-
Notifications
You must be signed in to change notification settings - Fork 7
/
fieldize.cpp
133 lines (127 loc) · 5.56 KB
/
fieldize.cpp
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
/* Copyright (c) 2009, Simeon Bird <[email protected]>
*
* Permission to use, copy, modify, and/or distribute this software for any
* purpose with or without fee is hereby granted, provided that the above
* copyright notice and this permission notice appear in all copies.
*
* THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
* WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
* MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
* ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
* WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
* ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
* OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. */
/* Fieldize. positions should be an array of size 3*segment_particles
* (like the output of read_gadget_float3)
* out is an array of size [dims*dims*dims]*/
#include <cmath>
#include <stdint.h>
#include "gen-pk.h"
#include <cassert>
/** Number of particles to keep in the thread-local buffer*/
#define IL 64
#ifndef M_PI
/** C99 does not define PI!*/
#define M_PI 3.1415926535897932384626433832795
#endif
/** \file
* Defines fieldize()
* Fieldize: this does cloud-in-cell interpolation, assuming periodic boundary conditions.
* @param boxsize The physical size of the grid. This is only used so to convert the returned data to physical units
* @param dims The number of grid points to use
* @param out Pointer to the grid to interpolate onto.
* Should be an array of size dims^3 (or slightly larger if extra is set, see below.
* @param segment_particles The number of particles to interpolate on this call. The length of the positions array.
* @param positions Array of particle positions. Assumed to have dimensions 3*segment_particles
* @param masses Array containing particle masses, if variable. Null if particle masses are constant.
* @param mass Particle mass if constant for all of this type.
* @param extra This switch should be 0 or 1. If set to one, will assume that the output
* is about to be handed to an FFTW in-place routine,
* and skip the last 2 places of each row in the last dimension (of out)
*/
int fieldize(double boxsize, int dims, GENFLOAT *out, int64_t segment_particles, GENPK_FLOAT_TYPE *positions, GENPK_FLOAT_TYPE * masses, double mass, int extra)
{
const size_t fdims=2*(dims/2+extra);
/*If extra is on, we want to leave space for FFTW
* to put the extra bits, so skip a couple of places.*/
const size_t dims2=fdims*dims;
const GENFLOAT units=dims/boxsize;
#pragma omp parallel for
for(int64_t index=0;index<segment_particles;index+=IL) {
GENFLOAT temp[IL][8];
size_t temp2[IL][8];
const int64_t il=(index+IL<segment_particles ? IL : segment_particles-index);
for(int64_t k=0; k<il; k++)
{
GENFLOAT dx[3],tx[3], x[3];
int fx[3],nex[3];
/* This is one over density.*/
const GENFLOAT invrho = (masses ? masses[index+k] : mass);
for(int i=0; i<3; i++)
{
x[i]=positions[3*(index+k)+i]*units;
fx[i]=floor(x[i]);
dx[i]=x[i]-fx[i];
tx[i]=1.0-dx[i];
nex[i]=(fx[i]+1)%dims;
if(nex[i]<0)
nex[i]+=dims;
fx[i]%=dims;
if(fx[i]<0)
fx[i]+=dims;
}
temp[k][0]=invrho*tx[0]*tx[1]*tx[2];
temp[k][1]=invrho*dx[0]*tx[1]*tx[2];
temp[k][2]=invrho*tx[0]*dx[1]*tx[2];
temp[k][3]=invrho*dx[0]*dx[1]*tx[2];
temp[k][4]=invrho*tx[0]*tx[1]*dx[2];
temp[k][5]=invrho*dx[0]*tx[1]*dx[2];
temp[k][6]=invrho*tx[0]*dx[1]*dx[2];
temp[k][7]=invrho*dx[0]*dx[1]*dx[2];
temp2[k][0]=dims2*fx[0] +fdims*fx[1] + fx[2];
temp2[k][1]=dims2*nex[0]+fdims*fx[1] + fx[2];
temp2[k][2]=dims2*fx[0] +fdims*nex[1]+ fx[2];
temp2[k][3]=dims2*nex[0]+fdims*nex[1]+ fx[2];
temp2[k][4]=dims2*fx[0] +fdims*fx[1] +nex[2];
temp2[k][5]=dims2*nex[0]+fdims*fx[1] +nex[2];
temp2[k][6]=dims2*fx[0] +fdims*nex[1]+nex[2];
temp2[k][7]=dims2*nex[0]+fdims*nex[1]+nex[2];
for(int c=0; c<8; c++)
assert(temp2[k][c] >= 0 && temp2[k][c] < dims2*dims);
}
/*The store operation may only be done by one thread at a time,
*to ensure synchronisation.*/
#pragma omp critical
{
for(int64_t k=0; k<il; k++)
{
out[temp2[k][0]]+=temp[k][0];
out[temp2[k][1]]+=temp[k][1];
out[temp2[k][2]]+=temp[k][2];
out[temp2[k][3]]+=temp[k][3];
out[temp2[k][4]]+=temp[k][4];
out[temp2[k][5]]+=temp[k][5];
out[temp2[k][6]]+=temp[k][6];
out[temp2[k][7]]+=temp[k][7];
}
}
}
return 0;
}
//Helper function for 1D window function.
inline GENFLOAT onedinvwindow(int64_t kx, int64_t n)
{
//Return \pi x /(n sin(\pi x n)) unless x = 0, in which case return 1.
return kx ? M_PI*kx/(n*sin(M_PI*kx/(float)n)) : 1.0;
}
//The window function of the CiC procedure above. Need to deconvolve this for the power spectrum.
//Only has an effect for k > Nyquist/4.
GENFLOAT invwindow(int64_t kx, int64_t ky, int64_t kz, int64_t n)
{
if(n == 0)
return 0;
float iwx = onedinvwindow(kx, n);
float iwy = onedinvwindow(ky, n);
float iwz = onedinvwindow(kz, n);
return pow(iwx*iwy*iwz,2);
}