host_utils.cpp File Reference

#include <limits>
#include <complex>
#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#include <short.h>
#include <comm_quda.h>
#include <mpi_comm_handle.h>
#include <color_spinor_field.h>
#include <unitarization_links.h>
#include <dslash_quda.h>
#include <llfat_utils.h>
#include <staggered_gauge_utils.h>
#include <host_utils.h>
#include <command_line_params.h>
#include <misc.h>
#include <qio_field.h>

Go to the source code of this file.

Macros
#define	MAX(a, b)   ((a) > (b) ? (a) : (b))
#define	XUP   0
#define	YUP   1
#define	ZUP   2
#define	TUP   3

Functions
void	setQudaPrecisions ()
void	setQudaMgSolveTypes ()
void	setQudaDefaultMgTestParams ()
void	constructQudaGaugeField (void **gauge, int type, QudaPrecision precision, QudaGaugeParam *param)
void	constructHostGaugeField (void **gauge, QudaGaugeParam &gauge_param, int argc, char **argv)
void	constructHostCloverField (void *clover, void *clover_inv, QudaInvertParam &inv_param)
void	constructQudaCloverField (void *clover, double norm, double diag, QudaPrecision precision)
void	constructWilsonTestSpinorParam (quda::ColorSpinorParam *cs_param, const QudaInvertParam *inv_param, const QudaGaugeParam *gauge_param)
void	constructRandomSpinorSource (void *v, int nSpin, int nColor, QudaPrecision precision, QudaSolutionType sol_type, const int *const x, quda::RNG &rng)
void	initComms (int argc, char **argv, std::array< int, 4 > &commDims)
void	initComms (int argc, char **argv, int *const commDims)
void	finalizeComms ()
void	initRand ()
void	setDims (int *X)
void	dw_setDims (int *X, const int L5)
int	dimPartitioned (int dim)
bool	last_node_in_t ()
int	index_4d_cb_from_coordinate_4d (const int coordinate[4], const int dim[4])
void	coordinate_from_shrinked_index (int coordinate[4], int shrinked_index, const int shrinked_dim[4], const int shift[4], int parity)
int	neighborIndex (int i, int oddBit, int dx4, int dx3, int dx2, int dx1)
int	neighborIndex (int dim[4], int index, int oddBit, int dx[4])
int	neighborIndex_mg (int i, int oddBit, int dx4, int dx3, int dx2, int dx1)
int	neighborIndexFullLattice (int i, int dx4, int dx3, int dx2, int dx1)
int	neighborIndexFullLattice (int dim[4], int index, int dx[4])
int	neighborIndexFullLattice_mg (int i, int dx4, int dx3, int dx2, int dx1)
void	printSpinorElement (void *spinor, int X, QudaPrecision precision)
void	printGaugeElement (void *gauge, int X, QudaPrecision precision)
int	fullLatticeIndex (int dim[4], int index, int oddBit)
int	fullLatticeIndex (int i, int oddBit)
const char *	__asan_default_options ()
	Set the default ASAN options. This ensures that QUDA just works when SANITIZE is enabled without requiring ASAN_OPTIONS to be set. More...
void	get_size_from_env (int *const dims, const char env[])
int	lex_rank_from_coords_t (const int *coords, void *fdata)
int	lex_rank_from_coords_x (const int *coords, void *fdata)
template<typename Float >
void	printVector (Float *v)
int	getOddBit (int Y)
template<typename Float >
void	complexAddTo (Float *a, Float *b)
template<typename Float >
void	complexProduct (Float *a, Float *b, Float *c)
template<typename Float >
void	complexConjugateProduct (Float *a, Float *b, Float *c)
template<typename Float >
void	complexDotProduct (Float *a, Float *b, Float *c)
template<typename Float >
void	accumulateComplexProduct (Float *a, Float *b, Float *c, Float sign)
template<typename Float >
void	accumulateComplexDotProduct (Float *a, Float *b, Float *c)
template<typename Float >
void	accumulateConjugateProduct (Float *a, Float *b, Float *c, int sign)
template<typename Float >
void	su3Construct12 (Float *mat)
template<typename Float >
void	su3Construct8 (Float *mat)
void	su3_construct (void *mat, QudaReconstructType reconstruct, QudaPrecision precision)
void	su3_reconstruct (void *mat, int dir, int ga_idx, QudaReconstructType reconstruct, QudaPrecision precision, QudaGaugeParam *param)
int	compare_floats (void *a, void *b, int len, double epsilon, QudaPrecision precision)
int	fullLatticeIndex_4d (int i, int oddBit)
int	fullLatticeIndex_5d (int i, int oddBit)
int	fullLatticeIndex_5d_4dpc (int i, int oddBit)
int	x4_from_full_index (int i)
template<typename Float >
void	applyGaugeFieldScaling (Float **gauge, int Vh, QudaGaugeParam *param)
template<typename Float >
void	constructUnitGaugeField (Float **res, QudaGaugeParam *param)
template void	constructUnitGaugeField (float **res, QudaGaugeParam *param)
template void	constructUnitGaugeField (double **res, QudaGaugeParam *param)
template<typename Float >
void	constructRandomGaugeField (Float **res, QudaGaugeParam *param, QudaDslashType dslash_type)
template void	constructRandomGaugeField (float **res, QudaGaugeParam *param, QudaDslashType dslash_type)
template void	constructRandomGaugeField (double **res, QudaGaugeParam *param, QudaDslashType dslash_type)
template<typename Float >
void	constructUnitaryGaugeField (Float **res)
template<typename Float >
void	constructCloverField (Float *res, double norm, double diag)
void	check_gauge (void **oldG, void **newG, double epsilon, QudaPrecision precision)
void	createSiteLinkCPU (void **link, QudaPrecision precision, int phase)
template<typename Float >
int	compareLink (Float **linkA, Float **linkB, int len)
int	strong_check_link (void **linkA, const char *msgA, void **linkB, const char *msgB, int len, QudaPrecision prec)
void	createMomCPU (void *mom, QudaPrecision precision)
void	createHwCPU (void *hw, QudaPrecision precision)
template<typename Float >
int	compare_mom (Float *momA, Float *momB, int len)
int	strong_check_mom (void *momA, void *momB, int len, QudaPrecision prec)
template<typename real >
double	mom_action (real *mom_, int len)
double	mom_action (void *mom, QudaPrecision prec, int len)
	Host reference implementation of the momentum action contribution. More...
void	stopwatchStart ()
double	stopwatchReadSeconds ()
void	performanceStats (std::vector< double > &time, std::vector< double > &gflops, std::vector< int > &iter)

Variables
int	Z [4]
int	V
int	Vh
int	Vs_x
int	Vs_y
int	Vs_z
int	Vs_t
int	Vsh_x
int	Vsh_y
int	Vsh_z
int	Vsh_t
int	faceVolume [4]
int	E1
int	E1h
int	E2
int	E3
int	E4
int	E [4]
int	V_ex
int	Vh_ex
int	Ls
int	V5
int	V5h
double	kappa5
float	fat_link_max
QudaPrecision	local_prec = QUDA_DOUBLE_PRECISION
QudaPrecision &	cpu_prec = local_prec
QudaPrecision &	cuda_prec = prec
QudaPrecision &	cuda_prec_sloppy = prec_sloppy
QudaPrecision &	cuda_prec_refinement_sloppy = prec_refinement_sloppy
QudaPrecision &	cuda_prec_precondition = prec_precondition
QudaPrecision &	cuda_prec_eigensolver = prec_eigensolver
QudaPrecision &	cuda_prec_ritz = prec_ritz
size_t	host_gauge_data_type_size = (cpu_prec == QUDA_DOUBLE_PRECISION) ? sizeof(double) : sizeof(float)
size_t	host_spinor_data_type_size = (cpu_prec == QUDA_DOUBLE_PRECISION) ? sizeof(double) : sizeof(float)
size_t	host_clover_data_type_size = (cpu_prec == QUDA_DOUBLE_PRECISION) ? sizeof(double) : sizeof(float)

Macro Definition Documentation

MAX

#define MAX	(		a,
			b
	)		((a) > (b) ? (a) : (b))

Definition at line 27 of file host_utils.cpp.

TUP

#define TUP   3

Definition at line 34 of file host_utils.cpp.

XUP

#define XUP   0

Definition at line 31 of file host_utils.cpp.

YUP

#define YUP   1

Definition at line 32 of file host_utils.cpp.

ZUP

#define ZUP   2

Definition at line 33 of file host_utils.cpp.

Function Documentation

__asan_default_options()

const char* __asan_default_options

(

)

Set the default ASAN options. This ensures that QUDA just works when SANITIZE is enabled without requiring ASAN_OPTIONS to be set.

Definition at line 638 of file host_utils.cpp.

accumulateComplexDotProduct()

template<typename Float >

void accumulateComplexDotProduct ( Float *  a, Float *  b, Float *  c  )

inline

Definition at line 727 of file host_utils.cpp.

accumulateComplexProduct()

template<typename Float >

void accumulateComplexProduct ( Float *  a, Float *  b, Float *  c, Float  sign  )

inline

Definition at line 720 of file host_utils.cpp.

accumulateConjugateProduct()

template<typename Float >

void accumulateConjugateProduct ( Float *  a, Float *  b, Float *  c, int  sign  )

inline

Definition at line 733 of file host_utils.cpp.

applyGaugeFieldScaling()

template<typename Float >

void applyGaugeFieldScaling	(	Float **	gauge,
		int	Vh,
		QudaGaugeParam *	param
	)

Definition at line 968 of file host_utils.cpp.

check_gauge()

void check_gauge	(	void **	oldG,
		void **	newG,
		double	epsilon,
		QudaPrecision	precision
	)

Definition at line 1255 of file host_utils.cpp.

compare_floats()

int compare_floats	(	void *	a,
		void *	b,
		int	len,
		double	epsilon,
		QudaPrecision	precision
	)

Definition at line 889 of file host_utils.cpp.

compare_mom()

template<typename Float >

int compare_mom	(	Float *	momA,
		Float *	momB,
		int	len
	)

Definition at line 1504 of file host_utils.cpp.

compareLink()

template<typename Float >

int compareLink	(	Float **	linkA,
		Float **	linkB,
		int	len
	)

Definition at line 1367 of file host_utils.cpp.

complexAddTo()

template<typename Float >

void complexAddTo ( Float *  a, Float *  b  )

inline

Definition at line 692 of file host_utils.cpp.

complexConjugateProduct()

template<typename Float >

void complexConjugateProduct ( Float *  a, Float *  b, Float *  c  )

inline

Definition at line 706 of file host_utils.cpp.

complexDotProduct()

template<typename Float >

void complexDotProduct ( Float *  a, Float *  b, Float *  c  )

inline

Definition at line 713 of file host_utils.cpp.

complexProduct()

template<typename Float >

void complexProduct ( Float *  a, Float *  b, Float *  c  )

inline

Definition at line 699 of file host_utils.cpp.

constructCloverField()

template<typename Float >

void constructCloverField	(	Float *	res,
		double	norm,
		double	diag
	)

Definition at line 1189 of file host_utils.cpp.

constructHostCloverField()

void constructHostCloverField	(	void *	clover,
		void *	clover_inv,
		QudaInvertParam &	inv_param
	)

Definition at line 186 of file host_utils.cpp.

constructHostGaugeField()

void constructHostGaugeField	(	void **	gauge,
		QudaGaugeParam &	gauge_param,
		int	argc,
		char **	argv
	)

Definition at line 166 of file host_utils.cpp.

constructQudaCloverField()

void constructQudaCloverField	(	void *	clover,
		double	norm,
		double	diag,
		QudaPrecision	precision
	)

Definition at line 199 of file host_utils.cpp.

constructQudaGaugeField()

void constructQudaGaugeField	(	void **	gauge,
		int	type,
		QudaPrecision	precision,
		QudaGaugeParam *	param
	)

Definition at line 146 of file host_utils.cpp.

constructRandomGaugeField() [1/3]

template void constructRandomGaugeField	(	double **	res,
		QudaGaugeParam *	param,
		QudaDslashType	dslash_type
	)

constructRandomGaugeField() [2/3]

template<typename Float >

void constructRandomGaugeField	(	Float **	res,
		QudaGaugeParam *	param,
		QudaDslashType	dslash_type
	)

Definition at line 1049 of file host_utils.cpp.

constructRandomGaugeField() [3/3]

template void constructRandomGaugeField	(	float **	res,
		QudaGaugeParam *	param,
		QudaDslashType	dslash_type
	)

constructRandomSpinorSource()

void constructRandomSpinorSource	(	void *	v,
		int	nSpin,
		int	nColor,
		QudaPrecision	precision,
		QudaSolutionType	sol_type,
		const int *const	x,
		quda::RNG &	rng
	)

Definition at line 235 of file host_utils.cpp.

constructUnitaryGaugeField()

template<typename Float >

void constructUnitaryGaugeField

(

Float **

res

)

Definition at line 1130 of file host_utils.cpp.

constructUnitGaugeField() [1/3]

template void constructUnitGaugeField	(	double **	res,
		QudaGaugeParam *	param
	)

constructUnitGaugeField() [2/3]

template<typename Float >

void constructUnitGaugeField	(	Float **	res,
		QudaGaugeParam *	param
	)

Definition at line 1006 of file host_utils.cpp.

constructUnitGaugeField() [3/3]

template void constructUnitGaugeField	(	float **	res,
		QudaGaugeParam *	param
	)

constructWilsonTestSpinorParam()

void constructWilsonTestSpinorParam	(	quda::ColorSpinorParam *	cs_param,
		const QudaInvertParam *	inv_param,
		const QudaGaugeParam *	gauge_param
	)

Definition at line 207 of file host_utils.cpp.

coordinate_from_shrinked_index()

void coordinate_from_shrinked_index	(	int	coordinate[4],
		int	shrinked_index,
		const int	shrinked_dim[4],
		const int	shift[4],
		int	parity
	)

Definition at line 393 of file host_utils.cpp.

createHwCPU()

void createHwCPU	(	void *	hw,
		QudaPrecision	precision
	)

Definition at line 1485 of file host_utils.cpp.

createMomCPU()

void createMomCPU	(	void *	mom,
		QudaPrecision	precision
	)

Definition at line 1450 of file host_utils.cpp.

createSiteLinkCPU()

void createSiteLinkCPU	(	void **	link,
		QudaPrecision	precision,
		int	phase
	)

Definition at line 1263 of file host_utils.cpp.

dimPartitioned()

int dimPartitioned

(

int

dim

)

Definition at line 376 of file host_utils.cpp.

dw_setDims()

void dw_setDims	(	int *	X,
		const int	L5
	)

Definition at line 353 of file host_utils.cpp.

finalizeComms()

void finalizeComms

(

)

Definition at line 292 of file host_utils.cpp.

fullLatticeIndex() [1/2]

int fullLatticeIndex	(	int	dim[4],
		int	index,
		int	oddBit
	)

Definition at line 591 of file host_utils.cpp.

fullLatticeIndex() [2/2]

int fullLatticeIndex	(	int	i,
		int	oddBit
	)

Definition at line 605 of file host_utils.cpp.

fullLatticeIndex_4d()

int fullLatticeIndex_4d	(	int	i,
		int	oddBit
	)

Definition at line 902 of file host_utils.cpp.

fullLatticeIndex_5d()

int fullLatticeIndex_5d	(	int	i,
		int	oddBit
	)

Definition at line 940 of file host_utils.cpp.

fullLatticeIndex_5d_4dpc()

int fullLatticeIndex_5d_4dpc	(	int	i,
		int	oddBit
	)

Definition at line 947 of file host_utils.cpp.

get_size_from_env()

void get_size_from_env	(	int *const	dims,
		const char	env[]
	)

For MPI, the default node mapping is lexicographical with t varying fastest.

Definition at line 645 of file host_utils.cpp.

getOddBit()

int getOddBit

(

int

Y

)

Definition at line 682 of file host_utils.cpp.

index_4d_cb_from_coordinate_4d()

int index_4d_cb_from_coordinate_4d	(	const int	coordinate[4],
		const int	dim[4]
	)

Definition at line 388 of file host_utils.cpp.

initComms() [1/2]

void initComms	(	int	argc,
		char **	argv,
		int *const	commDims
	)

Definition at line 257 of file host_utils.cpp.

initComms() [2/2]

void initComms	(	int	argc,
		char **	argv,
		std::array< int, 4 > &	commDims
	)

Definition at line 255 of file host_utils.cpp.

initRand()

void initRand

(

)

Definition at line 302 of file host_utils.cpp.

last_node_in_t()

bool last_node_in_t

(

)

Definition at line 378 of file host_utils.cpp.

lex_rank_from_coords_t()

int lex_rank_from_coords_t	(	const int *	coords,
		void *	fdata
	)

Definition at line 662 of file host_utils.cpp.

lex_rank_from_coords_x()

int lex_rank_from_coords_x	(	const int *	coords,
		void *	fdata
	)

Definition at line 669 of file host_utils.cpp.

mom_action() [1/2]

template<typename real >

double mom_action	(	real *	mom_,
		int	len
	)

Definition at line 1586 of file host_utils.cpp.

mom_action() [2/2]

double mom_action	(	void *	mom,
		QudaPrecision	prec,
		int	len
	)

Host reference implementation of the momentum action contribution.

Definition at line 1601 of file host_utils.cpp.

neighborIndex() [1/2]

int neighborIndex	(	int	dim[4],
		int	index,
		int	oddBit,
		int	dx[4]
	)

Definition at line 441 of file host_utils.cpp.

neighborIndex() [2/2]

int neighborIndex	(	int	i,
		int	oddBit,
		int	dx4,
		int	dx3,
		int	dx2,
		int	dx1
	)

Definition at line 423 of file host_utils.cpp.

neighborIndex_mg()

int neighborIndex_mg	(	int	i,
		int	oddBit,
		int	dx4,
		int	dx3,
		int	dx2,
		int	dx1
	)

Definition at line 457 of file host_utils.cpp.

neighborIndexFullLattice() [1/2]

int neighborIndexFullLattice	(	int	dim[4],
		int	index,
		int	dx[4]
	)

Definition at line 508 of file host_utils.cpp.

neighborIndexFullLattice() [2/2]

int neighborIndexFullLattice	(	int	i,
		int	dx4,
		int	dx3,
		int	dx2,
		int	dx1
	)

Definition at line 490 of file host_utils.cpp.

neighborIndexFullLattice_mg()

int neighborIndexFullLattice_mg	(	int	i,
		int	dx4,
		int	dx3,
		int	dx2,
		int	dx1
	)

Definition at line 528 of file host_utils.cpp.

performanceStats()

void performanceStats	(	std::vector< double > &	time,
		std::vector< double > &	gflops,
		std::vector< int > &	iter
	)

Definition at line 1627 of file host_utils.cpp.

printGaugeElement()

void printGaugeElement	(	void *	gauge,
		int	X,
		QudaPrecision	precision
	)

Definition at line 575 of file host_utils.cpp.

printSpinorElement()

void printSpinorElement	(	void *	spinor,
		int	X,
		QudaPrecision	precision
	)

Definition at line 566 of file host_utils.cpp.

printVector()

template<typename Float >

void printVector

(

Float *

v

)

Definition at line 676 of file host_utils.cpp.

setDims()

void setDims

(

int *

X

)

Definition at line 315 of file host_utils.cpp.

setQudaDefaultMgTestParams()

void setQudaDefaultMgTestParams

(

)

Definition at line 89 of file host_utils.cpp.

setQudaMgSolveTypes()

void setQudaMgSolveTypes

(

)

Definition at line 81 of file host_utils.cpp.

setQudaPrecisions()

void setQudaPrecisions

(

)

Definition at line 69 of file host_utils.cpp.

stopwatchReadSeconds()

double stopwatchReadSeconds

(

)

Definition at line 1617 of file host_utils.cpp.

stopwatchStart()

void stopwatchStart

(

)

Definition at line 1615 of file host_utils.cpp.

strong_check_link()

int strong_check_link	(	void **	linkA,
		const char *	msgA,
		void **	linkB,
		const char *	msgB,
		int	len,
		QudaPrecision	prec
	)

Definition at line 1428 of file host_utils.cpp.

strong_check_mom()

int strong_check_mom	(	void *	momA,
		void *	momB,
		int	len,
		QudaPrecision	prec
	)

Definition at line 1551 of file host_utils.cpp.

su3_construct()

void su3_construct	(	void *	mat,
		QudaReconstructType	reconstruct,
		QudaPrecision	precision
	)

Definition at line 758 of file host_utils.cpp.

su3_reconstruct()

void su3_reconstruct	(	void *	mat,
		int	dir,
		int	ga_idx,
		QudaReconstructType	reconstruct,
		QudaPrecision	precision,
		QudaGaugeParam *	param
	)

Definition at line 861 of file host_utils.cpp.

su3Construct12()

template<typename Float >

void su3Construct12 ( Float *  mat )

inline

Definition at line 739 of file host_utils.cpp.

su3Construct8()

template<typename Float >

void su3Construct8 ( Float *  mat )

inline

Definition at line 751 of file host_utils.cpp.

x4_from_full_index()

int x4_from_full_index

(

int

i

)

Definition at line 953 of file host_utils.cpp.

Variable Documentation

cpu_prec

QudaPrecision& cpu_prec = local_prec

Definition at line 57 of file host_utils.cpp.

cuda_prec

QudaPrecision& cuda_prec = prec

Definition at line 58 of file host_utils.cpp.

cuda_prec_eigensolver

QudaPrecision& cuda_prec_eigensolver = prec_eigensolver

Definition at line 62 of file host_utils.cpp.

cuda_prec_precondition

QudaPrecision& cuda_prec_precondition = prec_precondition

Definition at line 61 of file host_utils.cpp.

cuda_prec_refinement_sloppy

QudaPrecision& cuda_prec_refinement_sloppy = prec_refinement_sloppy

Definition at line 60 of file host_utils.cpp.

cuda_prec_ritz

QudaPrecision& cuda_prec_ritz = prec_ritz

Definition at line 63 of file host_utils.cpp.

cuda_prec_sloppy

QudaPrecision& cuda_prec_sloppy = prec_sloppy

Definition at line 59 of file host_utils.cpp.

E

int E[4]

Definition at line 45 of file host_utils.cpp.

E1

int E1

Definition at line 44 of file host_utils.cpp.

E1h

int E1h

Definition at line 44 of file host_utils.cpp.

E2

int E2

Definition at line 44 of file host_utils.cpp.

E3

int E3

Definition at line 44 of file host_utils.cpp.

E4

int E4

Definition at line 44 of file host_utils.cpp.

faceVolume

int faceVolume[4]

Definition at line 41 of file host_utils.cpp.

fat_link_max

float fat_link_max

extern

host_clover_data_type_size

size_t host_clover_data_type_size = (cpu_prec == QUDA_DOUBLE_PRECISION) ? sizeof(double) : sizeof(float)

Definition at line 67 of file host_utils.cpp.

host_gauge_data_type_size

size_t host_gauge_data_type_size = (cpu_prec == QUDA_DOUBLE_PRECISION) ? sizeof(double) : sizeof(float)

Definition at line 65 of file host_utils.cpp.

host_spinor_data_type_size

size_t host_spinor_data_type_size = (cpu_prec == QUDA_DOUBLE_PRECISION) ? sizeof(double) : sizeof(float)

Definition at line 66 of file host_utils.cpp.

kappa5

double kappa5

Definition at line 51 of file host_utils.cpp.

local_prec

QudaPrecision local_prec = QUDA_DOUBLE_PRECISION

Definition at line 56 of file host_utils.cpp.

Ls

int Ls

Definition at line 48 of file host_utils.cpp.

V

int V

Definition at line 37 of file host_utils.cpp.

V5

int V5

Definition at line 49 of file host_utils.cpp.

V5h

int V5h

Definition at line 50 of file host_utils.cpp.

V_ex

int V_ex

Definition at line 46 of file host_utils.cpp.

Vh

int Vh

Definition at line 38 of file host_utils.cpp.

Vh_ex

int Vh_ex

Definition at line 46 of file host_utils.cpp.

Vs_t

int Vs_t

Definition at line 39 of file host_utils.cpp.

Vs_x

int Vs_x

Definition at line 39 of file host_utils.cpp.

Vs_y

int Vs_y

Definition at line 39 of file host_utils.cpp.

Vs_z

int Vs_z

Definition at line 39 of file host_utils.cpp.

Vsh_t

int Vsh_t

Definition at line 40 of file host_utils.cpp.

Vsh_x

int Vsh_x

Definition at line 40 of file host_utils.cpp.

Vsh_y

int Vsh_y

Definition at line 40 of file host_utils.cpp.

Vsh_z

int Vsh_z

Definition at line 40 of file host_utils.cpp.

Z

int Z[4]

Definition at line 36 of file host_utils.cpp.