70 use iso_c_binding,
only : c_ptr
75 use mpi_f08,
only : mpi_allreduce, mpi_in_place, mpi_sum
96 s13, s23, p, coef, visc)
97 integer,
intent(in) :: tstep
99 type(
coef_t),
intent(inout) :: coef
100 real(kind=
rp),
intent(inout) :: s11(coef%Xh%lx, coef%Xh%ly, coef%Xh%lz, &
102 real(kind=
rp),
intent(inout) :: s22(coef%Xh%lx, coef%Xh%ly, coef%Xh%lz, &
104 real(kind=
rp),
intent(inout) :: s33(coef%Xh%lx, coef%Xh%ly, coef%Xh%lz, &
106 real(kind=
rp),
intent(inout) :: s12(coef%Xh%lx, coef%Xh%ly, coef%Xh%lz, &
108 real(kind=
rp),
intent(inout) :: s13(coef%Xh%lx, coef%Xh%ly, coef%Xh%lz, &
110 real(kind=
rp),
intent(inout) :: s23(coef%Xh%lx, coef%Xh%ly, coef%Xh%lz, &
112 type(
field_t),
intent(inout) :: p
113 real(kind=
rp),
intent(in) :: visc, center(3)
114 real(kind=
rp) :: dgtq(3,4)
115 real(kind=
rp) :: dragpx = 0.0_rp
116 real(kind=
rp) :: dragpy = 0.0_rp
117 real(kind=
rp) :: dragpz = 0.0_rp
118 real(kind=
rp) :: dragvx = 0.0_rp
119 real(kind=
rp) :: dragvy = 0.0_rp
120 real(kind=
rp) :: dragvz = 0.0_rp
121 real(kind=
rp) :: torqpx = 0.0_rp
122 real(kind=
rp) :: torqpy = 0.0_rp
123 real(kind=
rp) :: torqpz = 0.0_rp
124 real(kind=
rp) :: torqvx = 0.0_rp
125 real(kind=
rp) :: torqvy = 0.0_rp
126 real(kind=
rp) :: torqvz = 0.0_rp
127 real(kind=
rp) :: dragx, dragy, dragz
128 integer :: ie, ifc, mem, ierr
143 ie = zone%facet_el(mem)%x(2)
144 ifc = zone%facet_el(mem)%x(1)
147 s11, s22, s33, s12, s13, s23,&
148 p%x, visc, ifc, ie, coef, coef%Xh)
150 dragpx = dragpx + dgtq(1,1)
151 dragpy = dragpy + dgtq(2,1)
152 dragpz = dragpz + dgtq(3,1)
154 dragvx = dragvx + dgtq(1,2)
155 dragvy = dragvy + dgtq(2,2)
156 dragvz = dragvz + dgtq(3,2)
158 torqpx = torqpx + dgtq(1,3)
159 torqpy = torqpy + dgtq(2,3)
160 torqpz = torqpz + dgtq(3,3)
162 torqvx = torqvx + dgtq(1,4)
163 torqvy = torqvy + dgtq(2,4)
164 torqvz = torqvz + dgtq(3,4)
169 call mpi_allreduce(mpi_in_place,dragpx, 1, &
171 call mpi_allreduce(mpi_in_place,dragpy, 1, &
173 call mpi_allreduce(mpi_in_place,dragpz, 1, &
175 call mpi_allreduce(mpi_in_place,dragvx, 1, &
177 call mpi_allreduce(mpi_in_place,dragvy, 1, &
179 call mpi_allreduce(mpi_in_place,dragvz, 1, &
182 call mpi_allreduce(mpi_in_place,torqpx, 1, &
184 call mpi_allreduce(mpi_in_place,torqpy, 1, &
186 call mpi_allreduce(mpi_in_place,torqpz, 1, &
188 call mpi_allreduce(mpi_in_place,torqvx, 1, &
190 call mpi_allreduce(mpi_in_place,torqvy, 1, &
192 call mpi_allreduce(mpi_in_place,torqvz, 1, &
225 s11, s22, s33, s12, s13, s23,&
226 pm1,visc,f,e, coef, Xh)
227 type(
coef_t),
intent(in) :: coef
228 type(
space_t),
intent(in) :: xh
229 real(kind=
rp),
intent(out) :: dgtq(3,4)
230 real(kind=
rp),
intent(in) :: center(3)
231 real(kind=
rp),
intent(in) :: xm0(xh%lx,xh%ly,xh%lz,coef%msh%nelv)
232 real(kind=
rp),
intent(in) :: ym0(xh%lx,xh%ly,xh%lz,coef%msh%nelv)
233 real(kind=
rp),
intent(in) :: zm0(xh%lx,xh%ly,xh%lz,coef%msh%nelv)
234 real(kind=
rp),
intent(in) :: s11(xh%lx,xh%ly,xh%lz,coef%msh%nelv)
235 real(kind=
rp),
intent(in) :: s22(xh%lx,xh%ly,xh%lz,coef%msh%nelv)
236 real(kind=
rp),
intent(in) :: s33(xh%lx,xh%ly,xh%lz,coef%msh%nelv)
237 real(kind=
rp),
intent(in) :: s12(xh%lx,xh%ly,xh%lz,coef%msh%nelv)
238 real(kind=
rp),
intent(in) :: s13(xh%lx,xh%ly,xh%lz,coef%msh%nelv)
239 real(kind=
rp),
intent(in) :: s23(xh%lx,xh%ly,xh%lz,coef%msh%nelv)
240 real(kind=
rp),
intent(in) :: pm1(xh%lx,xh%ly,xh%lz,coef%msh%nelv)
241 real(kind=
rp),
intent(in) :: visc
242 integer,
intent(in) :: f,e
243 integer :: pf, i, j1, j2
244 real(kind=
rp) :: n1,n2,n3, a, v, dgtq_i(3,4)
245 integer :: skpdat(6,6), nx, ny, nz
252 real(kind=
rp) :: s11_, s12_, s22_, s13_, s23_, s33_
259 skpdat(2,1) = nx*(ny-1) + 1
262 skpdat(5,1) = ny*(nz-1) + 1
265 skpdat(1,2) = 1 + (nx-1)
266 skpdat(2,2) = nx*(ny-1)+1 + (nx-1)
269 skpdat(5,2) = ny*(nz-1)+1
276 skpdat(5,3) = ny*(nz-1)+1
279 skpdat(1,4) = 1 + nx*(ny-1)
280 skpdat(2,4) = nx + nx*(ny-1)
283 skpdat(5,4) = ny*(nz-1)+1
293 skpdat(1,6) = 1 + nx*ny*(nz-1)
294 skpdat(2,6) = nx + nx*ny*(nz-1)
302 jskip1 = skpdat(3,pf)
305 jskip2 = skpdat(6,pf)
309 do j2 = js2, jf2, jskip2
310 do j1 = js1, jf1, jskip1
312 n1 = coef%nx(i,1,f,e)*coef%area(i,1,f,e)
313 n2 = coef%ny(i,1,f,e)*coef%area(i,1,f,e)
314 n3 = coef%nz(i,1,f,e)*coef%area(i,1,f,e)
315 a = a + coef%area(i,1,f,e)
317 s11_ = s11(j1, j2, 1, e)
318 s12_ = s12(j1, j2, 1, e)
319 s22_ = s22(j1, j2, 1, e)
320 s13_ = s13(j1, j2, 1, e)
321 s23_ = s23(j1, j2, 1, e)
322 s33_ = s33(j1, j2, 1, e)
323 call drag_torque_pt(dgtq_i,xm0(j1, j2, 1, e), ym0(j1, j2, 1, e), &
324 zm0(j1, j2, 1, e), center,&
325 s11_, s22_, s33_, s12_, s13_, s23_,&
326 pm1(j1, j2, 1, e), n1, n2, n3, v)
344 subroutine drag_torque_pt(dgtq, x, y, z, center, s11, s22, s33, s12, s13, s23,&
346 real(kind=
rp),
intent(inout) :: dgtq(3, 4)
347 real(kind=
rp),
intent(in) :: x
348 real(kind=
rp),
intent(in) :: y
349 real(kind=
rp),
intent(in) :: z
350 real(kind=
rp),
intent(in) :: p
351 real(kind=
rp),
intent(in) :: v
352 real(kind=
rp),
intent(in) :: n1, n2, n3, center(3)
353 real(kind=
rp),
intent(in) :: s11, s12, s22, s13, s23, s33
354 real(kind=
rp) :: s21, s31, s32, r1, r2, r3
365 dgtq(1,2) = -2*v*(s11*n1 + s12*n2 + s13*n3)
366 dgtq(2,2) = -2*v*(s21*n1 + s22*n2 + s23*n3)
367 dgtq(3,2) = -2*v*(s31*n1 + s32*n2 + s33*n3)
372 dgtq(1,3) = r2*dgtq(3,1) - r3*dgtq(2,1)
373 dgtq(2,3) = r3*dgtq(1,1) - r1*dgtq(3,1)
374 dgtq(3,3) = r1*dgtq(2,1) - r2*dgtq(1,1)
376 dgtq(1,4) = r2*dgtq(3,2) - r3*dgtq(2,2)
377 dgtq(2,4) = r3*dgtq(1,2) - r1*dgtq(3,2)
378 dgtq(3,4) = r1*dgtq(2,2) - r2*dgtq(1,2)
391 s11, s22, s33, s12, s13, s23,&
392 p, n1, n2, n3, mu, n_pts)
394 real(kind=
rp),
intent(inout),
dimension(n_pts) :: force1, force2, force3
395 real(kind=
rp),
intent(inout),
dimension(n_pts) :: force4, force5, force6
396 real(kind=
rp),
intent(in) :: p(n_pts)
397 real(kind=
rp),
intent(in) :: mu(n_pts)
398 real(kind=
rp),
intent(in) :: n1(n_pts), n2(n_pts), n3(n_pts)
399 real(kind=
rp),
intent(in),
dimension(n_pts) :: s11, s12, s22, s13, s23, s33
400 real(kind=
rp) :: v2(n_pts)
401 call rzero(force4, n_pts)
402 call rzero(force5, n_pts)
403 call rzero(force6, n_pts)
405 call col3(force1, p, n1, n_pts)
406 call col3(force2, p, n2, n_pts)
407 call col3(force3, p, n3, n_pts)
410 call vdot3(force4, s11, s12, s13, n1, n2, n3, n_pts)
411 call vdot3(force5, s12, s22, s23, n1, n2, n3, n_pts)
412 call vdot3(force6, s13, s23, s33, n1, n2, n3, n_pts)
413 call col2(force4, v2, n_pts)
414 call col2(force5, v2, n_pts)
415 call col2(force6, v2, n_pts)
428 force4, force5, force6,&
429 s11, s22, s33, s12, s13, s23,&
430 p, n1, n2, n3, mu, n_pts)
432 type(c_ptr) :: force1, force2, force3
433 type(c_ptr) :: force4, force5, force6
434 type(c_ptr) :: p, n1, n2, n3
435 type(c_ptr) :: s11, s12, s22, s13, s23, s33
446 call device_vdot3(force4, s11, s12, s13, n1, n2, n3, n_pts)
447 call device_vdot3(force5, s12, s22, s23, n1, n2, n3, n_pts)
448 call device_vdot3(force6, s13, s23, s33, n1, n2, n3, n_pts)
456 call neko_error(
'error in drag_torque, no device bcklnd configured')
473 real(kind=
rp),
dimension(n_pts) :: n1, n2, n3
474 integer ::
mask(0:n_pts), facets(0:n_pts), fid, idx(4)
475 real(kind=
rp) :: normal(3), area(3)
481 normal = coef%get_normal(idx(1), idx(2), idx(3), idx(4), fid)
482 area = coef%get_area(idx(1), idx(2), idx(3), idx(4), fid)
483 n1(i) = normal(1)*area(1)
484 n2(i) = normal(2)*area(2)
485 n3(i) = normal(3)*area(3)
__inline__ __device__ void nonlinear_index(const int idx, const int lx, int *index)
type(mpi_datatype), public mpi_real_precision
MPI type for working precision of REAL types.
type(mpi_comm), public neko_comm
MPI communicator.
subroutine, public device_rzero(a_d, n, strm)
Zero a real vector.
subroutine, public device_cmult(a_d, c, n, strm)
Multiplication by constant c .
subroutine, public device_vdot3(dot_d, u1_d, u2_d, u3_d, v1_d, v2_d, v3_d, n, strm)
Compute a dot product (3-d version) assuming vector components etc.
subroutine, public device_col2(a_d, b_d, n, strm)
Vector multiplication .
subroutine, public device_col3(a_d, b_d, c_d, n, strm)
Vector multiplication with 3 vectors .
subroutine, public drag_torque_zone(dgtq, tstep, zone, center, s11, s22, s33, s12, s13, s23, p, coef, visc)
Some functions to calculate the lift/drag and torque Calculation can be done on a zone,...
subroutine, public setup_normals(coef, mask, facets, n1, n2, n3, n_pts)
Computes the normals for a given set of boundary points accessed by the mask.
subroutine, public drag_torque_facet(dgtq, xm0, ym0, zm0, center, s11, s22, s33, s12, s13, s23, pm1, visc, f, e, coef, xh)
Calculate drag and torque over a facet.
subroutine, public calc_force_array(force1, force2, force3, force4, force5, force6, s11, s22, s33, s12, s13, s23, p, n1, n2, n3, mu, n_pts)
Calculate drag and torque from array of points.
subroutine, public device_calc_force_array(force1, force2, force3, force4, force5, force6, s11, s22, s33, s12, s13, s23, p, n1, n2, n3, mu, n_pts)
Calculate drag and torque from array of points.
subroutine, public drag_torque_pt(dgtq, x, y, z, center, s11, s22, s33, s12, s13, s23, p, n1, n2, n3, v)
Calculate drag and torque from one point.
Defines a zone as a subset of facets in a mesh.
Object for handling masks in Neko.
subroutine, public col2(a, b, n)
Vector multiplication .
subroutine, public col3(a, b, c, n)
Vector multiplication with 3 vectors .
subroutine, public vdot3(dot, u1, u2, u3, v1, v2, v3, n)
Compute a dot product (3-d version) assuming vector components etc.
subroutine, public rzero(a, n)
Zero a real vector.
integer, parameter neko_bcknd_device
integer, parameter, public rp
Global precision used in computations.
Defines a function space.
Coefficients defined on a given (mesh, ) tuple. Arrays use indices (i,j,k,e): element e,...
The function space for the SEM solution fields.