes of direction d.

Call graph

add_gravity_edge_correction_dev

Add the faces beyond the block to the gravity energy correction of direction d (issue #73, M7-G2a; device twin of the CPU add_gravity_edge_correction): cell 1 takes -g_d F^rho / 24 of face -1, cell n_d that of face n_d + 1, each face computed as compute_face_fluxes_dev does (the same body, its mass component kept).

subroutine add_gravity_edge_correction_dev(d, di, dj, dk, ni, nj, nk, ngc, blocks_number, S, gamma, ch, is_characteristic, weno_a_gpu, weno_p_gpu, weno_d_gpu, weno_zeps, weno_sigma, ir, gd, q_gpu, q_aux_gpu, gc_gpu)

Arguments

Name	Type	Intent	Attributes	Description
d	integer(kind=I4P)	in		Direction, 1=x, 2=y, 3=z.
di	integer(kind=I4P)	in		Unit step along d.
dj	integer(kind=I4P)	in		Unit step along d.
dk	integer(kind=I4P)	in		Unit step along d.
ni	integer(kind=I4P)	in		Grid dimensions.
nj	integer(kind=I4P)	in		Grid dimensions.
nk	integer(kind=I4P)	in		Grid dimensions.
ngc	integer(kind=I4P)	in		Grid dimensions.
blocks_number	integer(kind=I4P)	in		Actual blocks number.
S	integer(kind=I4P)	in		WENO stencil half-width.
gamma	real(kind=R8P)	in		Specific heats ratio.
ch	real(kind=R8P)	in		GLM cleaning speed (models with GLM only).
is_characteristic	logical	in		Characteristic (or conservative) variables.
weno_a_gpu	real(kind=R8P)	in		WENO optimal weights.
weno_p_gpu	real(kind=R8P)	in		WENO polynomials coefficients.
weno_d_gpu	real(kind=R8P)	in		WENO smoothness indicators coefficients.
weno_zeps	real(kind=R8P)	in		WENO parameter avoiding division by zero.
weno_sigma	real(kind=R8P)	in		WENO descaler switch (0 Jiang-Shu, 1 SI).
ir	integer(kind=I4P)	in		Index of the density (mass) variable.
gd	real(kind=R8P)	in		Gravity along d.
q_gpu	real(kind=R8P)	in		Conservative variables.
q_aux_gpu	real(kind=R8P)	in		Auxiliary variables.
gc_gpu	real(kind=R8P)	inout		Gravity energy correction [..., 1].

Call graph

compute_q_aux_dev

Compute the auxiliary variables on every cell, ghost cells included.

subroutine compute_q_aux_dev(ni, nj, nk, ngc, blocks_number, gamma, R, q_gpu, q_aux_gpu)

Arguments

Name	Type	Intent	Attributes	Description
ni	integer(kind=I4P)	in		Grid dimensions.
nj	integer(kind=I4P)	in		Grid dimensions.
nk	integer(kind=I4P)	in		Grid dimensions.
ngc	integer(kind=I4P)	in		Grid dimensions.
blocks_number	integer(kind=I4P)	in		Actual blocks number.
gamma	real(kind=R8P)	in		Specific heats ratio.
R	real(kind=R8P)	in		Gas constant.
q_gpu	real(kind=R8P)	in		Conservative variables.
q_aux_gpu	real(kind=R8P)	inout		Auxiliary variables.

Call graph

count_nonfinite_dev

Count the non-finite (NaN or infinite) conservative values of the interior cells (issue #45): the exponent field has all its bits set. Tested on the bit pattern, so that no floating point mode (-fast) can fold the test away.

subroutine count_nonfinite_dev(ni, nj, nk, ngc, blocks_number, q_gpu, n)

Arguments

Name	Type	Intent	Attributes	Description
ni	integer(kind=I4P)	in		Grid dimensions.
nj	integer(kind=I4P)	in		Grid dimensions.
nk	integer(kind=I4P)	in		Grid dimensions.
ngc	integer(kind=I4P)	in		Grid dimensions.
blocks_number	integer(kind=I4P)	in		Actual blocks number.
q_gpu	real(kind=R8P)	in		Conservative variables.
n	integer(kind=I8P)	out		Non-finite values number.

Call graph

compute_positivity_factors_dev

Compute the cell factor Lambda of every interior cell into lam_gpu(:,:,:,:,1) and count the cells whose backbone is not admissible and those with a non-finite high-order face flux (device twin of the CPU compute_positivity_factors).

subroutine compute_positivity_factors_dev(ni, nj, nk, ngc, blocks_number, gamma, ch, damping, hs, dt, g, has_phi, phi_gpu, dxyz_gpu, is_null, q_gpu, q_aux_gpu, flx_gpu, fly_gpu, flz_gpu, gc_gpu, lam_gpu, bad, nonfinite)

Arguments

Name	Type	Intent	Attributes	Description
ni	integer(kind=I4P)	in		Grid dimensions.
nj	integer(kind=I4P)	in		Grid dimensions.
nk	integer(kind=I4P)	in		Grid dimensions.
ngc	integer(kind=I4P)	in		Grid dimensions.
blocks_number	integer(kind=I4P)	in		Actual blocks number.
gamma	real(kind=R8P)	in		Specific heats ratio.
ch	real(kind=R8P)	in		GLM cleaning speed (EGLM only).
damping	real(kind=R8P)	in		GLM damping rate (EGLM only).
hs	integer(kind=I4P)	in		Half stencil of the sources, 1 <= hs <= S_MAX.
dt	real(kind=R8P)	in		Forward-Euler step of the stage.
g	real(kind=R8P)	in		Gravity (issue #73), zero without it.
has_phi	logical	in		The cell potential is applied (issue #81).
phi_gpu	real(kind=R8P)	in		Gravitational potential (read with has_phi).
dxyz_gpu	real(kind=R8P)	in		Blocks space steps [nb, 3].
is_null	logical	in		Null directions.
q_gpu	real(kind=R8P)	in		Conservative variables (ghosts updated).
q_aux_gpu	real(kind=R8P)	in		Auxiliary variables.
flx_gpu	real(kind=R8P)	in		X-face high-order fluxes.
fly_gpu	real(kind=R8P)	in		Y-face high-order fluxes.
flz_gpu	real(kind=R8P)	in		Z-face high-order fluxes.
gc_gpu	real(kind=R8P)	in		Gravity energy correction (zero without it).
lam_gpu	real(kind=R8P)	inout		Cell factors (component 1).
bad	integer(kind=I4P)	out		Cells with an inadmissible backbone.
nonfinite	integer(kind=I4P)	out		Cells with a non-finite high-order face flux.

Call graph

blend_positivity_fluxes_dev

Blend the face fluxes of direction d with the backbone, theta the smaller factor of the two cells; faces with theta = 1 untouched, faces with theta = 0 set to the backbone flux (device twin of the CPU blend_positivity_fluxes).

subroutine blend_positivity_fluxes_dev(d, di, dj, dk, ni, nj, nk, ngc, blocks_number, gamma, ch, q_gpu, lam_gpu, fl_gpu, limited)

Arguments

Name	Type	Intent	Attributes	Description
d	integer(kind=I4P)	in		Direction, 1=x, 2=y, 3=z.
di	integer(kind=I4P)	in		Unit step along d.
dj	integer(kind=I4P)	in		Unit step along d.
dk	integer(kind=I4P)	in		Unit step along d.
ni	integer(kind=I4P)	in		Grid dimensions.
nj	integer(kind=I4P)	in		Grid dimensions.
nk	integer(kind=I4P)	in		Grid dimensions.
ngc	integer(kind=I4P)	in		Grid dimensions.
blocks_number	integer(kind=I4P)	in		Actual blocks number.
gamma	real(kind=R8P)	in		Specific heats ratio.
ch	real(kind=R8P)	in		GLM cleaning speed (EGLM only).
q_gpu	real(kind=R8P)	in		Conservative variables.
lam_gpu	real(kind=R8P)	in		Cell factors (component 1).
fl_gpu	real(kind=R8P)	inout		Face fluxes of direction d.
limited	integer(kind=I4P)	out		Faces with theta < 1.

Call graph

compute_seam_positivity_factors_host

Recompute on the host the factor of the cells beside a 2:1 seam with the seam faces' backbone and high-order fluxes overridden (issue #50, D1; the host twin of the CPU compute_seam_positivity_factors, fed by the device gathers of the seam cells: their axis stencils qs(:, l + hs + 1, d, n), auxiliary variables qa(:, n) and face fluxes fh(:, fec, n), fec 1..6 = -x, +x, -y, +y, -z, +z). dbad, dnonfinite: the change of the counts against the factors without the overrides, which the device kernel has already counted.

subroutine compute_seam_positivity_factors_host(ncells, hs, gamma, ch, damping, dt, g, has_phi, ps, ds, is_null, qs, qa, fh, omask, gcs, lsrc, olo, ohi, lam, dbad, dnonfinite)

Arguments

Name	Type	Intent	Attributes	Description
ncells	integer(kind=I4P)	in		Seam cells.
hs	integer(kind=I4P)	in		Stencil half width, 1 <= hs <= S_MAX.
gamma	real(kind=R8P)	in		Specific heats ratio.
ch	real(kind=R8P)	in		GLM cleaning speed (EGLM only).
damping	real(kind=R8P)	in		GLM damping rate (EGLM only).
dt	real(kind=R8P)	in		Forward-Euler step of the stage.
g	real(kind=R8P)	in		Gravity (issue #73), zero without it.
has_phi	logical	in		The cell potential is applied (issue #81).
ps	real(kind=R8P)	in		Potential axis stencils [1, 3, 3, ncells] (with has_phi).
ds	real(kind=R8P)	in		Space steps of the cells' blocks [3, ncells].
is_null	logical	in		Null directions.
qs	real(kind=R8P)	in		Axis stencils [nv, 2 hs + 1, 3, ncells].
qa	real(kind=R8P)	in		Auxiliary variables [nv_aux, ncells].
fh	real(kind=R8P)	in		High-order face fluxes [nv, 6, ncells].
omask	logical	in		Overridden faces [6, ncells].
gcs	real(kind=R8P)	in		Gravity energy correction [ncells] (M7-G2a).
lsrc	real(kind=R8P)	out		Gravity face-form excess L [ncells] (issue #73).
olo	real(kind=R8P)	in		Override backbone fluxes [nv, 6, ncells].
ohi	real(kind=R8P)	in		Override high-order fluxes [nv, 6, ncells].
lam	real(kind=R8P)	out		Cell factors [ncells].
dbad	integer(kind=I4P)	out		Change of the inadmissible backbones count.
dnonfinite	integer(kind=I4P)	out		Change of the non-finite cells count.

Call graph

compute_backbone_fluxes_host

Compute on the host the backbone flux of n face state pairs of direction d (issue #50: the fine faces of a 2:1 seam take the coarse donor cell as their outer state).

subroutine compute_backbone_fluxes_host(n, d, gamma, ch, qL, qR, flo)

Arguments

Name	Type	Intent	Attributes	Description
n	integer(kind=I4P)	in		Faces number.
d	integer(kind=I4P)	in		Direction, 1=x, 2=y, 3=z.
gamma	real(kind=R8P)	in		Specific heats ratio.
ch	real(kind=R8P)	in		GLM cleaning speed (EGLM only).
qL	real(kind=R8P)	in		Left states [nv, n].
qR	real(kind=R8P)	in		Right states [nv, n].
flo	real(kind=R8P)	out		Backbone fluxes [nv, n].

Call graph

positivity_cell_factor

The factor of one cell from its backbone update, antidiffusive differences and source difference (the decision of compute_positivity_factors_dev, for the host seam pass): status 1 inadmissible backbone, 2 non-finite difference (lc = 0 both), else 0 with lc the largest factor of the corner box.

Attributes: pure

subroutine positivity_cell_factor(nf, qb, dlt, dsrc, lc, status)

Arguments

Name	Type	Intent	Attributes	Description
nf	integer(kind=I4P)	in		Active faces.
qb	real(kind=R8P)	in		Backbone update.
dlt	real(kind=R8P)	in		Antidiffusive differences.
dsrc	real(kind=R8P)	in		Source difference.
lc	real(kind=R8P)	out		Cell factor.
status	integer(kind=I4P)	out		0 admissible, 1 inadmissible backbone, 2 non-finite.

Call graph

blend_inadmissible_ghosts_dev

Blend every inadmissible face ghost of q_gpu toward the adjacent interior cell and count the ghosts blended (device twin of the CPU blend_inadmissible_ghosts).

subroutine blend_inadmissible_ghosts_dev(ni, nj, nk, ngc, blocks_number, is_null, q_gpu, blended)

Arguments

Name	Type	Intent	Attributes	Description
ni	integer(kind=I4P)	in		Grid dimensions.
nj	integer(kind=I4P)	in		Grid dimensions.
nk	integer(kind=I4P)	in		Grid dimensions.
ngc	integer(kind=I4P)	in		Grid dimensions.
blocks_number	integer(kind=I4P)	in		Actual blocks number.
is_null	logical	in		Null directions.
q_gpu	real(kind=R8P)	inout		Conservative variables (ghosts updated).
blended	integer(kind=I4P)	out		Ghosts blended.

Call graph

compute_conservation_dev

Compute the volume integrals of the conservative variables (interior cells). The cell volume includes the null directions: the tree splits them too, so a refined block's cells are smaller along them (issue #37).

subroutine compute_conservation_dev(ni, nj, nk, ngc, blocks_number, dxyz_gpu, q_gpu, integrals)

Arguments

Name	Type	Intent	Attributes	Description
ni	integer(kind=I4P)	in		Grid dimensions.
nj	integer(kind=I4P)	in		Grid dimensions.
nk	integer(kind=I4P)	in		Grid dimensions.
ngc	integer(kind=I4P)	in		Grid dimensions.
blocks_number	integer(kind=I4P)	in		Actual blocks number.
dxyz_gpu	real(kind=R8P)	in		Blocks space steps [nb, 3].
q_gpu	real(kind=R8P)	in		Conservative variables.
integrals	real(kind=R8P)	out		Volume integrals.

Call graph

compute_lambda_max_dev

Compute max(sum_d (|u_d| + a) / dx_d) over the interior cells (null directions excluded).

subroutine compute_lambda_max_dev(ni, nj, nk, ngc, blocks_number, gamma, R, dxyz_gpu, is_null, q_gpu, lambda_max)

Arguments

Name	Type	Intent	Attributes	Description
ni	integer(kind=I4P)	in		Grid dimensions.
nj	integer(kind=I4P)	in		Grid dimensions.
nk	integer(kind=I4P)	in		Grid dimensions.
ngc	integer(kind=I4P)	in		Grid dimensions.
blocks_number	integer(kind=I4P)	in		Actual blocks number.
gamma	real(kind=R8P)	in		Specific heats ratio.
R	real(kind=R8P)	in		Gas constant.
dxyz_gpu	real(kind=R8P)	in		Blocks space steps [nb, 3].
is_null	logical	in		Null directions.
q_gpu	real(kind=R8P)	in		Conservative variables.
lambda_max	real(kind=R8P)	out		Maximum of sum_d (|u_d| + a) / dx_d.

Call graph

compute_lambda_max_dissipative_dev

Device twin of the CPU compute_lambda_max_dissipative (issue #65, D-M4-3): the maximum of the hyperbolic plus diffusive cell sum, the maxima of each part and the minimum cell Reynolds number.

subroutine compute_lambda_max_dissipative_dev(ni, nj, nk, ngc, blocks_number, gamma, R, cp, mu0, k0, eta, t_wall, tref, omega_mu, omega_k, dxyz_gpu, is_null, q_gpu, lambda_max, lambda_hyp, lambda_dif, re_cell_min)

Arguments

Name	Type	Intent	Attributes	Description
ni	integer(kind=I4P)	in		Grid dimensions.
nj	integer(kind=I4P)	in		Grid dimensions.
nk	integer(kind=I4P)	in		Grid dimensions.
ngc	integer(kind=I4P)	in		Grid dimensions.
blocks_number	integer(kind=I4P)	in		Actual blocks number.
gamma	real(kind=R8P)	in		Specific heats ratio, gas constant, cp.
R	real(kind=R8P)	in		Specific heats ratio, gas constant, cp.
cp	real(kind=R8P)	in		Specific heats ratio, gas constant, cp.
mu0	real(kind=R8P)	in		Dissipative coefficients.
k0	real(kind=R8P)	in		Dissipative coefficients.
eta	real(kind=R8P)	in		Dissipative coefficients.
t_wall	real(kind=R8P)	in		Hottest isothermal wall temperature, 0 if none.
tref	real(kind=R8P)	in		Temperature laws.
omega_mu	real(kind=R8P)	in		Temperature laws.
omega_k	real(kind=R8P)	in		Temperature laws.
dxyz_gpu	real(kind=R8P)	in		Blocks space steps [nb, 3].
is_null	logical	in		Null directions.
q_gpu	real(kind=R8P)	in		Conservative variables.
lambda_max	real(kind=R8P)	out		Maximum of the sum.
lambda_hyp	real(kind=R8P)	out		Maximum of the hyperbolic part.
lambda_dif	real(kind=R8P)	out		Maximum of the diffusive part.
re_cell_min	real(kind=R8P)	out		Minimum cell Reynolds number.

Call graph

face_split_fluxes

Split adapter of the shared face kernel (issue #41, M2-P3): the Euler split, ch unused.

Attributes: pure

subroutine face_split_fluxes(gamma, ch, d, S, is_characteristic, qs, qas, fsplit, er, mu)

Arguments

Name	Type	Intent	Attributes	Description
gamma	real(kind=R8P)	in		Specific heats ratio.
ch	real(kind=R8P)	in		GLM cleaning speed.
d	integer(kind=I4P)	in		Direction, 1=x, 2=y, 3=z.
S	integer(kind=I4P)	in		WENO stencil half-width, S <= S_MAX.
is_characteristic	logical	in		Characteristic (or conservative) variables.
qs	real(kind=R8P)	in		Stencil conservative variables.
qas	real(kind=R8P)	in		Stencil auxiliary variables.
fsplit	real(kind=R8P)	out		Split fields in the WENO upwind layout.
er	real(kind=R8P)	out		Right eigenvectors.
mu	real(kind=R8P)	out		Magnitude of each field (WENO descaler).

Call graph

backbone_face_flux

Backbone adapter of the positivity limiter: the Rusanov flux, ch unused.

Attributes: pure

subroutine backbone_face_flux(gamma, ch, d, qL, qR, f)

Arguments

Name	Type	Intent	Attributes	Description
gamma	real(kind=R8P)	in		Specific heats ratio.
ch	real(kind=R8P)	in		GLM cleaning speed.
d	integer(kind=I4P)	in		Direction, 1=x, 2=y, 3=z.
qL	real(kind=R8P)	in		Left state.
qR	real(kind=R8P)	in		Right state.
f	real(kind=R8P)	out		Flux.

Call graph

cell_sources

Sources adapter of the positivity limiter: none (Euler).

Attributes: pure

subroutine cell_sources(hs, damping, ds, w, qsx, qsy, qsz, qac, s_lo, s_hi)

Arguments

Name	Type	Intent	Attributes	Description
hs	integer(kind=I4P)	in		Half stencil of the high-order sources.
damping	real(kind=R8P)	in		GLM damping rate.
ds	real(kind=R8P)	in		Block space steps.
w	real(kind=R8P)	in		Direction weights: 1 active, 0 null.
qsx	real(kind=R8P)	in		Stencil along x.
qsy	real(kind=R8P)	in		Stencil along y.
qsz	real(kind=R8P)	in		Stencil along z.
qac	real(kind=R8P)	in		Cell auxiliary variables.
s_lo	real(kind=R8P)	out		Backbone sources.
s_hi	real(kind=R8P)	out		High-order sources.

Call graph

Functions

internal_energy

Internal-energy adapter of the positivity limiter: E minus the kinetic energy, per unit volume.

Attributes: pure

Returns: real(kind=R8P)

function internal_energy(q) result(e)

Arguments

Name	Type	Intent	Attributes	Description
q	real(kind=R8P)	in		Conservative variables.

Call graph