39 opr_cpu_conv1, opr_cpu_convect_scalar, opr_cpu_cdtp, &
42 opr_sx_conv1, opr_sx_convect_scalar, opr_sx_cdtp, &
43 opr_sx_dudxyz, opr_sx_lambda2, opr_sx_set_convect_rst
64 use mpi_f08,
only : mpi_allreduce, mpi_in_place, mpi_max, mpi_sum
65 use,
intrinsic :: iso_c_binding, only : c_ptr
81 subroutine dudxyz (du, u, dr, ds, dt, coef)
82 type(
coef_t),
intent(in),
target :: coef
83 real(kind=
rp),
dimension(coef%Xh%lx, coef%Xh%ly, coef%Xh%lz, &
coef%msh%nelv),
intent(inout) :: du
84 real(kind=
rp),
dimension(coef%Xh%lx, coef%Xh%ly, coef%Xh%lz, &
coef%msh%nelv),
intent(in) :: u, dr, ds, dt
87 call opr_sx_dudxyz(du, u, dr, ds, dt, coef)
93 call opr_cpu_dudxyz(du, u, dr, ds, dt, coef)
104 subroutine div(res, ux, uy, uz, coef)
105 type(
coef_t),
intent(in),
target :: coef
106 real(kind=
rp),
dimension(coef%Xh%lx, coef%Xh%ly, coef%Xh%lz, &
coef%msh%nelv),
intent(inout) :: res
107 real(kind=
rp),
dimension(coef%Xh%lx, coef%Xh%ly, coef%Xh%lz, &
coef%msh%nelv),
intent(in) :: ux, uy, uz
119 call dudxyz(res, ux, coef%drdx, coef%dsdx, coef%dtdx, coef)
122 call dudxyz(work%x, uy, coef%drdy, coef%dsdy, coef%dtdy, coef)
126 call add2(res, work%x, work%size())
130 call dudxyz(work%x, uz, coef%drdz, coef%dsdz, coef%dtdz, coef)
134 call add2(res, work%x, work%size())
147 subroutine grad(ux, uy, uz, u, coef)
148 type(
coef_t),
intent(in) :: coef
149 real(kind=
rp),
dimension(coef%Xh%lxyz, coef%msh%nelv),
intent(inout) :: ux
150 real(kind=
rp),
dimension(coef%Xh%lxyz, coef%msh%nelv),
intent(inout) :: uy
151 real(kind=
rp),
dimension(coef%Xh%lxyz, coef%msh%nelv),
intent(inout) :: uz
152 real(kind=
rp),
dimension(coef%Xh%lxyz, coef%msh%nelv),
intent(in) :: u
154 call dudxyz(ux, u, coef%drdx, coef%dsdx, coef%dtdx, coef)
155 call dudxyz(uy, u, coef%drdy, coef%dsdy, coef%dtdy, coef)
156 call dudxyz(uz, u, coef%drdz, coef%dsdz, coef%dtdz, coef)
172 subroutine opgrad(ux, uy, uz, u, coef, es, ee)
173 type(
coef_t),
intent(in) :: coef
174 real(kind=
rp),
dimension(coef%Xh%lxyz, coef%msh%nelv),
intent(inout) :: ux
175 real(kind=
rp),
dimension(coef%Xh%lxyz, coef%msh%nelv),
intent(inout) :: uy
176 real(kind=
rp),
dimension(coef%Xh%lxyz, coef%msh%nelv),
intent(inout) :: uz
177 real(kind=
rp),
dimension(coef%Xh%lxyz, coef%msh%nelv),
intent(in) :: u
178 integer,
optional :: es, ee
179 integer :: eblk_start, eblk_end
181 if (
present(es))
then
187 if (
present(ee))
then
190 eblk_end = coef%msh%nelv
194 call opr_sx_opgrad(ux, uy, uz, u, coef)
200 call opr_cpu_opgrad(ux, uy, uz, u, coef, eblk_start, eblk_end)
209 subroutine ortho(x, glb_n_points, n)
210 integer,
intent(in) :: n
211 integer(kind=i8),
intent(in) :: glb_n_points
212 real(kind=
rp),
dimension(n),
intent(inout) :: x
220 c =
glsum(x, n)/glb_n_points
237 subroutine cdtp (dtx, x, dr, ds, dt, coef, es, ee)
238 type(
coef_t),
intent(in) :: coef
239 real(kind=
rp),
dimension(coef%Xh%lxyz, coef%msh%nelv),
intent(inout) :: dtx
240 real(kind=
rp),
dimension(coef%Xh%lxyz, coef%msh%nelv),
intent(inout) :: x
241 real(kind=
rp),
dimension(coef%Xh%lxyz, coef%msh%nelv),
intent(in) :: dr
242 real(kind=
rp),
dimension(coef%Xh%lxyz, coef%msh%nelv),
intent(in) :: ds
243 real(kind=
rp),
dimension(coef%Xh%lxyz, coef%msh%nelv),
intent(in) :: dt
244 integer,
optional :: es, ee
245 integer :: eblk_start, eblk_end
247 if (
present(es))
then
253 if (
present(ee))
then
256 eblk_end = coef%msh%nelv
260 call opr_sx_cdtp(dtx, x, dr, ds, dt, coef)
266 call opr_cpu_cdtp(dtx, x, dr, ds, dt, coef, eblk_start, eblk_end)
281 subroutine conv1(du, u, vx, vy, vz, Xh, coef, es, ee)
282 type(
space_t),
intent(in) :: xh
283 type(
coef_t),
intent(in) :: coef
284 real(kind=
rp),
intent(inout) :: du(xh%lxyz, coef%msh%nelv)
285 real(kind=
rp),
intent(inout) :: u(xh%lx, xh%ly, xh%lz, coef%msh%nelv)
286 real(kind=
rp),
intent(inout) :: vx(xh%lx, xh%ly, xh%lz, coef%msh%nelv)
287 real(kind=
rp),
intent(inout) :: vy(xh%lx, xh%ly, xh%lz, coef%msh%nelv)
288 real(kind=
rp),
intent(inout) :: vz(xh%lx, xh%ly, xh%lz, coef%msh%nelv)
289 integer,
optional :: es, ee
290 integer :: eblk_end, eblk_start
292 associate(nelv => coef%msh%nelv, gdim => coef%msh%gdim)
293 if (
present(es))
then
299 if (
present(ee))
then
302 eblk_end = coef%msh%nelv
306 call opr_sx_conv1(du, u, vx, vy, vz, xh, coef, nelv)
312 call opr_cpu_conv1(du, u, vx, vy, vz, xh, coef, eblk_start, eblk_end)
335 subroutine convect_scalar(du, u, cr, cs, ct, Xh_GLL, Xh_GL, coef_GLL, &
337 type(space_t),
intent(in) :: xh_gl
338 type(space_t),
intent(in) :: xh_gll
339 type(coef_t),
intent(in) :: coef_GLL
340 type(coef_t),
intent(in) :: coef_GL
341 type(interpolator_t),
intent(inout) :: GLL_to_GL
342 real(kind=rp),
intent(inout) :: &
343 du(xh_gll%lx, xh_gll%ly, xh_gll%lz, coef_gl%msh%nelv)
344 real(kind=rp),
intent(inout) :: &
345 u(xh_gl%lx, xh_gl%lx, xh_gl%lx, coef_gl%msh%nelv)
346 type(field_t),
intent(inout) :: cr, cs, ct
349 if (neko_bcknd_sx .eq. 1)
then
350 call opr_sx_convect_scalar(du, u, cr%x, cs%x, ct%x, &
351 xh_gll, xh_gl, coef_gll, coef_gl, gll_to_gl)
352 else if (neko_bcknd_xsmm .eq. 1)
then
353 call opr_xsmm_convect_scalar(du, u, cr%x, cs%x, ct%x, &
354 xh_gll, xh_gl, coef_gll, coef_gl, gll_to_gl)
355 else if (neko_bcknd_device .eq. 1)
then
356 u_d = device_get_ptr(u)
357 call opr_device_convect_scalar(du, u_d, cr%x_d, cs%x_d, ct%x_d, &
358 xh_gll, xh_gl, coef_gll, coef_gl, gll_to_gl)
360 call opr_cpu_convect_scalar(du, u, cr%x, cs%x, ct%x, &
361 xh_gll, xh_gl, coef_gll, coef_gl, gll_to_gl)
376 subroutine curl(w1, w2, w3, u1, u2, u3, work1, work2, coef, event)
377 type(field_t),
intent(inout) :: w1
378 type(field_t),
intent(inout) :: w2
379 type(field_t),
intent(inout) :: w3
380 type(field_t),
intent(in) :: u1
381 type(field_t),
intent(in) :: u2
382 type(field_t),
intent(in) :: u3
383 type(field_t),
intent(inout) :: work1
384 type(field_t),
intent(inout) :: work2
385 type(coef_t),
intent(in) :: coef
386 type(c_ptr),
optional,
intent(inout) :: event
388 if (neko_bcknd_sx .eq. 1)
then
389 call opr_sx_curl(w1, w2, w3, u1, u2, u3, work1, work2, coef)
390 else if (neko_bcknd_xsmm .eq. 1)
then
391 call opr_xsmm_curl(w1, w2, w3, u1, u2, u3, work1, work2, coef)
392 else if (neko_bcknd_device .eq. 1)
then
393 if (
present(event))
then
394 call opr_device_curl(w1, w2, w3, u1, u2, u3, &
395 work1, work2, coef, event)
397 call opr_device_curl(w1, w2, w3, u1, u2, u3, work1, work2, coef)
400 call opr_cpu_curl(w1, w2, w3, u1, u2, u3, work1, work2, coef)
414 function cfl(dt, u, v, w, Xh, coef, nelv, gdim)
415 type(space_t),
intent(in) :: xh
416 type(coef_t),
intent(in) :: coef
417 integer,
intent(in) :: nelv, gdim
418 real(kind=rp),
intent(in) :: dt
419 real(kind=rp),
dimension(Xh%lx, Xh%ly, Xh%lz, nelv),
intent(in) :: u, v, w
423 if (neko_bcknd_sx .eq. 1)
then
424 cfl = opr_sx_cfl(dt, u, v, w, xh, coef, nelv)
425 else if (neko_bcknd_device .eq. 1)
then
426 cfl = opr_device_cfl(dt, u, v, w, xh, coef, nelv, gdim)
428 cfl = opr_cpu_cfl(dt, u, v, w, xh, coef, nelv, gdim)
431 if (.not. neko_device_mpi)
then
432 call mpi_allreduce(mpi_in_place,
cfl, 1, &
433 mpi_real_precision, mpi_max, neko_comm, ierr)
446 type(space_t),
intent(in) :: xh
447 type(coef_t),
intent(in) :: coef
448 integer,
intent(in) :: nelv, gdim
449 real(kind=rp),
intent(in) :: dt
450 real(kind=rp),
dimension(Xh%lx, Xh%ly, Xh%lz, nelv),
intent(in) :: max_wave_speed
453 type(field_t),
pointer :: zero_vector
456 n = xh%lx * xh%ly * xh%lz * nelv
459 call neko_scratch_registry%request_field(zero_vector, ind)
462 call field_rzero(zero_vector)
465 cfl_compressible =
cfl(dt, max_wave_speed, zero_vector%x, zero_vector%x, xh, coef, nelv, gdim)
468 call neko_scratch_registry%relinquish_field(ind)
484 subroutine strain_rate(s11, s22, s33, s12, s13, s23, u, v, w, coef)
485 type(field_t),
intent(in) :: u, v, w
486 type(coef_t),
intent(in) :: coef
487 real(kind=rp),
intent(inout) :: s11(u%Xh%lx, u%Xh%ly, u%Xh%lz, u%msh%nelv)
488 real(kind=rp),
intent(inout) :: s22(u%Xh%lx, u%Xh%ly, u%Xh%lz, u%msh%nelv)
489 real(kind=rp),
intent(inout) :: s33(u%Xh%lx, u%Xh%ly, u%Xh%lz, u%msh%nelv)
490 real(kind=rp),
intent(inout) :: s12(u%Xh%lx, u%Xh%ly, u%Xh%lz, u%msh%nelv)
491 real(kind=rp),
intent(inout) :: s13(u%Xh%lx, u%Xh%ly, u%Xh%lz, u%msh%nelv)
492 real(kind=rp),
intent(inout) :: s23(u%Xh%lx, u%Xh%ly, u%Xh%lz, u%msh%nelv)
494 type(c_ptr) :: s11_d, s22_d, s33_d, s12_d, s23_d, s13_d
496 integer :: nelv, lxyz
498 if (neko_bcknd_device .eq. 1)
then
499 s11_d = device_get_ptr(s11)
500 s22_d = device_get_ptr(s22)
501 s33_d = device_get_ptr(s33)
502 s12_d = device_get_ptr(s12)
503 s23_d = device_get_ptr(s23)
504 s13_d = device_get_ptr(s13)
511 call dudxyz (s12, u%x, coef%drdy, coef%dsdy, coef%dtdy, coef)
512 call dudxyz (s11, v%x, coef%drdx, coef%dsdx, coef%dtdx, coef)
513 if (neko_bcknd_device .eq. 1)
then
514 call device_add2(s12_d, s11_d, nelv*lxyz)
516 call add2(s12, s11, nelv*lxyz)
519 call dudxyz (s13, u%x, coef%drdz, coef%dsdz, coef%dtdz, coef)
520 call dudxyz (s11, w%x, coef%drdx, coef%dsdx, coef%dtdx, coef)
521 if (neko_bcknd_device .eq. 1)
then
522 call device_add2(s13_d, s11_d, nelv*lxyz)
524 call add2(s13, s11, nelv*lxyz)
527 call dudxyz (s23, v%x, coef%drdz, coef%dsdz, coef%dtdz, coef)
528 call dudxyz (s11, w%x, coef%drdy, coef%dsdy, coef%dtdy, coef)
529 if (neko_bcknd_device .eq. 1)
then
530 call device_add2(s23_d, s11_d, nelv*lxyz)
532 call add2(s23, s11, nelv*lxyz)
535 call dudxyz (s11, u%x, coef%drdx, coef%dsdx, coef%dtdx, coef)
536 call dudxyz (s22, v%x, coef%drdy, coef%dsdy, coef%dtdy, coef)
537 call dudxyz (s33, w%x, coef%drdz, coef%dsdz, coef%dtdz, coef)
539 if (neko_bcknd_device .eq. 1)
then
540 call device_cmult(s12_d, 0.5_rp, nelv*lxyz)
541 call device_cmult(s13_d, 0.5_rp, nelv*lxyz)
542 call device_cmult(s23_d, 0.5_rp, nelv*lxyz)
544 call cmult(s12, 0.5_rp, nelv*lxyz)
545 call cmult(s13, 0.5_rp, nelv*lxyz)
546 call cmult(s23, 0.5_rp, nelv*lxyz)
557 subroutine lambda2op(lambda2, u, v, w, coef)
558 type(coef_t),
intent(in) :: coef
559 type(field_t),
intent(inout) ::
lambda2
560 type(field_t),
intent(in) :: u, v, w
562 if (neko_bcknd_sx .eq. 1)
then
563 call opr_sx_lambda2(
lambda2, u, v, w, coef)
564 else if (neko_bcknd_device .eq. 1)
then
565 call opr_device_lambda2(
lambda2, u, v, w, coef)
567 call opr_cpu_lambda2(
lambda2, u, v, w, coef)
583 type(space_t),
intent(inout) :: xh
584 type(coef_t),
intent(inout) :: coef
585 type(field_t),
intent(inout) :: cr, cs, ct
586 real(kind=rp),
dimension(Xh%lxyz, coef%msh%nelv), &
587 intent(in) :: cx, cy, cz
588 type(c_ptr) :: cx_d, cy_d, cz_d
590 if (neko_bcknd_sx .eq. 1)
then
591 call opr_sx_set_convect_rst(cr%x, cs%x, ct%x, cx, cy, cz, xh, coef)
592 else if (neko_bcknd_xsmm .eq. 1)
then
593 call opr_xsmm_set_convect_rst(cr%x, cs%x, ct%x, cx, cy, cz, xh, coef)
594 else if (neko_bcknd_device .eq. 1)
then
595 cx_d = device_get_ptr(cx)
596 cy_d = device_get_ptr(cy)
597 cz_d = device_get_ptr(cz)
598 call opr_device_set_convect_rst(cr%x_d, cs%x_d, ct%x_d, &
599 cx_d, cy_d, cz_d, xh, coef)
601 call opr_cpu_set_convect_rst(cr%x, cs%x, ct%x, cx, cy, cz, xh, coef)
621 subroutine runge_kutta(phi, conv_k1, conv_k23, conv_k4, Xh_GLL, Xh_GL, &
622 coef, coef_GL, GLL_to_GL, tau, dtau, n, nel, n_GL)
623 type(space_t),
intent(in) :: xh_gll
624 type(space_t),
intent(inout) :: xh_gl
625 type(coef_t),
intent(in) :: coef
626 type(coef_t),
intent(inout) :: coef_gl
627 type(interpolator_t) :: gll_to_gl
628 real(kind=rp),
intent(inout) :: tau, dtau
629 integer,
intent(in) :: n, nel, n_gl
630 type(field_t),
intent(inout) :: phi
631 type(field_list_t) :: conv_k1, conv_k23, conv_k4
632 real(kind=rp) :: c1, c2, c3
633 type(field_t),
pointer :: u1, k1, k2, k3, k4
634 real(kind=rp),
dimension(n_GL) :: u1_gl
635 integer :: ind(5), i, e
636 type(c_ptr) :: u1_gl_d
638 call neko_scratch_registry%request_field(u1, ind(1))
639 call neko_scratch_registry%request_field(k1, ind(2))
640 call neko_scratch_registry%request_field(k2, ind(3))
641 call neko_scratch_registry%request_field(k3, ind(4))
642 call neko_scratch_registry%request_field(k4, ind(5))
648 if (neko_bcknd_device .eq. 1)
then
649 call device_map(u1_gl, u1_gl_d, n_gl)
652 call device_invcol3(u1%x_d, phi%x_d, coef%B_d, n)
653 call gll_to_gl%map(u1_gl, u1%x, nel, xh_gl)
655 conv_k1%items(2)%ptr, conv_k1%items(3)%ptr, &
656 xh_gll, xh_gl, coef, coef_gl, gll_to_gl)
657 call device_col2(k1%x_d, coef%B_d, n)
660 call device_add3s2(u1%x_d, phi%x_d, k1%x_d, c1, c2, n)
661 call device_invcol2(u1%x_d, coef%B_d, n)
662 call gll_to_gl%map(u1_gl, u1%x, nel, xh_gl)
664 conv_k23%items(2)%ptr, conv_k23%items(3)%ptr, &
665 xh_gll, xh_gl, coef, coef_gl, gll_to_gl)
666 call device_col2(k2%x_d, coef%B_d, n)
669 call device_add3s2(u1%x_d, phi%x_d, k2%x_d, c1, c2, n)
670 call device_invcol2(u1%x_d, coef%B_d, n)
671 call gll_to_gl%map(u1_gl, u1%x, nel, xh_gl)
673 conv_k23%items(2)%ptr, conv_k23%items(3)%ptr, &
674 xh_gll, xh_gl, coef, coef_gl, gll_to_gl)
675 call device_col2(k3%x_d, coef%B_d, n)
678 call device_add3s2(u1%x_d, phi%x_d, k3%x_d, c1, c3, n)
679 call device_invcol2(u1%x_d, coef%B_d, n)
680 call gll_to_gl%map(u1_gl, u1%x, nel, xh_gl)
682 conv_k4%items(2)%ptr, conv_k4%items(3)%ptr, &
683 xh_gll, xh_gl, coef, coef_gl, gll_to_gl)
684 call device_col2(k4%x_d, coef%B_d, n)
689 call device_add5s4(phi%x_d, k1%x_d, k2%x_d, k3%x_d, k4%x_d, &
692 call device_free(u1_gl_d)
697 call invcol3(u1%x, phi%x, coef%B, n)
698 call gll_to_gl%map(u1_gl, u1%x, nel, xh_gl)
700 conv_k1%items(2)%ptr, conv_k1%items(3)%ptr, &
701 xh_gll, xh_gl, coef, coef_gl, gll_to_gl)
702 call col2(k1%x, coef%B, n)
705 call add3s2(u1%x, phi%x, k1%x, c1, c2, n)
706 call invcol2(u1%x, coef%B, n)
707 call gll_to_gl%map(u1_gl, u1%x, nel, xh_gl)
709 conv_k23%items(2)%ptr, conv_k23%items(3)%ptr, &
710 xh_gll, xh_gl, coef, coef_gl, gll_to_gl)
711 call col2(k2%x, coef%B, n)
714 call add3s2(u1%x, phi%x, k2%x, c1, c2, n)
715 call invcol2(u1%x, coef%B, n)
716 call gll_to_gl%map(u1_gl, u1%x, nel, xh_gl)
718 conv_k23%items(2)%ptr, conv_k23%items(3)%ptr, &
719 xh_gll, xh_gl, coef, coef_gl, gll_to_gl)
720 call col2(k3%x, coef%B, n)
723 call add3s2(u1%x, phi%x, k3%x, c1, c3, n)
724 call invcol2(u1%x, coef%B, n)
725 call gll_to_gl%map(u1_gl, u1%x, nel, xh_gl)
727 conv_k4%items(2)%ptr, conv_k4%items(3)%ptr, &
728 xh_gll, xh_gl, coef, coef_gl, gll_to_gl)
729 call col2(k4%x, coef%B, n)
733 call add5s4(phi%x, k1%x, k2%x, k3%x, k4%x, c1, c2, c2, c1, n)
736 call neko_scratch_registry%relinquish_field(ind)
Return the device pointer for an associated Fortran array.
Map a Fortran array to a device (allocate and associate)
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_add2(a_d, b_d, n, strm)
Vector addition .
real(kind=rp) function, public device_glsum(a_d, n, strm)
Sum a vector of length n.
subroutine, public device_add3s2(a_d, b_d, c_d, c1, c2, n, strm)
Returns .
subroutine, public device_cmult(a_d, c, n, strm)
Multiplication by constant c .
subroutine, public device_copy(a_d, b_d, n, strm)
Copy a vector .
subroutine, public device_invcol3(a_d, b_d, c_d, n, strm)
Vector division .
subroutine, public device_col2(a_d, b_d, n, strm)
Vector multiplication .
subroutine, public device_add5s4(a_d, b_d, c_d, d_d, e_d, c1, c2, c3, c4, n, strm)
Returns .
subroutine, public device_invcol2(a_d, b_d, n, strm)
Vector division .
Device abstraction, common interface for various accelerators.
subroutine, public device_free(x_d)
Deallocate memory on the device.
subroutine, public field_rzero(a, n)
Zero a real vector.
Routines to interpolate between different spaces.
A simulation component that computes lambda2 The values are stored in the field registry under the na...
subroutine, public cmult(a, c, n)
Multiplication by constant c .
subroutine, public invcol2(a, b, n)
Vector division .
subroutine, public cadd(a, s, n)
Add a scalar to vector .
real(kind=rp) function, public glsum(a, n)
Sum a vector of length n.
subroutine, public add2(a, b, n)
Vector addition .
subroutine, public invcol3(a, b, c, n)
Invert a vector .
subroutine, public add3s2(a, b, c, c1, c2, n)
Returns .
subroutine, public col2(a, b, n)
Vector multiplication .
subroutine, public copy(a, b, n)
Copy a vector .
subroutine, public add5s4(a, b, c, d, e, c1, c2, c3, c4, n)
Returns .
subroutine, public rzero(a, n)
Zero a real vector.
integer, parameter neko_bcknd_sx
integer, parameter neko_bcknd_device
logical, parameter neko_device_mpi
integer, parameter neko_bcknd_xsmm
integer, parameter, public i8
integer, parameter, public rp
Global precision used in computations.
subroutine, public set_convect_rst(cr, cs, ct, cx, cy, cz, xh, coef)
Transforms the convecting velocity field to the rst form of the GL space.
subroutine, public ortho(x, glb_n_points, n)
Othogonalize with regard to vector (1,1,1,1,1,1...,1)^T.
real(kind=rp) function, public cfl(dt, u, v, w, xh, coef, nelv, gdim)
subroutine, public opgrad(ux, uy, uz, u, coef, es, ee)
Compute the weak gradient of a scalar field, i.e. the gradient multiplied by the mass matrix.
subroutine convect_scalar(du, u, cr, cs, ct, xh_gll, xh_gl, coef_gll, coef_gl, gll_to_gl)
Apply the convecting velocity c to the to the scalar field u, used in the OIFS scheme.
subroutine, public div(res, ux, uy, uz, coef)
Compute the divergence of a vector field.
real(kind=rp) function, public cfl_compressible(dt, max_wave_speed, xh, coef, nelv, gdim)
subroutine, public conv1(du, u, vx, vy, vz, xh, coef, es, ee)
Compute the advection term.
subroutine, public grad(ux, uy, uz, u, coef)
Compute the gradient of a scalar field.
subroutine, public curl(w1, w2, w3, u1, u2, u3, work1, work2, coef, event)
subroutine, public strain_rate(s11, s22, s33, s12, s13, s23, u, v, w, coef)
Compute the strain rate tensor, i.e 0.5 * du_i/dx_j + du_j/dx_i.
subroutine, public lambda2op(lambda2, u, v, w, coef)
Compute the Lambda2 field for a given velocity field.
subroutine, public cdtp(dtx, x, dr, ds, dt, coef, es, ee)
Apply D^T to a scalar field, where D is the derivative matrix.
subroutine, public dudxyz(du, u, dr, ds, dt, coef)
Compute derivative of a scalar field along a single direction.
subroutine, public runge_kutta(phi, conv_k1, conv_k23, conv_k4, xh_gll, xh_gl, coef, coef_gl, gll_to_gl, tau, dtau, n, nel, n_gl)
Compute one step of Runge Kutta time interpolation for OIFS scheme.
subroutine, public opr_cpu_curl(w1, w2, w3, u1, u2, u3, work1, work2, c_xh)
real(kind=rp) function, public opr_cpu_cfl(dt, u, v, w, xh, coef, nelv, gdim)
subroutine, public opr_cpu_lambda2(lambda2, u, v, w, coef)
Operators accelerator backends.
subroutine, public opr_device_convect_scalar(du, u_d, cr_d, cs_d, ct_d, xh_gll, xh_gl, coef_gll, coef_gl, gll_to_gl)
subroutine, public opr_device_cdtp(dtx, x, dr, ds, dt, coef)
real(kind=rp) function, public opr_device_cfl(dt, u, v, w, xh, coef, nelv, gdim)
subroutine, public opr_device_curl(w1, w2, w3, u1, u2, u3, work1, work2, c_xh, event)
subroutine, public opr_device_set_convect_rst(cr_d, cs_d, ct_d, cx_d, cy_d, cz_d, xh, coef)
subroutine, public opr_device_dudxyz(du, u, dr, ds, dt, coef)
subroutine, public opr_device_opgrad(ux, uy, uz, u, coef)
subroutine, public opr_device_conv1(du, u, vx, vy, vz, xh, coef, nelv, gdim)
subroutine, public opr_device_lambda2(lambda2, u, v, w, coef)
Operators SX-Aurora backend.
subroutine, public opr_sx_curl(w1, w2, w3, u1, u2, u3, work1, work2, c_xh)
Operators libxsmm backend.
subroutine, public opr_xsmm_convect_scalar(du, u, cr, cs, ct, xh_gll, xh_gl, coef_gll, coef_gl, gll_to_gl)
subroutine, public opr_xsmm_conv1(du, u, vx, vy, vz, xh, coef, nelv, gdim)
subroutine, public opr_xsmm_cdtp(dtx, x, dr, ds, dt, coef)
subroutine, public opr_xsmm_curl(w1, w2, w3, u1, u2, u3, work1, work2, c_xh)
subroutine, public opr_xsmm_dudxyz(du, u, dr, ds, dt, coef)
subroutine, public opr_xsmm_opgrad(ux, uy, uz, u, coef)
subroutine, public opr_xsmm_set_convect_rst(cr, cs, ct, cx, cy, cz, xh, coef)
Defines a registry for storing and requesting temporary fields This can be used when you have a funct...
type(scratch_registry_t), target, public neko_scratch_registry
Global scratch registry.
Defines a function space.
Coefficients defined on a given (mesh, ) tuple. Arrays use indices (i,j,k,e): element e,...
field_list_t, To be able to group fields together
Interpolation between two space::space_t.
The function space for the SEM solution fields.