fastlisaresponse 1.0.8__tar.gz → 1.0.10__tar.gz

{fastlisaresponse-1.0.8 → fastlisaresponse-1.0.10}/PKG-INFO

@@ -1,6 +1,6 @@
-Metadata-Version: 2.1
+Metadata-Version: 2.2
 Name: fastlisaresponse
-Version: 1.0.8
+Version: 1.0.10
 Home-page: https://github.com/mikekatz04/lisa-on-gpu
 Author: Michael Katz
 Author-email: mikekatz04@gmail.com
@@ -13,6 +13,13 @@ Classifier: Programming Language :: Cython
 Classifier: Programming Language :: Python :: 3.12
 Requires-Python: >=3.6
 Description-Content-Type: text/markdown
+Dynamic: author
+Dynamic: author-email
+Dynamic: classifier
+Dynamic: description
+Dynamic: description-content-type
+Dynamic: home-page
+Dynamic: requires-python
 
 # fastlisaresponse: Generic LISA response function for GPUs
 
@@ -40,7 +47,7 @@ See [examples notebook](https://github.com/mikekatz04/lisa-on-gpu/blob/master/ex
 
 ### Prerequisites
 
-Now (version 1.0.7) `fastlisaresponse` requires the newest version of [LISA Analysis Tools](github.com/mikekatz04/LISAanalysistools). You can run `pip install lisaanalysistools`.
+Now (version 1.0.10) `fastlisaresponse` requires the newest version of [LISA Analysis Tools](github.com/mikekatz04/LISAanalysistools). You can run `pip install lisaanalysistools`.
 
 To install this software for CPU usage, you need Python >3.4 and NumPy. To run the examples, you will also need jupyter and matplotlib. We generally recommend installing everything, including gcc and g++ compilers, in the conda environment as is shown in the examples here. This generally helps avoid compilation and linking issues. If you use your own chosen compiler, you will need to make sure all necessary information is passed to the setup command (see below). You also may need to add information to the `setup.py` file.
 
@@ -105,7 +112,7 @@ Please read [CONTRIBUTING.md](CONTRIBUTING.md) for details on our code of conduc
 
 We use [SemVer](http://semver.org/) for versioning. For the versions available, see the [tags on this repository](https://github.com/mikekatz04/lisa-on-gpu/tags).
 
-Current Version: 1.0.7
+Current Version: 1.0.10
 
 ## Authors
 

{fastlisaresponse-1.0.8 → fastlisaresponse-1.0.10}/README.md

@@ -24,7 +24,7 @@ See [examples notebook](https://github.com/mikekatz04/lisa-on-gpu/blob/master/ex
 
 ### Prerequisites
 
-Now (version 1.0.7) `fastlisaresponse` requires the newest version of [LISA Analysis Tools](github.com/mikekatz04/LISAanalysistools). You can run `pip install lisaanalysistools`.
+Now (version 1.0.10) `fastlisaresponse` requires the newest version of [LISA Analysis Tools](github.com/mikekatz04/LISAanalysistools). You can run `pip install lisaanalysistools`.
 
 To install this software for CPU usage, you need Python >3.4 and NumPy. To run the examples, you will also need jupyter and matplotlib. We generally recommend installing everything, including gcc and g++ compilers, in the conda environment as is shown in the examples here. This generally helps avoid compilation and linking issues. If you use your own chosen compiler, you will need to make sure all necessary information is passed to the setup command (see below). You also may need to add information to the `setup.py` file.
 
@@ -89,7 +89,7 @@ Please read [CONTRIBUTING.md](CONTRIBUTING.md) for details on our code of conduc
 
 We use [SemVer](http://semver.org/) for versioning. For the versions available, see the [tags on this repository](https://github.com/mikekatz04/lisa-on-gpu/tags).
 
-Current Version: 1.0.7
+Current Version: 1.0.10
 
 ## Authors
 

fastlisaresponse-1.0.10/fastlisaresponse/_version.py

@@ -0,0 +1 @@
+__version__ = "1.0.10"

{fastlisaresponse-1.0.8 → fastlisaresponse-1.0.10/fastlisaresponse/cutils}/include/LISAResponse.hh

@@ -16,7 +16,7 @@ void get_response(double *y_gw, double *t_data, double *k_in, double *u_in, doub
                   int projections_start_ind,
                   Orbits *orbits);
 
-void get_tdi_delays(double *delayed_links, double *input_links, int num_inputs, int num_delays, double *t_arr, int *tdi_base_link, int *tdi_link_combinations, int *tdi_signs_in, int *channels, int num_units, int num_channels,
+void get_tdi_delays(double *delayed_links, double *input_links, int num_inputs, int num_delays, double *t_arr, int *unit_starts, int *unit_lengths, int *tdi_base_link, int *tdi_link_combinations, int *tdi_signs_in, int *channels, int num_units, int num_channels,
                     int order, double sampling_frequency, int buffer_integer, double *A_in, double deps, int num_A, double *E_in, int tdi_start_ind, Orbits *orbits);
 
 #endif // __LISA_RESPONSE__

{fastlisaresponse-1.0.8 → fastlisaresponse-1.0.10/fastlisaresponse/cutils}/src/LISAResponse.cpp

@@ -236,11 +236,11 @@ void interp(double *result_hp, double *result_hc, cmplx *input, int h, int d, do
 #define MAX_ORDER 40
 
 CUDA_KERNEL
-void TDI_delay(double *delayed_links, double *input_links, int num_inputs, int num_delays, double *t_arr, int *tdi_base_link, int *tdi_link_combinations, int *tdi_signs_in, int *channels, int num_units, int num_channels,
+void TDI_delay(double *delayed_links, double *input_links, int num_inputs, int num_delays, double *t_arr, int *unit_starts, int *unit_lengths, int *tdi_base_link, int *tdi_link_combinations, int *tdi_signs_in, int *channels, int num_units, int num_channels,
                int order, double sampling_frequency, int buffer_integer, double *A_in, double deps, int num_A, double *E_in, int tdi_start_ind, Orbits *orbits_in)
 {
     Orbits orbits = *orbits_in;
-
+    int start, end, increment;
 #ifdef __CUDACC__
     CUDA_SHARED double input[BUFFER_SIZE];
 #endif
@@ -251,7 +251,6 @@ void TDI_delay(double *delayed_links, double *input_links, int num_inputs, int n
     CUDA_SHARED double A_arr[MAX_A_VALS];
     CUDA_SHARED double E_arr[MAX_ORDER];
 
-    int start, increment;
 #ifdef __CUDACC__
     start = threadIdx.x;
     increment = blockDim.x;
@@ -297,60 +296,61 @@ void TDI_delay(double *delayed_links, double *input_links, int num_inputs, int n
 // #pragma  omp parallel for
 #endif
 
-    for (int unit_i = start1; unit_i < num_units; unit_i += increment1)
-    {
-        int base_link = tdi_base_link[unit_i];
-        int base_link_index = orbits.get_link_ind(base_link);
-
-        int combination_link = tdi_link_combinations[unit_i];
-
-        int combination_link_index;
-        if (combination_link == -11)
-        {
-            combination_link_index = -1;
-        }
-        else
-        {
-            combination_link_index = orbits.get_link_ind(combination_link);
-        }
-        int sign = tdi_signs_in[unit_i];
-        int channel = channels[unit_i];
+    double t, L, delay;
 
-        int point_count = order + 1;
-        int half_point_count = int(point_count / 2);
+    double large_factor, pre_factor;
+    double clipped_delay, out, fraction;
+    double link_delayed_out;
+    int integer_delay, max_integer_delay, min_integer_delay;
 
-        int start2, increment2;
+    int point_count = order + 1;
+    int half_point_count = int(point_count / 2);
+    int start2, increment2;
 #ifdef __CUDACC__
-        start2 = tdi_start_ind + threadIdx.x + blockDim.x * blockIdx.x;
-        increment2 = blockDim.x * gridDim.x;
+    start2 = tdi_start_ind + threadIdx.x + blockDim.x * blockIdx.x;
+    increment2 = blockDim.x * gridDim.x;
 #else
-        start2 = tdi_start_ind;
-        increment2 = 1;
+    start2 = tdi_start_ind;
+    increment2 = 1;
 // #pragma  omp parallel for
 #endif
-        for (int i = start2;
-             i < num_delays - tdi_start_ind;
-             i += increment2)
-        {
-            double t, L, delay;
-
-            double large_factor, pre_factor;
-            double clipped_delay, out, fraction;
-            double link_delayed_out;
-            int integer_delay, max_integer_delay, min_integer_delay;
-            int start, end, increment;
+    for (int i = start2;
+         i < num_delays - tdi_start_ind;
+         i += increment2)
+    {
 
-            // at i = 0, delay ind should be at TDI_buffer = total_buffer - projection_buffer
-            t = t_arr[i];
-            if (combination_link == -11)
-            {
-                delay = t;
-            }
-            else
+        t = t_arr[i];
+        for (int unit_i = start1; unit_i < num_units; unit_i += increment1)
+        {
+            int unit_start = unit_starts[unit_i];
+            int unit_length = unit_lengths[unit_i];
+            int base_link = tdi_base_link[unit_i];
+            int base_link_index = orbits.get_link_ind(base_link);
+            int channel = channels[unit_i];
+            int sign = tdi_signs_in[unit_i];
+            delay = t;
+            for (int sub_i = 0; sub_i < unit_length; sub_i += 1)
             {
-                delay = t - orbits.get_light_travel_time(t, combination_link);
+                int combination_index = unit_start + sub_i;
+                int combination_link = tdi_link_combinations[combination_index];
+                int combination_link_index;
+                if (combination_link == -11)
+                {
+                    combination_link_index = -1;
+                }
+                else
+                {
+                    combination_link_index = orbits.get_link_ind(combination_link);
+                }
+
+                if (combination_link != -11)
+                {
+                    delay -= orbits.get_light_travel_time(t, combination_link);
+                }
             }
 
+            // at i = 0, delay ind should be at TDI_buffer = total_buffer - projection_buffer
+
             // delays are still with respect to projection start
             clipped_delay = delay;
             integer_delay = (int)ceil(clipped_delay * sampling_frequency) - 1;
@@ -412,7 +412,7 @@ void TDI_delay(double *delayed_links, double *input_links, int num_inputs, int n
     }
 }
 
-void get_tdi_delays(double *delayed_links, double *input_links, int num_inputs, int num_delays, double *t_arr, int *tdi_base_link, int *tdi_link_combinations, int *tdi_signs_in, int *channels, int num_units, int num_channels,
+void get_tdi_delays(double *delayed_links, double *input_links, int num_inputs, int num_delays, double *t_arr, int *unit_starts, int *unit_lengths, int *tdi_base_link, int *tdi_link_combinations, int *tdi_signs_in, int *channels, int num_units, int num_channels,
                     int order, double sampling_frequency, int buffer_integer, double *A_in, double deps, int num_A, double *E_in, int tdi_start_ind, Orbits *orbits_in)
 {
 
@@ -426,7 +426,7 @@ void get_tdi_delays(double *delayed_links, double *input_links, int num_inputs,
     gpuErrchk(cudaMemcpy(orbits_gpu, orbits_in, sizeof(Orbits), cudaMemcpyHostToDevice));
 
     // printf("RUNNING: %d\n", i);
-    TDI_delay<<<gridDim, NUM_THREADS>>>(delayed_links, input_links, num_inputs, num_delays, t_arr, tdi_base_link, tdi_link_combinations, tdi_signs_in, channels, num_units, num_channels,
+    TDI_delay<<<gridDim, NUM_THREADS>>>(delayed_links, input_links, num_inputs, num_delays, t_arr, unit_starts, unit_lengths, tdi_base_link, tdi_link_combinations, tdi_signs_in, channels, num_units, num_channels,
                                         order, sampling_frequency, buffer_integer, A_in, deps, num_A, E_in, tdi_start_ind, orbits_gpu);
 
     cudaDeviceSynchronize();
@@ -434,7 +434,7 @@ void get_tdi_delays(double *delayed_links, double *input_links, int num_inputs,
     gpuErrchk(cudaFree(orbits_gpu));
 
 #else
-    TDI_delay(delayed_links, input_links, num_inputs, num_delays, t_arr, tdi_base_link, tdi_link_combinations, tdi_signs_in, channels, num_units, num_channels,
+    TDI_delay(delayed_links, input_links, num_inputs, num_delays, t_arr, unit_starts, unit_lengths, tdi_base_link, tdi_link_combinations, tdi_signs_in, channels, num_units, num_channels,
               order, sampling_frequency, buffer_integer, A_in, deps, num_A, E_in, tdi_start_ind, orbits_in);
 
 #endif