UpdatePosition.H File Reference

#include "Utils/WarpXConst.H"
#include <AMReX.H>
#include <AMReX_FArrayBox.H>
#include <AMReX_REAL.H>

Go to the source code of this file.

Functions
AMREX_GPU_HOST_DEVICE AMREX_INLINE void	UpdatePosition (amrex::ParticleReal &x, amrex::ParticleReal &y, amrex::ParticleReal &z, const amrex::ParticleReal ux, const amrex::ParticleReal uy, const amrex::ParticleReal uz, const amrex::Real dt, amrex::ParticleReal const mass)
	Push the particle position over one time step, given the value of its momenta ux, uy, uz, using the standard leapfrog algorithm x(t+dt) = x(t) + v(t+dt/2)*dt.
AMREX_GPU_HOST_DEVICE AMREX_FORCE_INLINE amrex::ParticleReal	GetImplicitGammaInverse (const amrex::ParticleReal uxp_n, const amrex::ParticleReal uyp_n, const amrex::ParticleReal uzp_n, const amrex::ParticleReal uxp_nph, const amrex::ParticleReal uyp_nph, const amrex::ParticleReal uzp_nph) noexcept
	Compute the inverse Lorentz factor for the position update in the implicit methods, which is the average of gamma at time levels n and n+1.
AMREX_GPU_HOST_DEVICE AMREX_INLINE void	UpdatePositionImplicit (amrex::ParticleReal &x, amrex::ParticleReal &y, amrex::ParticleReal &z, const amrex::ParticleReal ux_n, const amrex::ParticleReal uy_n, const amrex::ParticleReal uz_n, const amrex::ParticleReal ux, const amrex::ParticleReal uy, const amrex::ParticleReal uz, const amrex::Real dt)
	Push the particle's positions over one timestep, given the value of its momenta ux, uy, uz. The implicit version is the Crank-Nicolson scheme, x^{n+1} - x^{n} = dt*(u^{n+1} + u^{n})/(gamma^{n+1} + gamma^{n}) See Eqs. 15 and 17 in Chen, JCP 407 (2020) 109228.
AMREX_GPU_HOST_DEVICE AMREX_INLINE void	PositionNorm (const amrex::ParticleReal dxp, const amrex::ParticleReal dyp, const amrex::ParticleReal dzp, const amrex::ParticleReal dxp_save, const amrex::ParticleReal dyp_save, const amrex::ParticleReal dzp_save, const amrex::ParticleReal idxg2, const amrex::ParticleReal idyg2, const amrex::ParticleReal idzg2, amrex::ParticleReal &step_norm)
	Check particle position for convergence. This is used by the Picard method used to achieve a self-consistent time-centered update of the particles for given electric and magnetic fields on the grid.

Function Documentation

GetImplicitGammaInverse()

AMREX_GPU_HOST_DEVICE AMREX_FORCE_INLINE amrex::ParticleReal GetImplicitGammaInverse ( const amrex::ParticleReal uxp_n, const amrex::ParticleReal uyp_n, const amrex::ParticleReal uzp_n, const amrex::ParticleReal uxp_nph, const amrex::ParticleReal uyp_nph, const amrex::ParticleReal uzp_nph )

noexcept

Compute the inverse Lorentz factor for the position update in the implicit methods, which is the average of gamma at time levels n and n+1.

PositionNorm()

AMREX_GPU_HOST_DEVICE AMREX_INLINE void PositionNorm	(	const amrex::ParticleReal	dxp,
		const amrex::ParticleReal	dyp,
		const amrex::ParticleReal	dzp,
		const amrex::ParticleReal	dxp_save,
		const amrex::ParticleReal	dyp_save,
		const amrex::ParticleReal	dzp_save,
		const amrex::ParticleReal	idxg2,
		const amrex::ParticleReal	idyg2,
		const amrex::ParticleReal	idzg2,
		amrex::ParticleReal &	step_norm )

Check particle position for convergence. This is used by the Picard method used to achieve a self-consistent time-centered update of the particles for given electric and magnetic fields on the grid.

UpdatePosition()

AMREX_GPU_HOST_DEVICE AMREX_INLINE void UpdatePosition	(	amrex::ParticleReal &	x,
		amrex::ParticleReal &	y,
		amrex::ParticleReal &	z,
		const amrex::ParticleReal	ux,
		const amrex::ParticleReal	uy,
		const amrex::ParticleReal	uz,
		const amrex::Real	dt,
		amrex::ParticleReal const	mass )

Push the particle position over one time step, given the value of its momenta ux, uy, uz, using the standard leapfrog algorithm x(t+dt) = x(t) + v(t+dt/2)*dt.

UpdatePositionImplicit()

AMREX_GPU_HOST_DEVICE AMREX_INLINE void UpdatePositionImplicit	(	amrex::ParticleReal &	x,
		amrex::ParticleReal &	y,
		amrex::ParticleReal &	z,
		const amrex::ParticleReal	ux_n,
		const amrex::ParticleReal	uy_n,
		const amrex::ParticleReal	uz_n,
		const amrex::ParticleReal	ux,
		const amrex::ParticleReal	uy,
		const amrex::ParticleReal	uz,
		const amrex::Real	dt )

Push the particle's positions over one timestep, given the value of its momenta ux, uy, uz. The implicit version is the Crank-Nicolson scheme, x^{n+1} - x^{n} = dt*(u^{n+1} + u^{n})/(gamma^{n+1} + gamma^{n}) See Eqs. 15 and 17 in Chen, JCP 407 (2020) 109228.