Classes
struct	Chroma::HBParams
	Heat-bath params. More...

Functions
void	Chroma::mciter (multi1d< LatticeColorMatrix > &u, const LinearGaugeAction &S_g, const HBParams &hbp)
	One heatbath interation of updating the gauge field configuration. More...

void	Chroma::mciter32 (multi1d< LatticeColorMatrix > &u, int n_over, int nheat, int NmaxHB, int &ntrials, int &nfails, const Set &ss, const multi3d< int > &neighsubl)
	One heatbath interation of updating the gauge field configuration. More...

void	Chroma::monte (multi1d< LatticeColorMatrix > &u(Nd), int NOver, int NHeat, int j_decay, int GlueP, int WilsonP, int &NumTry, int &NumFail, int PrtMes, int MesItr, int NumItr, int TotalItr, int TopolP, int TopolItr, Real TopAccu, Real ActAccu, int NumTop, int NumCool, Real BlkAccu, int BlkMax, int HQPotP, int HQPotItr, Real FacSmea, int NumSmea, int GFixP, int GFixItr, int ORlxDo, Real OrPara,)
	Routine for doing the hybrid (monte carlo) algorithm. More...

void	Chroma::su3hb (LatticeColorMatrix &u, const LatticeColorMatrix &w, int su2_index, int nheat, int &ntrials, int &nfails, HeatbathType algorithm, const Subset &sub)
	Do one SU(2) subgroup heatbath update of SU(Nc) matrix U with action. More...

void	Chroma::su3over (LatticeColorMatrix &u, const LatticeColorMatrix &w, int su2_index, const Subset &sub)
	Do one SU(2) subgroup microcanonical overrelaxation update of SU(Nc) matrix. More...

Detailed Description

Heatbath gauge field update routines

Function Documentation

◆ mciter()

void Chroma::mciter	(	multi1d< LatticeColorMatrix > &	u,
		const LinearGaugeAction &	S_g,
		const HBParams &	hbp
	)

One heatbath interation of updating the gauge field configuration.

Make one interation of updating the gauge field configuration: this consists of n_over overrelaxation sweeps followed by one heatbath sweep with nheat trials. In the case of SU(3), for each link we loop over the 3 SU(2) subgroups.

Warning: this works only for Nc = 2 and 3 !

Parameters

u	gauge field ( Modify )
S_g	gauge action ( Read )
hbp	heatbath parameters ( Read )

Definition at line 30 of file mciter.cc.

References Chroma::cb, Chroma::GaugeAction< P, Q >::createState(), Chroma::END_CODE(), Chroma::GaugeAction< P, Q >::getGaugeBC(), Chroma::GaugeAction< P, Q >::getSet(), Chroma::GaugeBC< P, Q >::modify(), mu, Nd, Chroma::HBParams::nmax(), Chroma::HBParams::nOver, Chroma::reunit(), Chroma::LinearGaugeAction::staple(), Chroma::START_CODE(), Chroma::state, Chroma::su2_hb_update(), su2_index, Chroma::su3over(), and Chroma::u.

Referenced by Chroma::doProd(), Chroma::doWarmUp(), and Chroma::monte().

◆ mciter32()

void Chroma::mciter32	(	multi1d< LatticeColorMatrix > &	u,
		int	n_over,
		int	nheat,
		int	NmaxHB,
		int &	ntrials,
		int &	nfails,
		const Set &	ss,
		const multi3d< int > &	neighsubl
	)

One heatbath interation of updating the gauge field configuration.

Make one interation of updating the gauge field configuration: for Wilson or Symmanzik improved pure gauge action: this consists of n_over overrelaxation sweeps followed by one heatbath sweep with nheat trials. In the case of SU(3), for each link we loop over the 3 SU(2) subgroups.

Because of the option for the Symanzik improved action we need 2^(d+1) sublattices, deviding first into 2^d hypercubes and then checkerboarding those.

Warning: this works only for Nc = 2 and 3 !

Parameters

u	gauge field ( Modify )
n_over	number of overrelaxation sweeps ( Read )
nheat	number of heatbath trials ( Read )
NmaxHB	max number of heatbath hits ( Read )
ntrials	total number of individual heatbath trials ( Modify )
nfails	total number of individual heatbath failures ( Modify )

Definition at line 53 of file mciter32.cc.

References BACKWARD, copymask(), Chroma::END_CODE(), Chroma::eta, FORWARD, GlueCoeffPG(), GlueCoeffRT(), GlueImp, igluetmp, mu, Nd, nu, Chroma::QDP_error_exit(), Chroma::reunit(), Chroma::START_CODE(), Chroma::su2_hb_update(), su2_index, Chroma::su3over(), tmp_1, tmp_2, and Chroma::u.

◆ monte()

void Chroma::monte	(	multi1d< LatticeColorMatrix > &	uNd,
		int	NOver,
		int	NHeat,
		int	j_decay,
		int	GlueP,
		int	WilsonP,
		int &	NumTry,
		int &	NumFail,
		int	PrtMes,
		int	MesItr,
		int	NumItr,
		int	TotalItr,
		int	TopolP,
		int	TopolItr,
		Real	TopAccu,
		Real	ActAccu,
		int	NumTop,
		int	NumCool,
		Real	BlkAccu,
		int	BlkMax,
		int	HQPotP,
		int	HQPotItr,
		Real	FacSmea,
		int	NumSmea,
		int	GFixP,
		int	GFixItr,
		int	ORlxDo,
		Real	OrPara
	)

Routine for doing the hybrid (monte carlo) algorithm.

Parameters

u	gauge field ( Modify )
NOver	number of overrelaxations steps ( Read )
NHeat	number of heatbath trials ( Read )
j_decay	direction of exponential decay — for sideways potl ! ( Read )
GlueP	flag for glueball measurements ( Read )
WilsonP	flag for Wilson loop measurements ( Read )
TopolP	flag for topology measurements ( Read )
PrtMes	measurements per printout to RESULT file
MesItr	number of iterations in mod for measurements ( Read )
NumTry	number of heatbath link trials performed ( Write )
NumFail	number of heatbath link trials failed ( Write )
NumItr	number of iterations to perform ( Read )
TotalItr	total number of iterations to perform ( Read )
BlkMax	maximum number of SU(3) trace maximimations allowed ( Read )
BlkAccu	blocking accuracy in glueball code ( Read )
TopolItr	number of iterations in mod for topology measurement ( Read )
TopAccu	accuracy for convergence of topological charge ( Read )
ActAccu	accuracy for convergence of action ratio ( Read )
NumTop	number of topological charge measurements ( Read )
NumCool	number of cooling sweeps per topological charge measurement ( Read )
HQPotP	flag for heavy quark potential measurements ( Read )
HQPotItr	number of iterations in mod for potential measurement ( Read )
FacSmea	smearing factor for potential measurement ( Read )
NumSmea	number of smearing iterations in potential measurement ( Read )
GFixP	flag for Landau gauge fixing ( Read )
GFixItr	number of iterations in mod for Landau gauge fixing ( Read )
ORlxDo	do over-relaxation in gauge fixing ( Read )
OrPara	over-relaxation parameter in gauge-fixing ( Read )

Definition at line 56 of file monte.cc.

References ActAccu, Chroma::END_CODE(), Chroma::fuzglue(), Chroma::i, j_decay, link, Chroma::mciter(), Chroma::MesPlq(), mu, Nd, NumCool, NumTop, Chroma::polylp(), Chroma::pop(), Chroma::push(), s_plaq, Chroma::START_CODE(), t_plaq, TopAccu, Chroma::u, w_plaq, Chroma::wilslp(), Chroma::write(), and YES.

◆ su3hb()

void Chroma::su3hb	(	LatticeColorMatrix &	u,
		const LatticeColorMatrix &	w,
		int	su2_index,
		int	nheat,
		int &	ntrials,
		int &	nfails,
		HeatbathType	algorithm,
		const Subset &	sub
	)

Do one SU(2) subgroup heatbath update of SU(Nc) matrix U with action.

BetaMC * [tr(U*W) + hc] / (2Nc). Do at most nheat update trials of the Kennedy-Pstd::endleton or Creutz SU(2) heatbath algorithm.

Parameters

u	field to be updated ( Modify )
w	"staple" field in the action ( Read )
su2_index	SU(2) subgroup index ( Read )
nheat	maximal number of update trials ( Read )
ntrials	total number of link trials ( Write )
nfails	total number of failed trials ( Write )
sub	Subset for updating ( Read )