|
Nek5000
SEM for Incompressible NS
|
All Classes Namespaces Files Functions Variables Typedefs Enumerations Enumerator Macros Modules Pages
|
Nek5000
SEM for Incompressible NS
|
Go to the source code of this file.
Functions/Subroutines | |
| subroutine | hmholtz (name, u, rhs, h1, h2, mask, mult, imsh, tli, maxit, isd) |
| subroutine | axhelm (au, u, helm1, helm2, imesh, isd) |
| Compute the (Helmholtz) matrix-vector product, AU = helm1*[A]u + helm2*[B]u, for NEL elements. More... | |
| subroutine | setfast (helm1, helm2, imesh) |
| Set logicals for fast evaluation of A*x. More... | |
| subroutine | sfastax () |
| For undeformed elements, set up appropriate elemental matrices and geometric factors for fast evaluation of Ax. More... | |
| subroutine | setprec (dpcm1, helm1, helm2, imsh, isd) |
| Generate diagonal preconditioner for the Helmholtz operator. More... | |
| subroutine | chktcg1 (tol, res, h1, h2, mask, mult, imesh, isd) |
| Check that the tolerances are not too small for the CG-solver. Important when calling the CG-solver (Gauss-Lobatto mesh) with zero Neumann b.c. More... | |
| subroutine | cggo (x, f, h1, h2, mask, mult, imsh, tin, maxit, isd, binv, name) |
| Solve the Helmholtz equation, H*U = RHS, using preconditioned conjugate gradient iteration. Preconditioner: diag(H). More... | |
| real(dp) function | vlsc32 (r, b, m, n) |
| subroutine | set_fdm_prec_h1b (d, h1, h2, nel) |
| subroutine | generalev (a, b, lam, n, w) |
| Solve the generalized eigenvalue problem A x = lam B x. More... | |
| subroutine | outmat2 (a, m, n, k, name) |
| subroutine axhelm | ( | real(dp), dimension (lx1,ly1,lz1,*), intent(out) | au, |
| real(dp), dimension (lx1,ly1,lz1,*), intent(in) | u, | ||
| real(dp), dimension (lx1,ly1,lz1,*), intent(in) | helm1, | ||
| real(dp), dimension (lx1,ly1,lz1,*), intent(in) | helm2, | ||
| integer, intent(in) | imesh, | ||
| integer, intent(in) | isd | ||
| ) |
Compute the (Helmholtz) matrix-vector product, AU = helm1*[A]u + helm2*[B]u, for NEL elements.
| [out] | au | H u |
| [in] | helm1 | coefficient of stif |
| [in] | helm2 | coefficient of mass |
| [in] | imesh | mesh index (v or t) |
| [in] | isd | axi-symmetric flag of sorts |
Definition at line 79 of file hmholtz.F90.
References ctimer::dnekclock(), i, mxm(), and setfast().
Referenced by ax(), cdscal(), cggo(), chktcg1(), crespsp(), gammam1(), get_local_crs_galerkin(), helmholtz::hconj(), normsc(), ophx(), helmholtz::projh(), and wlaplacian().
Here is the call graph for this function: Here is the caller graph for this function:| subroutine cggo | ( | real(dp), dimension(lx1*ly1*lz1,*), intent(out) | x, |
| real(dp), dimension(lx1*ly1*lz1,*), intent(in) | f, | ||
| real(dp), dimension(lx1*ly1*lz1,*), intent(in) | h1, | ||
| real(dp), dimension(lx1*ly1*lz1,*), intent(in) | h2, | ||
| real(dp), dimension(lx1*ly1*lz1,*), intent(in) | mask, | ||
| real(dp), dimension(lx1*ly1*lz1,*), intent(in) | mult, | ||
| integer, intent(in) | imsh, | ||
| real(dp), intent(in) | tin, | ||
| integer, intent(in) | maxit, | ||
| integer, intent(in) | isd, | ||
| real(dp), dimension(lx1*ly1*lz1,*), intent(in) | binv, | ||
| character(4) | name | ||
| ) |
Solve the Helmholtz equation, H*U = RHS, using preconditioned conjugate gradient iteration. Preconditioner: diag(H).
| [out] | x | solution vector |
| [in] | f | residual vector |
| [in] | h1 | coefficient of A (stiffness) |
| [in] | h2 | coefficient of M (mass) |
| [in] | mask | mask array |
| [in] | mult | multiplicity array |
| [in] | binv | inverse of mass matrix |
| [in] | tin | input tolerance |
Definition at line 744 of file hmholtz.F90.
References axhelm(), copy(), ctimer::dnekclock(), dssum(), gop(), hmh_gmres(), setfast(), setprec(), and sum().
Referenced by hmholtz(), and helmholtz::hmhzpf().
Here is the call graph for this function: Here is the caller graph for this function:| subroutine chktcg1 | ( | real(dp), intent(out) | tol, |
| real(dp), dimension (lx1,ly1,lz1,*), intent(in) | res, | ||
| real(dp), dimension (lx1,ly1,lz1,*), intent(in) | h1, | ||
| real(dp), dimension (lx1,ly1,lz1,*), intent(in) | h2, | ||
| real(dp), dimension (lx1,ly1,lz1,*), intent(in) | mask, | ||
| real(dp), dimension (lx1,ly1,lz1,*), intent(in) | mult, | ||
| integer, intent(in) | imesh, | ||
| integer, intent(in) | isd | ||
| ) |
Check that the tolerances are not too small for the CG-solver. Important when calling the CG-solver (Gauss-Lobatto mesh) with zero Neumann b.c.
Definition at line 637 of file hmholtz.F90.
References axhelm(), and ctimer::dnekclock().
Referenced by hmh_gmres(), hmholtz(), and helmholtz::hmhzpf().
Here is the call graph for this function: Here is the caller graph for this function:| subroutine generalev | ( | real(dp), dimension(n,n), intent(inout) | a, |
| real(dp), dimension(n,n), intent(inout) | b, | ||
| real(dp), dimension(n), intent(out) | lam, | ||
| integer, intent(in) | n, | ||
| real(dp), dimension(n,n), intent(in) | w | ||
| ) |
Solve the generalized eigenvalue problem A x = lam B x.
A – symm. B – symm., pos. definite
"SIZE" is included here only to deduce WDSIZE, the working precision, in bytes, so as to know whether dsygv or ssygv should be called.
| [in] | w | unused |
Definition at line 1101 of file hmholtz.F90.
References copy(), exitt(), and outmat2().
Referenced by hsmg_routines::hsmg_setup_fast1d().
Here is the call graph for this function: Here is the caller graph for this function:| subroutine hmholtz | ( | character(4) | name, |
| real(dp), dimension (lx1,ly1,lz1,*) | u, | ||
| real(dp), dimension (lx1,ly1,lz1,*) | rhs, | ||
| real(dp), dimension (lx1,ly1,lz1,*) | h1, | ||
| real(dp), dimension (lx1,ly1,lz1,*) | h2, | ||
| real(dp), dimension (lx1,ly1,lz1,*) | mask, | ||
| real(dp), dimension (lx1,ly1,lz1,*) | mult, | ||
| integer | imsh, | ||
| real(dp) | tli, | ||
| integer | maxit, | ||
| integer | isd | ||
| ) |
Definition at line 2 of file hmholtz.F90.
References cggo(), chcopy(), chktcg1(), ctimer::dnekclock(), and dssum().
Referenced by helmholtz::hsolve().
Here is the call graph for this function: Here is the caller graph for this function:| subroutine outmat2 | ( | real(dp), dimension(m,n) | a, |
| integer | m, | ||
| integer | n, | ||
| integer | k, | ||
| character(4) | name | ||
| ) |
Definition at line 1150 of file hmholtz.F90.
Referenced by generalev().
Here is the caller graph for this function:| subroutine set_fdm_prec_h1b | ( | real(dp), dimension (nx1,ny1,nz1,1), intent(out) | d, |
| real(dp), dimension(nx1,ny1,nz1,1), intent(in) | h1, | ||
| real(dp), dimension(nx1,ny1,nz1,1), intent(in) | h2, | ||
| integer, intent(in) | nel | ||
| ) |
Definition at line 989 of file hmholtz.F90.
| subroutine setfast | ( | real(dp), dimension(nx1,ny1,nz1,*), intent(in) | helm1, |
| real(dp), dimension(nx1,ny1,nz1,*), intent(in) | helm2, | ||
| integer, intent(in) | imesh | ||
| ) |
Set logicals for fast evaluation of A*x.
Definition at line 329 of file hmholtz.F90.
References ctimer::dnekclock().
Referenced by axhelm(), and cggo().
Here is the call graph for this function: Here is the caller graph for this function:| subroutine setprec | ( | real(dp), dimension (lx1,ly1,lz1,*), intent(out) | dpcm1, |
| real(dp), dimension(nx1,ny1,nz1,*), intent(in) | helm1, | ||
| real(dp), dimension(nx1,ny1,nz1,*), intent(in) | helm2, | ||
| integer, intent(in) | imsh, | ||
| integer, intent(in) | isd | ||
| ) |
Generate diagonal preconditioner for the Helmholtz operator.
Definition at line 464 of file hmholtz.F90.
References ctimer::dnekclock(), dssum(), and mxm().
Referenced by cggo().
Here is the call graph for this function: Here is the caller graph for this function:| subroutine sfastax | ( | ) |
For undeformed elements, set up appropriate elemental matrices and geometric factors for fast evaluation of Ax.
Definition at line 387 of file hmholtz.F90.
Referenced by gengeom(), and geom_reset().
Here is the caller graph for this function:
1.8.8