Aquila-Consortium/libsharp2
0
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

cleanup

Browse source
This commit is contained in:
Martin Reinecke 2019-01-10 10:30:07 +01:00
parent 48cafe5c1c
commit ecd6c1b48b
5 changed files with 14 additions and 113 deletions
Download patch file Download diff file Expand all files Collapse all files

1
Makefile.am
View file

@ -19,7 +19,6 @@ src_sharp = \
libsharp/sharp_internal.h \
libsharp/sharp_legendre_roots.h \
libsharp/sharp_vecsupport.h \
libsharp/sharp_vecutil.h \
libsharp/sharp_ylmgen_c.h
include_HEADERS = \

61
libsharp/sharp_complex_hacks.h
View file

@ -39,13 +39,6 @@
#if (VLEN==1)
static inline _Complex double vhsum_cmplx(Tv a, Tv b)
{ return a+_Complex_I*b; }
static inline void vhsum_cmplx2 (Tv a, Tv b, Tv c, Tv d,
_Complex double * restrict c1, _Complex double * restrict c2)
{ *c1 += a+_Complex_I*b; *c2 += c+_Complex_I*d; }
static inline void vhsum_cmplx_special (Tv a, Tv b, Tv c, Tv d,
_Complex double * restrict cc)
{ cc[0] += a+_Complex_I*b; cc[1] += c+_Complex_I*d; }
@ -54,18 +47,6 @@ static inline void vhsum_cmplx_special (Tv a, Tv b, Tv c, Tv d,
#if (VLEN==2)
static inline _Complex double vhsum_cmplx (Tv a, Tv b)
{
#if defined(__SSE3__)
Tv tmp = _mm_hadd_pd(a,b);
#else
Tv tmp = _mm_shuffle_pd(a,b,_MM_SHUFFLE2(0,1)) +
_mm_shuffle_pd(a,b,_MM_SHUFFLE2(1,0));
#endif
union {Tv v; _Complex double c; } u;
u.v=tmp; return u.c;
}
static inline void vhsum_cmplx2 (Tv a, Tv b, Tv c,
Tv d, _Complex double * restrict c1, _Complex double * restrict c2)
{
@ -101,38 +82,6 @@ static inline void vhsum_cmplx_special (Tv a, Tv b, Tv c, Tv d,
#if (VLEN==4)
static inline _Complex double vhsum_cmplx (Tv a, Tv b)
{
Tv tmp=_mm256_hadd_pd(a,b);
Tv tmp2=_mm256_permute2f128_pd(tmp,tmp,1);
tmp=_mm256_add_pd(tmp,tmp2);
#ifdef UNSAFE_CODE
_Complex double ret;
*((__m128d *)&ret)=_mm256_extractf128_pd(tmp, 0);
return ret;
#else
union {Tv v; _Complex double c[2]; } u;
u.v=tmp; return u.c[0];
#endif
}
static inline void vhsum_cmplx2 (Tv a, Tv b, Tv c, Tv d,
_Complex double * restrict c1, _Complex double * restrict c2)
{
Tv tmp1=_mm256_hadd_pd(a,b), tmp2=_mm256_hadd_pd(c,d);
Tv tmp3=_mm256_permute2f128_pd(tmp1,tmp2,49),
tmp4=_mm256_permute2f128_pd(tmp1,tmp2,32);
tmp1=tmp3+tmp4;
#ifdef UNSAFE_CODE
*((__m128d *)c1)=_mm_add_pd(*((__m128d *)c1),_mm256_extractf128_pd(tmp1, 0));
*((__m128d *)c2)=_mm_add_pd(*((__m128d *)c2),_mm256_extractf128_pd(tmp1, 1));
#else
union {Tv v; _Complex double c[2]; } u;
u.v=tmp1;
*c1+=u.c[0]; *c2+=u.c[1];
#endif
}
static inline void vhsum_cmplx_special (Tv a, Tv b, Tv c, Tv d,
_Complex double * restrict cc)
{
@ -154,16 +103,6 @@ static inline void vhsum_cmplx_special (Tv a, Tv b, Tv c, Tv d,
#if (VLEN==8)
static inline _Complex double vhsum_cmplx(Tv a, Tv b)
{ return _mm512_reduce_add_pd(a)+_Complex_I*_mm512_reduce_add_pd(b); }
static inline void vhsum_cmplx2 (Tv a, Tv b, Tv c, Tv d,
_Complex double * restrict c1, _Complex double * restrict c2)
{
*c1 += _mm512_reduce_add_pd(a)+_Complex_I*_mm512_reduce_add_pd(b);
*c2 += _mm512_reduce_add_pd(c)+_Complex_I*_mm512_reduce_add_pd(d);
}
static inline void vhsum_cmplx_special (Tv a, Tv b, Tv c, Tv d,
_Complex double * restrict cc)
{ vhsum_cmplx2(a,b,c,d,cc,cc+1); }

15
libsharp/sharp_vecsupport.h
View file

@ -33,7 +33,20 @@
#define SHARP_VECSUPPORT_H
#include <math.h>
#include "sharp_vecutil.h"
#ifndef VLEN
#if (defined(__AVX512F__))
#define VLEN 8
#elif (defined (__AVX__))
#define VLEN 4
#elif (defined (__SSE2__))
#define VLEN 2
#else
#define VLEN 1
#endif
#endif
typedef double Ts;

49
libsharp/sharp_vecutil.h
View file

@ -1,49 +0,0 @@
/*
* This file is part of libc_utils.
*
* libc_utils is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* libc_utils is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with libc_utils; if not, write to the Free Software
* Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
*/
/*
* libc_utils is being developed at the Max-Planck-Institut fuer Astrophysik
* and financially supported by the Deutsches Zentrum fuer Luft- und Raumfahrt
* (DLR).
*/
/*! \file sharp_vecutil.h
* Functionality related to vector instruction support
*
* Copyright (C) 2012-2018 Max-Planck-Society
* \author Martin Reinecke
*/
#ifndef SHARP_VECUTIL_H
#define SHARP_VECUTIL_H
#ifndef VLEN
#if (defined(__AVX512F__))
#define VLEN 8
#elif (defined (__AVX__))
#define VLEN 4
#elif (defined (__SSE2__))
#define VLEN 2
#else
#define VLEN 1
#endif
#endif
#endif

1
libsharp/sharp_ylmgen_c.h
View file

@ -42,7 +42,6 @@ static const double sharp_ftol=0x1p-60;
static const double sharp_fbighalf=0x1p+400;
typedef struct { double f[2]; } sharp_ylmgen_dbl2;
typedef struct { double f[3]; } sharp_ylmgen_dbl3;
typedef struct
{