lisaanalysistools 1.0.4__cp312-cp312-macosx_10_9_x86_64.whl → 1.0.6__cp312-cp312-macosx_10_9_x86_64.whl

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.

Potentially problematic release.


This version of lisaanalysistools might be problematic. Click here for more details.

Files changed (30) hide show
  1. {lisaanalysistools-1.0.4.dist-info → lisaanalysistools-1.0.6.dist-info}/METADATA +12 -6
  2. lisaanalysistools-1.0.6.dist-info/RECORD +36 -0
  3. {lisaanalysistools-1.0.4.dist-info → lisaanalysistools-1.0.6.dist-info}/WHEEL +1 -1
  4. {lisaanalysistools-1.0.4.dist-info → lisaanalysistools-1.0.6.dist-info}/top_level.txt +0 -1
  5. lisatools/_version.py +1 -1
  6. lisatools/cutils/__init__.py +0 -0
  7. lisatools/cutils/detector_cpu.cpython-312-darwin.so +0 -0
  8. lisatools/cutils/include/Detector.hpp +78 -0
  9. lisatools/cutils/include/__init__.py +0 -0
  10. lisatools/cutils/include/global.hpp +28 -0
  11. lisatools/cutils/src/Detector.cpp +210 -0
  12. lisatools/cutils/src/Detector.cu +210 -0
  13. lisatools/cutils/src/__init__.py +0 -0
  14. lisatools/cutils/src/pycppdetector.pyx +199 -0
  15. lisatools/datacontainer.py +23 -6
  16. lisatools/detector.py +59 -15
  17. lisatools/sensitivity.py +8 -6
  18. lisatools/stochastic.py +1 -1
  19. lisaanalysistools-1.0.4.dist-info/RECORD +0 -37
  20. lisatools/cutils/detector.cpython-312-darwin.so +0 -0
  21. lisatools/glitch.py +0 -193
  22. lisatools/sampling/moves/gbgroupstretch.py +0 -53
  23. lisatools/sampling/moves/gbmultipletryrj.py +0 -1287
  24. lisatools/sampling/moves/gbspecialgroupstretch.py +0 -671
  25. lisatools/sampling/moves/gbspecialstretch.py +0 -1836
  26. lisatools/sampling/moves/mbhspecialmove.py +0 -286
  27. lisatools/sampling/moves/placeholder.py +0 -16
  28. lisatools/sampling/moves/specialforegroundmove.py +0 -564
  29. lisatools/utils/multigpudataholder.py +0 -730
  30. {lisaanalysistools-1.0.4.dist-info → lisaanalysistools-1.0.6.dist-info}/LICENSE +0 -0
@@ -1,7 +1,7 @@
1
1
  Metadata-Version: 2.1
2
2
  Name: lisaanalysistools
3
- Version: 1.0.4
4
- Home-page: https://github.com/mikekatz04/lisa-on-gpu
3
+ Version: 1.0.6
4
+ Home-page: https://github.com/mikekatz04/LISAanalysistools
5
5
  Author: Michael Katz
6
6
  Author-email: mikekatz04@gmail.com
7
7
  Classifier: Programming Language :: Python :: 3
@@ -36,7 +36,9 @@ pip install lisaanalysistools
36
36
 
37
37
  If you want to install all of the tools associated with LISA Analysis Tools (Fast EMRI Waveforms, BBHx, GBGPU, fastlisaresponse, eryn), see the following instructions.
38
38
 
39
- LISA Analysis Tools leverages conda environments to install and use necessary packages. If you do not have [Anaconda](https://www.anaconda.com/download) or [miniconda](https://docs.anaconda.com/free/miniconda/index.html) installed, you must do this first and load your `base` conda environment.
39
+ LISA Analysis Tools leverages conda environments to install and use necessary packages. If you do not have [Anaconda](https://www.anaconda.com/download) or [miniconda](https://docs.anaconda.com/free/miniconda/index.html) installed, you must do this first and load your `base` conda environment. Recommended components for install in your conda environment are `lapack`, `gsl`, `hdf5`, which are needed for various waveform packages.
40
+
41
+ For an easy full install, follow these instructions.
40
42
 
41
43
  First, clone the repo and `cd` to the `LATW` directory.:
42
44
  ```
@@ -50,7 +52,11 @@ bash install.sh
50
52
  ```
51
53
  Running `bash install.sh -h` will also give you some basic install options.
52
54
 
53
- If you want more flexibility, you can install each package given above separately. If you do this, you will also need # TODO: add.
55
+ If you want more flexibility, you can install each package given above separately.
56
+
57
+ To install this software for use with NVIDIA GPUs (compute capability >5.0), you need the [CUDA toolkit](https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html) and [CuPy](https://cupy.chainer.org/). The CUDA toolkit must have cuda version >8.0. Be sure to properly install CuPy within the correct CUDA toolkit version. Make sure the nvcc binary is on `$PATH` or set it as the `CUDA_HOME` environment variable.
58
+
59
+ We are currently working on building wheels and making the GPU version pip installable. For now, to work with GPUs, git clone the repository and install it from source.
54
60
 
55
61
  ## Contributing
56
62
 
@@ -58,9 +64,9 @@ Please read [CONTRIBUTING.md](CONTRIBUTING.md) for details on our code of conduc
58
64
 
59
65
  ## Versioning
60
66
 
61
- We use [SemVer](http://semver.org/) for versioning. For the versions available, see the [tags on this repository](https://github.com/BlackHolePerturbationToolkit/FastEMRIWaveforms/tags).
67
+ We use [SemVer](http://semver.org/) for versioning. For the versions available, see the [tags on this repository](https://github.com/mikekatz04/LISAanalysistools/tags).
62
68
 
63
- Current Version: 1.0.4
69
+ Current Version: 1.0.6
64
70
 
65
71
  ## Authors/Developers
66
72
 
@@ -0,0 +1,36 @@
1
+ lisatools/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
2
+ lisatools/_version.py,sha256=14imVuUDlv1B9HhbSkkcqdP6lnMYhzyYQVjZiFvzp6k,123
3
+ lisatools/analysiscontainer.py,sha256=ePwTBUTEBJn2TK93_afARate9SAqUKK8c8T6DcGUx1Y,15321
4
+ lisatools/datacontainer.py,sha256=m4gp8E2qOIFq1JPEid3jA450hdNo0yWQu770AVJdrDI,9809
5
+ lisatools/detector.py,sha256=4XYWi8bSdM_97IL1L0EKKInO1fjRM4-FuW3gE8xeEEM,15262
6
+ lisatools/diagnostic.py,sha256=CfPpfvDys1XyZRWmmTqTSWb0SY2eH0G_8TRnt1OxBFo,34174
7
+ lisatools/sensitivity.py,sha256=rckXeJ_5U3cRRgbYDmt0kaTyJRjkrYfgBwos12QJJfw,27297
8
+ lisatools/stochastic.py,sha256=XL6ZbqprfcM1rsDw3PhrqdflpQiCnZ5x4y_k4QZ9YGg,9423
9
+ lisatools/cutils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
10
+ lisatools/cutils/detector_cpu.cpython-312-darwin.so,sha256=QQFxyuqx9L7HciApuIiw2tLkh-7Qqs4PRkiOWkqcClM,126912
11
+ lisatools/cutils/include/Detector.hpp,sha256=dzdo8QUrrcAnfrTyytlZkQrYmVk_QWXXTL68p4COpZ8,1926
12
+ lisatools/cutils/include/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
13
+ lisatools/cutils/include/global.hpp,sha256=3VPiqglTMRrIBXlEvDUJO2-CjKy_SLUXZt-9A1fH6lQ,572
14
+ lisatools/cutils/src/Detector.cpp,sha256=RbPYrK70m0heUZyFha-2_FJOW8Fg9vKfZQzNIXvOqUQ,5188
15
+ lisatools/cutils/src/Detector.cu,sha256=RbPYrK70m0heUZyFha-2_FJOW8Fg9vKfZQzNIXvOqUQ,5188
16
+ lisatools/cutils/src/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
17
+ lisatools/cutils/src/pycppdetector.pyx,sha256=N0AyF5YPPqmHaBKuj4gZVXOt67z9qKwzpB6jlh9zWRY,5581
18
+ lisatools/sampling/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
19
+ lisatools/sampling/likelihood.py,sha256=G2kAQ43qlhAzIFWvYsrSmHXd7WKJAzcCN2o07vRE8vc,29585
20
+ lisatools/sampling/prior.py,sha256=1K1PMStpwO9WT0qG0aotKSyoNjuehXNbzTDtlk8Q15M,21407
21
+ lisatools/sampling/stopping.py,sha256=Q8q7nM0wnJervhRduf2tBXZZHUVza5kJiAUAMUQXP5o,9682
22
+ lisatools/sampling/utility.py,sha256=rOGotS0Aj8-DAWqsTVy2xWNsxsoz74BVrHEnG2mOkwU,14340
23
+ lisatools/sampling/moves/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
24
+ lisatools/sampling/moves/skymodehop.py,sha256=0nf72eFhFMGwi0dLJci6XZz-bIMGqco2B2_J72hQvf8,3348
25
+ lisatools/sources/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
26
+ lisatools/sources/emri/__init__.py,sha256=HjsWS4mjog_gRE4N90My6ahLKQGjNwrHIeA40R7qzw0,41
27
+ lisatools/sources/emri/tdiwaveform.py,sha256=jaKL2PO8RlJrHP45zCa4BpRniLfdnMIP1QsIB-6jVq4,2297
28
+ lisatools/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
29
+ lisatools/utils/constants.py,sha256=r1kVwkpbZS13JTOxj2iRxT5sMgTYX30y-S0JdVmD5Oo,1354
30
+ lisatools/utils/pointeradjust.py,sha256=2sT-7qeYWr1pd_sHk9leVHUTSJ7jJgYIRoWQOtYqguE,2995
31
+ lisatools/utils/utility.py,sha256=3mJoJKNGrm3KuNXIa2RUKi9WKd593V4q9XjjQZCQD0M,6831
32
+ lisaanalysistools-1.0.6.dist-info/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
33
+ lisaanalysistools-1.0.6.dist-info/METADATA,sha256=1T8XjHyoWcJMiuLG2DpQXqh181FMwc_K2adouVyiuO0,4109
34
+ lisaanalysistools-1.0.6.dist-info/WHEEL,sha256=wTFHiv6xiwExbw_CBGcw1b3hAKzFq_QaWIv4Plh5P0w,110
35
+ lisaanalysistools-1.0.6.dist-info/top_level.txt,sha256=qXUn8Xi8yvK0vr3QH0vvT5cmoccjSj-UPeKGLAxdYGo,10
36
+ lisaanalysistools-1.0.6.dist-info/RECORD,,
@@ -1,5 +1,5 @@
1
1
  Wheel-Version: 1.0
2
- Generator: bdist_wheel (0.43.0)
2
+ Generator: setuptools (72.1.0)
3
3
  Root-Is-Purelib: false
4
4
  Tag: cp312-cp312-macosx_10_9_x86_64
5
5
 
lisatools/_version.py CHANGED
@@ -1,4 +1,4 @@
1
- __version__ = '1.0.4'
1
+ __version__ = '1.0.6'
2
2
  __copyright__ = "Michael L. Katz 2024"
3
3
  __name__ = "lisaanalysistools"
4
4
  __author__ = "Michael L. Katz"
File without changes
@@ -0,0 +1,78 @@
1
+ #ifndef __DETECTOR_HPP__
2
+ #define __DETECTOR_HPP__
3
+
4
+ #include "global.hpp"
5
+ #include <iostream>
6
+
7
+ class Vec
8
+ {
9
+ public:
10
+ double x;
11
+ double y;
12
+ double z;
13
+
14
+ CUDA_DEVICE
15
+ Vec(double x_, double y_, double z_)
16
+ {
17
+ x = x_;
18
+ y = y_;
19
+ z = z_;
20
+ }
21
+ };
22
+
23
+ class Orbits
24
+ {
25
+ public:
26
+ double dt;
27
+ int N;
28
+ double *n_arr;
29
+ double *ltt_arr;
30
+ double *x_arr;
31
+ int nlinks;
32
+ int nspacecraft;
33
+ int *links;
34
+ int *sc_r;
35
+ int *sc_e;
36
+
37
+ Orbits(double dt_, int N_, double *n_arr_, double *ltt_arr_, double *x_arr_, int *links_, int *sc_r_, int *sc_e_)
38
+ {
39
+ dt = dt_;
40
+ N = N_;
41
+ n_arr = n_arr_;
42
+ ltt_arr = ltt_arr_;
43
+ x_arr = x_arr_;
44
+ nlinks = 6;
45
+ nspacecraft = 3;
46
+
47
+ sc_r = sc_r_;
48
+ sc_e = sc_e_;
49
+ links = links_;
50
+
51
+ // std::cout << " START " << std::endl;
52
+ // for (int i = 0; i < nlinks; i += 1)
53
+ // {
54
+ // sc_r[i] = sc_r_[i];
55
+ // sc_e[i] = sc_e_[i];
56
+ // links[i] = links_[i];
57
+ // // std::cout << i << " HAHAHAH " << sc_r_[i] << " " << sc_e_[i] << std::endl;
58
+ // }
59
+ };
60
+
61
+ CUDA_DEVICE int get_window(double t);
62
+ CUDA_DEVICE Vec get_normal_unit_vec(double t, int link);
63
+ CUDA_DEVICE double interpolate(double t, double *in_arr, int window, int major_ndim, int major_ind, int ndim, int pos);
64
+ CUDA_DEVICE int get_link_ind(int link);
65
+ CUDA_DEVICE int get_sc_ind(int sc);
66
+ CUDA_DEVICE double get_light_travel_time(double t, int link);
67
+ CUDA_DEVICE Vec get_pos(double t, int sc);
68
+ CUDA_DEVICE void get_normal_unit_vec_ptr(Vec *vec, double t, int link);
69
+ CUDA_DEVICE void get_pos_ptr(Vec *vec, double t, int sc);
70
+ void get_light_travel_time_arr(double *ltt, double *t, int *link, int num);
71
+ void dealloc(){
72
+ // delete[] links;
73
+ // delete[] sc_r;
74
+ // delete[] sc_e;
75
+ };
76
+ };
77
+
78
+ #endif // __DETECTOR_HPP__
File without changes
@@ -0,0 +1,28 @@
1
+ #ifndef __GLOBAL_HPP__
2
+ #define __GLOBAL_HPP__
3
+
4
+ #ifdef __CUDACC__
5
+ #define CUDA_KERNEL __global__
6
+ #define CUDA_DEVICE __device__
7
+
8
+ #else // __CUDACC__
9
+ #define CUDA_KERNEL
10
+ #define CUDA_DEVICE
11
+
12
+ #endif // __CUDACC__
13
+
14
+ #ifdef __CUDACC__
15
+ #define gpuErrchk(ans) { gpuAssert((ans), __FILE__, __LINE__); }
16
+ inline void gpuAssert(cudaError_t code, const char *file, int line, bool abort=true)
17
+ {
18
+ if (code != cudaSuccess)
19
+ {
20
+ fprintf(stderr,"GPUassert: %s %s %d\n", cudaGetErrorString(code), file, line);
21
+ if (abort) exit(code);
22
+ }
23
+ }
24
+
25
+ #endif
26
+
27
+
28
+ #endif // __GLOBAL_HPP__
@@ -0,0 +1,210 @@
1
+ #include "stdio.h"
2
+ #include "global.hpp"
3
+ #include "Detector.hpp"
4
+ #include <iostream>
5
+ #include <stdexcept>
6
+ #include <string>
7
+ #include <sstream>
8
+
9
+ CUDA_DEVICE
10
+ int Orbits::get_window(double t)
11
+ {
12
+ int out = int(t / dt);
13
+ if ((out < 0) || (out >= N))
14
+ return -1;
15
+ else
16
+ return out;
17
+ }
18
+
19
+ CUDA_DEVICE
20
+ int Orbits::get_link_ind(int link)
21
+ {
22
+ if (link == 12)
23
+ return 0;
24
+ else if (link == 23)
25
+ return 1;
26
+ else if (link == 31)
27
+ return 2;
28
+ else if (link == 13)
29
+ return 3;
30
+ else if (link == 32)
31
+ return 4;
32
+ else if (link == 21)
33
+ return 5;
34
+ else
35
+ #ifdef __CUDACC__
36
+ printf("BAD link ind. Must be 12, 23, 31, 13, 32, 21.");
37
+ #else
38
+ throw std::invalid_argument("Bad link ind. Must be 12, 23, 31, 13, 32, 21.");
39
+ #endif // __CUDACC__
40
+ return -1;
41
+ }
42
+
43
+ CUDA_DEVICE
44
+ int Orbits::get_sc_ind(int sc)
45
+ {
46
+ if (sc == 1)
47
+ return 0;
48
+ else if (sc == 2)
49
+ return 1;
50
+ else if (sc == 3)
51
+ return 2;
52
+ else
53
+ {
54
+ #ifdef __CUDACC__
55
+ printf("BAD sc ind. Must be 1,2,3. %d\n", sc);
56
+ #else
57
+ std::ostringstream oss;
58
+ int xxxx = 1;
59
+ oss << "Bad sc ind. Must be 1,2,3. Input sc is " << sc << " " << xxxx;
60
+ std::string var = oss.str();
61
+ throw std::invalid_argument(var);
62
+ #endif // __CUDACC__
63
+ }
64
+ return 0;
65
+ }
66
+
67
+ CUDA_DEVICE
68
+ double Orbits::interpolate(double t, double *in_arr, int window, int major_ndim, int major_ind, int ndim, int pos)
69
+ {
70
+ double up = in_arr[((window + 1) * major_ndim + major_ind) * ndim + pos]; // down_ind * ndim + pos];
71
+ double down = in_arr[(window * major_ndim + major_ind) * ndim + pos];
72
+
73
+ // m *(x - x0) + y0
74
+ double fin = ((up - down) / dt) * (t - (dt * window)) + down;
75
+ // if ((ndim == 1))
76
+ // printf("%d %e %e %e %e \n", window, fin, down, up, (t - (dt * window)));
77
+
78
+ return fin;
79
+ }
80
+
81
+ CUDA_DEVICE
82
+ void Orbits::get_normal_unit_vec_ptr(Vec *vec, double t, int link)
83
+ {
84
+ Vec _tmp = get_normal_unit_vec(t, link);
85
+ vec->x = _tmp.x;
86
+ vec->y = _tmp.y;
87
+ vec->z = _tmp.z;
88
+ }
89
+
90
+ CUDA_DEVICE
91
+ Vec Orbits::get_normal_unit_vec(double t, int link)
92
+ {
93
+ int window = get_window(t);
94
+ if (window == -1)
95
+ {
96
+ // out of bounds
97
+ return Vec(0.0, 0.0, 0.0);
98
+ }
99
+
100
+ int link_ind = get_link_ind(link);
101
+
102
+ int up_ind = (window + 1) * nlinks + link_ind;
103
+ int down_ind = window * nlinks + link_ind;
104
+
105
+ // x (pos = 0) ndim = 3
106
+ double x_out = interpolate(t, n_arr, window, nlinks, link_ind, 3, 0);
107
+ // y (pos = 1)
108
+ double y_out = interpolate(t, n_arr, window, nlinks, link_ind, 3, 1);
109
+ // z (pos = 2)
110
+ double z_out = interpolate(t, n_arr, window, nlinks, link_ind, 3, 2);
111
+
112
+ return Vec(x_out, y_out, z_out);
113
+ }
114
+
115
+ CUDA_DEVICE
116
+ double Orbits::get_light_travel_time(double t, int link)
117
+ {
118
+ int window = get_window(t);
119
+ if (window == -1)
120
+ {
121
+ // out of bounds
122
+ return 0.0;
123
+ }
124
+
125
+ int link_ind = get_link_ind(link);
126
+ if ((link_ind < 0) || (link_ind >= 6))
127
+ printf("BAD %d\n", link_ind);
128
+ int up_ind = (window + 1) * (nlinks + link_ind);
129
+ int down_ind = window * (nlinks + link_ind);
130
+
131
+ // x (pos = 0), ndim = 1
132
+ double ltt_out = interpolate(t, ltt_arr, window, nlinks, link_ind, 1, 0);
133
+
134
+ return ltt_out;
135
+ }
136
+
137
+ CUDA_DEVICE
138
+ Vec Orbits::get_pos(double t, int sc)
139
+ {
140
+ int window = get_window(t);
141
+ if (window == -1)
142
+ {
143
+ // out of bounds
144
+ return Vec(0.0, 0.0, 0.0);
145
+ }
146
+
147
+ int sc_ind = get_sc_ind(sc);
148
+
149
+ // x (pos = 0), ndim = 3
150
+ double x_out = interpolate(t, x_arr, window, nspacecraft, sc_ind, 3, 0);
151
+ // y (pos = 1), ndim = 3
152
+ double y_out = interpolate(t, x_arr, window, nspacecraft, sc_ind, 3, 1);
153
+ // z (pos = 2), ndim = 3
154
+ double z_out = interpolate(t, x_arr, window, nspacecraft, sc_ind, 3, 2);
155
+ return Vec(x_out, y_out, z_out);
156
+ }
157
+
158
+ CUDA_DEVICE
159
+ void Orbits::get_pos_ptr(Vec *vec, double t, int sc)
160
+ {
161
+ Vec _tmp = get_pos(t, sc);
162
+ vec->x = _tmp.x;
163
+ vec->y = _tmp.y;
164
+ vec->z = _tmp.z;
165
+ }
166
+
167
+ #define NUM_THREADS 64
168
+
169
+ CUDA_KERNEL
170
+ void get_light_travel_time_kernel(double *ltt, double *t, int *link, int num, Orbits& orbits)
171
+ {
172
+ int start, end, increment;
173
+ #ifdef __CUDACC__
174
+ start = blockIdx.x * blockDim.x + threadIdx.x;
175
+ end = num;
176
+ increment = gridDim.x * blockDim.x;
177
+ #else // __CUDACC__
178
+ start = 0;
179
+ end = num;
180
+ increment = 1;
181
+ #endif // __CUDACC__
182
+
183
+ for (int i = start; i < end; i += increment)
184
+ {
185
+ ltt[i] = orbits.get_light_travel_time(t[i], link[i]);
186
+ }
187
+ }
188
+
189
+ void Orbits::get_light_travel_time_arr(double *ltt, double *t, int *link, int num)
190
+ {
191
+ #ifdef __CUDACC__
192
+ int num_blocks = std::ceil((num + NUM_THREADS - 1) / NUM_THREADS);
193
+
194
+ // copy self to GPU
195
+ Orbits *orbits_gpu;
196
+ gpuErrchk(cudaMalloc(&orbits_gpu, sizeof(Orbits)));
197
+ gpuErrchk(cudaMemcpy(orbits_gpu, this, sizeof(Orbits), cudaMemcpyHostToDevice));
198
+
199
+ get_light_travel_time_kernel<<<num_blocks, NUM_THREADS>>>(ltt, t, link, num, *orbits_gpu);
200
+ cudaDeviceSynchronize();
201
+ gpuErrchk(cudaGetLastError());
202
+
203
+ gpuErrchk(cudaFree(orbits_gpu));
204
+
205
+ #else // __CUDACC__
206
+
207
+ get_light_travel_time_kernel(ltt, t, link, num, *this);
208
+
209
+ #endif // __CUDACC__
210
+ }
@@ -0,0 +1,210 @@
1
+ #include "stdio.h"
2
+ #include "global.hpp"
3
+ #include "Detector.hpp"
4
+ #include <iostream>
5
+ #include <stdexcept>
6
+ #include <string>
7
+ #include <sstream>
8
+
9
+ CUDA_DEVICE
10
+ int Orbits::get_window(double t)
11
+ {
12
+ int out = int(t / dt);
13
+ if ((out < 0) || (out >= N))
14
+ return -1;
15
+ else
16
+ return out;
17
+ }
18
+
19
+ CUDA_DEVICE
20
+ int Orbits::get_link_ind(int link)
21
+ {
22
+ if (link == 12)
23
+ return 0;
24
+ else if (link == 23)
25
+ return 1;
26
+ else if (link == 31)
27
+ return 2;
28
+ else if (link == 13)
29
+ return 3;
30
+ else if (link == 32)
31
+ return 4;
32
+ else if (link == 21)
33
+ return 5;
34
+ else
35
+ #ifdef __CUDACC__
36
+ printf("BAD link ind. Must be 12, 23, 31, 13, 32, 21.");
37
+ #else
38
+ throw std::invalid_argument("Bad link ind. Must be 12, 23, 31, 13, 32, 21.");
39
+ #endif // __CUDACC__
40
+ return -1;
41
+ }
42
+
43
+ CUDA_DEVICE
44
+ int Orbits::get_sc_ind(int sc)
45
+ {
46
+ if (sc == 1)
47
+ return 0;
48
+ else if (sc == 2)
49
+ return 1;
50
+ else if (sc == 3)
51
+ return 2;
52
+ else
53
+ {
54
+ #ifdef __CUDACC__
55
+ printf("BAD sc ind. Must be 1,2,3. %d\n", sc);
56
+ #else
57
+ std::ostringstream oss;
58
+ int xxxx = 1;
59
+ oss << "Bad sc ind. Must be 1,2,3. Input sc is " << sc << " " << xxxx;
60
+ std::string var = oss.str();
61
+ throw std::invalid_argument(var);
62
+ #endif // __CUDACC__
63
+ }
64
+ return 0;
65
+ }
66
+
67
+ CUDA_DEVICE
68
+ double Orbits::interpolate(double t, double *in_arr, int window, int major_ndim, int major_ind, int ndim, int pos)
69
+ {
70
+ double up = in_arr[((window + 1) * major_ndim + major_ind) * ndim + pos]; // down_ind * ndim + pos];
71
+ double down = in_arr[(window * major_ndim + major_ind) * ndim + pos];
72
+
73
+ // m *(x - x0) + y0
74
+ double fin = ((up - down) / dt) * (t - (dt * window)) + down;
75
+ // if ((ndim == 1))
76
+ // printf("%d %e %e %e %e \n", window, fin, down, up, (t - (dt * window)));
77
+
78
+ return fin;
79
+ }
80
+
81
+ CUDA_DEVICE
82
+ void Orbits::get_normal_unit_vec_ptr(Vec *vec, double t, int link)
83
+ {
84
+ Vec _tmp = get_normal_unit_vec(t, link);
85
+ vec->x = _tmp.x;
86
+ vec->y = _tmp.y;
87
+ vec->z = _tmp.z;
88
+ }
89
+
90
+ CUDA_DEVICE
91
+ Vec Orbits::get_normal_unit_vec(double t, int link)
92
+ {
93
+ int window = get_window(t);
94
+ if (window == -1)
95
+ {
96
+ // out of bounds
97
+ return Vec(0.0, 0.0, 0.0);
98
+ }
99
+
100
+ int link_ind = get_link_ind(link);
101
+
102
+ int up_ind = (window + 1) * nlinks + link_ind;
103
+ int down_ind = window * nlinks + link_ind;
104
+
105
+ // x (pos = 0) ndim = 3
106
+ double x_out = interpolate(t, n_arr, window, nlinks, link_ind, 3, 0);
107
+ // y (pos = 1)
108
+ double y_out = interpolate(t, n_arr, window, nlinks, link_ind, 3, 1);
109
+ // z (pos = 2)
110
+ double z_out = interpolate(t, n_arr, window, nlinks, link_ind, 3, 2);
111
+
112
+ return Vec(x_out, y_out, z_out);
113
+ }
114
+
115
+ CUDA_DEVICE
116
+ double Orbits::get_light_travel_time(double t, int link)
117
+ {
118
+ int window = get_window(t);
119
+ if (window == -1)
120
+ {
121
+ // out of bounds
122
+ return 0.0;
123
+ }
124
+
125
+ int link_ind = get_link_ind(link);
126
+ if ((link_ind < 0) || (link_ind >= 6))
127
+ printf("BAD %d\n", link_ind);
128
+ int up_ind = (window + 1) * (nlinks + link_ind);
129
+ int down_ind = window * (nlinks + link_ind);
130
+
131
+ // x (pos = 0), ndim = 1
132
+ double ltt_out = interpolate(t, ltt_arr, window, nlinks, link_ind, 1, 0);
133
+
134
+ return ltt_out;
135
+ }
136
+
137
+ CUDA_DEVICE
138
+ Vec Orbits::get_pos(double t, int sc)
139
+ {
140
+ int window = get_window(t);
141
+ if (window == -1)
142
+ {
143
+ // out of bounds
144
+ return Vec(0.0, 0.0, 0.0);
145
+ }
146
+
147
+ int sc_ind = get_sc_ind(sc);
148
+
149
+ // x (pos = 0), ndim = 3
150
+ double x_out = interpolate(t, x_arr, window, nspacecraft, sc_ind, 3, 0);
151
+ // y (pos = 1), ndim = 3
152
+ double y_out = interpolate(t, x_arr, window, nspacecraft, sc_ind, 3, 1);
153
+ // z (pos = 2), ndim = 3
154
+ double z_out = interpolate(t, x_arr, window, nspacecraft, sc_ind, 3, 2);
155
+ return Vec(x_out, y_out, z_out);
156
+ }
157
+
158
+ CUDA_DEVICE
159
+ void Orbits::get_pos_ptr(Vec *vec, double t, int sc)
160
+ {
161
+ Vec _tmp = get_pos(t, sc);
162
+ vec->x = _tmp.x;
163
+ vec->y = _tmp.y;
164
+ vec->z = _tmp.z;
165
+ }
166
+
167
+ #define NUM_THREADS 64
168
+
169
+ CUDA_KERNEL
170
+ void get_light_travel_time_kernel(double *ltt, double *t, int *link, int num, Orbits& orbits)
171
+ {
172
+ int start, end, increment;
173
+ #ifdef __CUDACC__
174
+ start = blockIdx.x * blockDim.x + threadIdx.x;
175
+ end = num;
176
+ increment = gridDim.x * blockDim.x;
177
+ #else // __CUDACC__
178
+ start = 0;
179
+ end = num;
180
+ increment = 1;
181
+ #endif // __CUDACC__
182
+
183
+ for (int i = start; i < end; i += increment)
184
+ {
185
+ ltt[i] = orbits.get_light_travel_time(t[i], link[i]);
186
+ }
187
+ }
188
+
189
+ void Orbits::get_light_travel_time_arr(double *ltt, double *t, int *link, int num)
190
+ {
191
+ #ifdef __CUDACC__
192
+ int num_blocks = std::ceil((num + NUM_THREADS - 1) / NUM_THREADS);
193
+
194
+ // copy self to GPU
195
+ Orbits *orbits_gpu;
196
+ gpuErrchk(cudaMalloc(&orbits_gpu, sizeof(Orbits)));
197
+ gpuErrchk(cudaMemcpy(orbits_gpu, this, sizeof(Orbits), cudaMemcpyHostToDevice));
198
+
199
+ get_light_travel_time_kernel<<<num_blocks, NUM_THREADS>>>(ltt, t, link, num, *orbits_gpu);
200
+ cudaDeviceSynchronize();
201
+ gpuErrchk(cudaGetLastError());
202
+
203
+ gpuErrchk(cudaFree(orbits_gpu));
204
+
205
+ #else // __CUDACC__
206
+
207
+ get_light_travel_time_kernel(ltt, t, link, num, *this);
208
+
209
+ #endif // __CUDACC__
210
+ }
File without changes