janitorial work
This commit is contained in:
parent
1d4823bfd2
commit
8bcc26a9f9
7 changed files with 26 additions and 24 deletions
|
@ -14,7 +14,7 @@ LIBOBJ:=$(LIBOBJ:%=$(OD)/%)
|
||||||
ALLOBJ:=$(ALLOBJ:%=$(OD)/%)
|
ALLOBJ:=$(ALLOBJ:%=$(OD)/%)
|
||||||
|
|
||||||
ODEP:=$(HDR_$(PKG)) $(HDR_libfftpack) $(HDR_c_utils)
|
ODEP:=$(HDR_$(PKG)) $(HDR_libfftpack) $(HDR_c_utils)
|
||||||
$(OD)/sharp_core.o: $(SD)/sharp_inchelper1.inc.c $(SD)/sharp_core_inc.c $(SD)/sharp_core_inc4.c
|
$(OD)/sharp_core.o: $(SD)/sharp_core_inchelper.c $(SD)/sharp_core_inc.c $(SD)/sharp_core_inc2.c
|
||||||
$(OD)/sharp.o: $(SD)/sharp_mpi.c
|
$(OD)/sharp.o: $(SD)/sharp_mpi.c
|
||||||
BDEP:=$(LIB_$(PKG)) $(LIB_libfftpack) $(LIB_c_utils)
|
BDEP:=$(LIB_$(PKG)) $(LIB_libfftpack) $(LIB_c_utils)
|
||||||
|
|
||||||
|
|
|
@ -50,27 +50,27 @@ typedef complex double dcmplx;
|
||||||
#define CONCAT3(a,b,c) XCONCAT3(a,b,c)
|
#define CONCAT3(a,b,c) XCONCAT3(a,b,c)
|
||||||
|
|
||||||
#define nvec 1
|
#define nvec 1
|
||||||
#include "sharp_inchelper1.inc.c"
|
#include "sharp_core_inchelper.c"
|
||||||
#undef nvec
|
#undef nvec
|
||||||
|
|
||||||
#define nvec 2
|
#define nvec 2
|
||||||
#include "sharp_inchelper1.inc.c"
|
#include "sharp_core_inchelper.c"
|
||||||
#undef nvec
|
#undef nvec
|
||||||
|
|
||||||
#define nvec 3
|
#define nvec 3
|
||||||
#include "sharp_inchelper1.inc.c"
|
#include "sharp_core_inchelper.c"
|
||||||
#undef nvec
|
#undef nvec
|
||||||
|
|
||||||
#define nvec 4
|
#define nvec 4
|
||||||
#include "sharp_inchelper1.inc.c"
|
#include "sharp_core_inchelper.c"
|
||||||
#undef nvec
|
#undef nvec
|
||||||
|
|
||||||
#define nvec 5
|
#define nvec 5
|
||||||
#include "sharp_inchelper1.inc.c"
|
#include "sharp_core_inchelper.c"
|
||||||
#undef nvec
|
#undef nvec
|
||||||
|
|
||||||
#define nvec 6
|
#define nvec 6
|
||||||
#include "sharp_inchelper1.inc.c"
|
#include "sharp_core_inchelper.c"
|
||||||
#undef nvec
|
#undef nvec
|
||||||
|
|
||||||
void inner_loop (sharp_job *job, const int *ispair,const double *cth,
|
void inner_loop (sharp_job *job, const int *ispair,const double *cth,
|
||||||
|
|
|
@ -70,7 +70,7 @@ static inline Tb Y(Tbprod)(Tb a, Tb b)
|
||||||
static inline void Y(Tbmuleq)(Tb * restrict a, Tb b)
|
static inline void Y(Tbmuleq)(Tb * restrict a, Tb b)
|
||||||
{ for (int i=0; i<nvec; ++i) vmuleq(a->v[i],b.v[i]); }
|
{ for (int i=0; i<nvec; ++i) vmuleq(a->v[i],b.v[i]); }
|
||||||
|
|
||||||
static inline void Y(Tbnormalize) (Tb * restrict val, Tb * restrict scale,
|
static void Y(Tbnormalize) (Tb * restrict val, Tb * restrict scale,
|
||||||
double maxval)
|
double maxval)
|
||||||
{
|
{
|
||||||
const Tv vfsmall=vload(sharp_fsmall), vfbig=vload(sharp_fbig);
|
const Tv vfsmall=vload(sharp_fsmall), vfbig=vload(sharp_fbig);
|
||||||
|
@ -94,7 +94,7 @@ static inline void Y(Tbnormalize) (Tb * restrict val, Tb * restrict scale,
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
static inline void Y(mypow) (Tb val, int npow, Tb * restrict resd,
|
static void Y(mypow) (Tb val, int npow, Tb * restrict resd,
|
||||||
Tb * restrict ress)
|
Tb * restrict ress)
|
||||||
{
|
{
|
||||||
Tb scale=Y(Tbconst)(0.), scaleint=Y(Tbconst)(0.), res=Y(Tbconst)(1.);
|
Tb scale=Y(Tbconst)(0.), scaleint=Y(Tbconst)(0.), res=Y(Tbconst)(1.);
|
||||||
|
@ -168,7 +168,7 @@ static inline int Y(TballGe)(Tb a,double b)
|
||||||
return vallTrue(res);
|
return vallTrue(res);
|
||||||
}
|
}
|
||||||
|
|
||||||
static inline void Y(getCorfac)(Tb scale, Tb * restrict corfac,
|
static void Y(getCorfac)(Tb scale, Tb * restrict corfac,
|
||||||
const double * restrict cf)
|
const double * restrict cf)
|
||||||
{
|
{
|
||||||
Y(Tbu) sc, corf;
|
Y(Tbu) sc, corf;
|
||||||
|
|
|
@ -22,7 +22,7 @@
|
||||||
* (DLR).
|
* (DLR).
|
||||||
*/
|
*/
|
||||||
|
|
||||||
/*! \file sharp_core_inc4.c
|
/*! \file sharp_core_inc2.c
|
||||||
* Type-dependent code for the computational core
|
* Type-dependent code for the computational core
|
||||||
*
|
*
|
||||||
* Copyright (C) 2012 Max-Planck-Society
|
* Copyright (C) 2012 Max-Planck-Society
|
|
@ -3,10 +3,10 @@
|
||||||
#include "sharp_core_inc.c"
|
#include "sharp_core_inc.c"
|
||||||
|
|
||||||
#if (SHARP_MAXTRANS>MAXJOB_SPECIAL)
|
#if (SHARP_MAXTRANS>MAXJOB_SPECIAL)
|
||||||
#define NJ1 ,int njobs
|
#define NJ1 , int njobs
|
||||||
#define NJ2 , njobs
|
#define NJ2 , njobs
|
||||||
#define Z(arg) CONCAT2(arg,nvec)
|
#define Z(arg) CONCAT2(arg,nvec)
|
||||||
#include "sharp_core_inc4.c"
|
#include "sharp_core_inc2.c"
|
||||||
#undef Z
|
#undef Z
|
||||||
#undef NJ1
|
#undef NJ1
|
||||||
#undef NJ2
|
#undef NJ2
|
||||||
|
@ -18,7 +18,7 @@
|
||||||
#if ((MAXJOB_SPECIAL>=1)&&(SHARP_MAXTRANS>=1))
|
#if ((MAXJOB_SPECIAL>=1)&&(SHARP_MAXTRANS>=1))
|
||||||
#define njobs 1
|
#define njobs 1
|
||||||
#define Z(arg) CONCAT3(arg,nvec,njobs)
|
#define Z(arg) CONCAT3(arg,nvec,njobs)
|
||||||
#include "sharp_core_inc4.c"
|
#include "sharp_core_inc2.c"
|
||||||
#undef Z
|
#undef Z
|
||||||
#undef njobs
|
#undef njobs
|
||||||
#endif
|
#endif
|
||||||
|
@ -26,7 +26,7 @@
|
||||||
#if ((MAXJOB_SPECIAL>=2)&&(SHARP_MAXTRANS>=2))
|
#if ((MAXJOB_SPECIAL>=2)&&(SHARP_MAXTRANS>=2))
|
||||||
#define njobs 2
|
#define njobs 2
|
||||||
#define Z(arg) CONCAT3(arg,nvec,njobs)
|
#define Z(arg) CONCAT3(arg,nvec,njobs)
|
||||||
#include "sharp_core_inc4.c"
|
#include "sharp_core_inc2.c"
|
||||||
#undef Z
|
#undef Z
|
||||||
#undef njobs
|
#undef njobs
|
||||||
#endif
|
#endif
|
||||||
|
@ -34,7 +34,7 @@
|
||||||
#if ((MAXJOB_SPECIAL>=3)&&(SHARP_MAXTRANS>=3))
|
#if ((MAXJOB_SPECIAL>=3)&&(SHARP_MAXTRANS>=3))
|
||||||
#define njobs 3
|
#define njobs 3
|
||||||
#define Z(arg) CONCAT3(arg,nvec,njobs)
|
#define Z(arg) CONCAT3(arg,nvec,njobs)
|
||||||
#include "sharp_core_inc4.c"
|
#include "sharp_core_inc2.c"
|
||||||
#undef Z
|
#undef Z
|
||||||
#undef njobs
|
#undef njobs
|
||||||
#endif
|
#endif
|
||||||
|
@ -42,7 +42,7 @@
|
||||||
#if ((MAXJOB_SPECIAL>=4)&&(SHARP_MAXTRANS>=4))
|
#if ((MAXJOB_SPECIAL>=4)&&(SHARP_MAXTRANS>=4))
|
||||||
#define njobs 4
|
#define njobs 4
|
||||||
#define Z(arg) CONCAT3(arg,nvec,njobs)
|
#define Z(arg) CONCAT3(arg,nvec,njobs)
|
||||||
#include "sharp_core_inc4.c"
|
#include "sharp_core_inc2.c"
|
||||||
#undef Z
|
#undef Z
|
||||||
#undef njobs
|
#undef njobs
|
||||||
#endif
|
#endif
|
||||||
|
@ -50,7 +50,7 @@
|
||||||
#if ((MAXJOB_SPECIAL>=5)&&(SHARP_MAXTRANS>=5))
|
#if ((MAXJOB_SPECIAL>=5)&&(SHARP_MAXTRANS>=5))
|
||||||
#define njobs 5
|
#define njobs 5
|
||||||
#define Z(arg) CONCAT3(arg,nvec,njobs)
|
#define Z(arg) CONCAT3(arg,nvec,njobs)
|
||||||
#include "sharp_core_inc4.c"
|
#include "sharp_core_inc2.c"
|
||||||
#undef Z
|
#undef Z
|
||||||
#undef njobs
|
#undef njobs
|
||||||
#endif
|
#endif
|
||||||
|
@ -58,7 +58,7 @@
|
||||||
#if ((MAXJOB_SPECIAL>=6)&&(SHARP_MAXTRANS>=6))
|
#if ((MAXJOB_SPECIAL>=6)&&(SHARP_MAXTRANS>=6))
|
||||||
#define njobs 6
|
#define njobs 6
|
||||||
#define Z(arg) CONCAT3(arg,nvec,njobs)
|
#define Z(arg) CONCAT3(arg,nvec,njobs)
|
||||||
#include "sharp_core_inc4.c"
|
#include "sharp_core_inc2.c"
|
||||||
#undef Z
|
#undef Z
|
||||||
#undef njobs
|
#undef njobs
|
||||||
#endif
|
#endif
|
|
@ -79,7 +79,8 @@ void sharp_make_gauss_geom_info (int nrings, int nphi, double phi0,
|
||||||
void sharp_make_fejer1_geom_info (int nrings, int nphi, double phi0,
|
void sharp_make_fejer1_geom_info (int nrings, int nphi, double phi0,
|
||||||
int stride_lon, int stride_lat, sharp_geom_info **geom_info);
|
int stride_lon, int stride_lat, sharp_geom_info **geom_info);
|
||||||
|
|
||||||
/*! Old name for sharp_make_fejer1_geom_info() */
|
/*! Old name for sharp_make_fejer1_geom_info()
|
||||||
|
\ingroup geominfogroup */
|
||||||
static inline void sharp_make_ecp_geom_info (int nrings, int nphi, double phi0,
|
static inline void sharp_make_ecp_geom_info (int nrings, int nphi, double phi0,
|
||||||
int stride_lon, int stride_lat, sharp_geom_info **geom_info)
|
int stride_lon, int stride_lat, sharp_geom_info **geom_info)
|
||||||
{
|
{
|
||||||
|
|
|
@ -154,11 +154,12 @@ static void get_errors (dcmplx **alm, ptrdiff_t nalms, int ncomp, double *sqsum,
|
||||||
double sum=0, maxdiff=0, sumtot, sqsumtot, maxdifftot;
|
double sum=0, maxdiff=0, sumtot, sqsumtot, maxdifftot;
|
||||||
for (ptrdiff_t j=0; j<nalms; ++j)
|
for (ptrdiff_t j=0; j<nalms; ++j)
|
||||||
{
|
{
|
||||||
sum+=creal(alm[i][j])*creal(alm[i][j])
|
double sqr=creal(alm[i][j])*creal(alm[i][j])
|
||||||
+cimag(alm[i][j])*cimag(alm[i][j]);
|
+cimag(alm[i][j])*cimag(alm[i][j]);
|
||||||
if (fabs(creal(alm[i][j]))>maxdiff) maxdiff=fabs(creal(alm[i][j]));
|
sum+=sqr;
|
||||||
if (fabs(cimag(alm[i][j]))>maxdiff) maxdiff=fabs(cimag(alm[i][j]));
|
if (sqr>maxdiff) maxdiff=sqr;
|
||||||
}
|
}
|
||||||
|
maxdiff=sqrt(maxdiff);
|
||||||
|
|
||||||
#ifdef USE_MPI
|
#ifdef USE_MPI
|
||||||
MPI_Allreduce(&sum,&sumtot,1,MPI_DOUBLE,MPI_SUM,MPI_COMM_WORLD);
|
MPI_Allreduce(&sum,&sumtot,1,MPI_DOUBLE,MPI_SUM,MPI_COMM_WORLD);
|
||||||
|
|
Loading…
Add table
Add a link
Reference in a new issue