navier8.F90 File Reference

Go to the source code of this file.

Functions/Subroutines
subroutine	set_vert (glo_num, ngv, nx, nel, vertex, ifcenter)
	Given global array, vertex, pointing to hex vertices, set up a new array of global pointers for an nx^ndim set of elements. More...
subroutine	set_up_h1_crs
	? More...
subroutine	set_jl_crs_mask (n, mask, se_to_gcrs)
subroutine	set_mat_ij (ia, ja, n, ne)
subroutine	ituple_sort (a, lda, n, key, nkey, ind, aa)
	Use Heap Sort (p 231 Num. Rec., 1st Ed.) More...
logical function	iftuple_ialtb (a, b, key, nkey)
logical function	iftuple_ianeb (a, b, key, nkey)
subroutine	specmpn (b, nb, a, na, ba, ab, if3d, w, ldw)
	Spectral interpolation from A to B via tensor products. More...
subroutine	irank (A, IND, N)
	Use Heap Sort (p 233 Num. Rec.), 5/26/93 pff. More...
subroutine	get_local_crs_galerkin (a, ncl, nxc, h1, h2, w1, w2)
	This routine generates Nelv submatrices of order ncl using Galerkin projection. More...
subroutine	gen_crs_basis (b, j)
subroutine	get_vertex
subroutine	assign_gllnid (gllnid, iunsort, nelgt, nelgv, np)
subroutine	get_vert
subroutine	get_vert_map (vertex, nlv, nel, suffix, ifgfdm)
subroutine	irank_vecn (ind, nn, a, m, n, key, nkey, aa)
	Compute rank of each unique entry a(1,i) Output: ind(i) i=1,...,n (global) rank of entry a(*,i) nn = max(rank) a(j,i) is permuted Input: a(j,i) j=1,...,m; i=1,...,n m : leading dim. of v (ldv must be .ge. m) key : sort key nkey : Although not mandatory, this ranking procedure is probably most effectively employed when the keys are pre-sorted. Thus, the option is provided to sort vi() prior to the ranking. More...
subroutine	gbtuple_rank (tuple, m, n, nmax, cr_h, nid, np, ind)
	Return a unique rank for each matched tuple set. Global. Balanced. tuple is destroyed. By "balanced" we mean that none of the tuple entries is likely to be much more uniquely populated than any other, so that any of the tuples can serve as an initial (parallel) sort key First two slots in tuple(:,i) assumed empty. More...
subroutine	setvert3d (glo_num, ngv, nx, nel, vertex, ifcenter)
	setup unique ids for dssum. note: total number of unique vertices, edges and faces has to be smaller than 2**31 (integer-4 limit). if nelgt < 2**31/12 we're ok for sure (independent of N)! More...
subroutine	check_p_bc (glo_num, nx, ny, nz, nel)

Function/Subroutine Documentation

subroutine assign_gllnid	(	integer, dimension(*)	gllnid,
		integer, dimension(*)	iunsort,
		integer	nelgt,
		integer	nelgv,
		integer	np
	)

Definition at line 519 of file navier8.F90.

References exitti(), isort(), and iswapt_ip().

Here is the call graph for this function:

subroutine check_p_bc	(	integer(i8), dimension(nx,ny,nz,nel)	glo_num,
		integer	nx,
		integer	ny,
		integer	nz,
		integer	nel
	)

Definition at line 1237 of file navier8.F90.

Referenced by set_vert().

Here is the caller graph for this function:

subroutine gbtuple_rank	(	integer, dimension(m,nmax)	tuple,
		integer	m,
		integer	n,
		integer	nmax,
		integer	cr_h,
		integer	nid,
		integer	np,
		integer, dimension(nmax)	ind
	)

Return a unique rank for each matched tuple set. Global. Balanced. tuple is destroyed. By "balanced" we mean that none of the tuple entries is likely to be much more uniquely populated than any other, so that any of the tuples can serve as an initial (parallel) sort key First two slots in tuple(:,i) assumed empty.

Definition at line 802 of file navier8.F90.

References exitt(), irank_vecn(), and ituple_sort().

Referenced by setvert3d().

Here is the call graph for this function:

Here is the caller graph for this function:

subroutine gen_crs_basis	(	real(dp), dimension(nx1,ny1,nz1)	b,
		integer	j
	)

Definition at line 449 of file navier8.F90.

References copy(), and speclib::zwgll().

Referenced by get_local_crs_galerkin().

Here is the call graph for this function:

Here is the caller graph for this function:

subroutine get_local_crs_galerkin	(	real(dp), dimension(ncl,ncl,*)	a,
		integer	ncl,
		integer	nxc,
		real(dp), dimension(*)	h1,
		real(dp), dimension(*)	h2,
		real(dp), dimension(nx1*ny1*nz1,nelv)	w1,
		real(dp), dimension(nx1*ny1*nz1,nelv)	w2
	)

This routine generates Nelv submatrices of order ncl using Galerkin projection.

Definition at line 407 of file navier8.F90.

References axhelm(), copy(), and gen_crs_basis().

Referenced by set_up_h1_crs().

Here is the call graph for this function:

Here is the caller graph for this function:

subroutine get_vert

(

)

Definition at line 607 of file navier8.F90.

References get_vert_map().

Referenced by get_vertex(), hsmg_routines::h1mg_setup_dssum(), hsmg_routines::hsmg_setup_dssum(), set_proc_map(), and setup_topo().

Here is the call graph for this function:

Here is the caller graph for this function:

subroutine get_vert_map	(	integer, dimension(nlv,*)	vertex,
		integer	nlv,
		integer	nel,
		character(4)	suffix,
		logical	ifgfdm
	)

Definition at line 635 of file navier8.F90.

References bcast(), blank(), chcopy(), exitt(), exitt0(), parallel::gllel(), parallel::gllnid(), mesh::ieg_to_xyz(), parallel::init_gllnid(), parallel::lglel(), string::ltrunc(), nekgsync(), and np.

Referenced by get_vert().

Here is the call graph for this function:

Here is the caller graph for this function:

subroutine get_vertex

(

)

Definition at line 499 of file navier8.F90.

References get_vert().

Referenced by set_up_h1_crs().

Here is the call graph for this function:

Here is the caller graph for this function:

logical function iftuple_ialtb	(	integer, dimension(*)	a,
		integer, dimension(*)	b,
		integer, dimension(nkey)	key,
		integer	nkey
	)

Definition at line 267 of file navier8.F90.

logical function iftuple_ianeb	(	integer, dimension(*)	a,
		integer, dimension(*)	b,
		integer, dimension(nkey)	key,
		integer	nkey
	)

Definition at line 290 of file navier8.F90.

subroutine irank	(	integer, dimension(*)	A,
		integer, dimension(*)	IND,
		integer	N
	)

Use Heap Sort (p 233 Num. Rec.), 5/26/93 pff.

Definition at line 355 of file navier8.F90.

Referenced by setvert3d().

Here is the caller graph for this function:

subroutine irank_vecn	(	integer, dimension(n)	ind,
		integer	nn,
		integer, dimension(m,n)	a,
		integer	m,
		integer	n,
		integer, dimension(nkey)	key,
		integer	nkey,
		integer, dimension(m)	aa
	)

Compute rank of each unique entry a(1,i) Output: ind(i) i=1,...,n (global) rank of entry a(*,i) nn = max(rank) a(j,i) is permuted Input: a(j,i) j=1,...,m; i=1,...,n m : leading dim. of v (ldv must be .ge. m) key : sort key nkey : Although not mandatory, this ranking procedure is probably most effectively employed when the keys are pre-sorted. Thus, the option is provided to sort vi() prior to the ranking.

Definition at line 766 of file navier8.F90.

References ituple_sort().

Referenced by gbtuple_rank().

Here is the call graph for this function:

Here is the caller graph for this function:

subroutine ituple_sort	(	integer, dimension(lda,n)	a,
		integer	lda,
		integer	n,
		integer, dimension(nkey)	key,
		integer	nkey,
		integer, dimension(*)	ind,
		integer, dimension(lda)	aa
	)

Use Heap Sort (p 231 Num. Rec., 1st Ed.)

Definition at line 204 of file navier8.F90.

Referenced by gbtuple_rank(), and irank_vecn().

Here is the caller graph for this function:

subroutine set_jl_crs_mask	(	integer	n,
		real(dp), dimension(*)	mask,
		integer(i8), dimension(*)	se_to_gcrs
	)

Definition at line 170 of file navier8.F90.

Referenced by set_up_h1_crs().

Here is the caller graph for this function:

subroutine set_mat_ij	(	integer, dimension(n,n,ne)	ia,
		integer, dimension(n,n,ne)	ja,
		integer	n,
		integer	ne
	)

Definition at line 185 of file navier8.F90.

Referenced by set_up_h1_crs().

Here is the caller graph for this function:

subroutine set_up_h1_crs

(

)

?

Definition at line 35 of file navier8.F90.

References crs_setup, ctimer::dnekclock(), facev(), get_local_crs_galerkin(), get_vertex(), gs_free, gs_op, gs_setup, np, set_jl_crs_mask(), set_mat_ij(), and set_vert().

Referenced by set_overlap().

Here is the call graph for this function:

Here is the caller graph for this function:

subroutine set_vert	(	integer(i8), dimension(1)	glo_num,
		integer(i8)	ngv,
		integer	nx,
		integer	nel,
		integer, dimension(*)	vertex,
		logical	ifcenter
	)

Given global array, vertex, pointing to hex vertices, set up a new array of global pointers for an nx^ndim set of elements.

Definition at line 4 of file navier8.F90.

References check_p_bc(), and setvert3d().

Referenced by set_up_h1_crs(), and setupds().

Here is the call graph for this function:

Here is the caller graph for this function:

subroutine setvert3d	(	integer(i8), dimension(*)	glo_num,
		integer(i8)	ngv,
		integer	nx,
		integer	nel,
		integer, dimension(0:1,0:1,0:1,*)	vertex,
		logical	ifcenter
	)

setup unique ids for dssum. note: total number of unique vertices, edges and faces has to be smaller than 2**31 (integer-4 limit). if nelgt < 2**31/12 we're ok for sure (independent of N)!

Definition at line 860 of file navier8.F90.

References dsset(), gbtuple_rank(), irank(), isort(), parallel::lglel(), and np.

Referenced by set_vert().

Here is the call graph for this function:

Here is the caller graph for this function:

subroutine specmpn	(	real(dp), dimension(nb,nb,nb), intent(out)	b,
		integer, intent(in)	nb,
		real(dp), dimension(na,na,na), intent(in)	a,
		integer, intent(in)	na,
		real(dp)	ba,
		real(dp)	ab,
		logical, intent(in)	if3d,
		real(dp), dimension(ldw), intent(out)	w,
		integer, intent(in)	ldw
	)

Spectral interpolation from A to B via tensor products.

• scratch arrays: w(na*na*nb + nb*nb*na) 5/3/00 – this routine replaces specmp in navier1.f, which has a potential memory problem

Parameters

[out]

w

work buffer

Definition at line 313 of file navier8.F90.

References exitt(), and mxm().

Referenced by convect_new(), set_convect_new(), and set_dealias_rx().

Here is the call graph for this function:

Here is the caller graph for this function: