make build pass

README.md

@@ -381,7 +381,9 @@ Overall, I'd describe the error handling capabilities of this library as pretty
 ### To do
 
 - [ ] Drive in a few more real-life applications
+  - [ ] US election modelling?
 - [ ] Look into using size_t instead of int for sample numbers
+- [ ] Reorganize code a little bit to reduce usage of gcc's nested functions
 
 ### Done
 

examples/core/01_one_sample/example.c

@@ -3,33 +3,13 @@
 #include <stdlib.h>
 
 // Estimate functions
-double sample_0(uint64_t* seed)
-{
-    UNUSED(seed);
-    return 0;
-}
 
-double sample_1(uint64_t* seed)
-{
-    UNUSED(seed);
-    return 1;
-}
+double sample_0(uint64_t* seed) { UNUSED(seed); return 0; }
+double sample_1(uint64_t* seed) { UNUSED(seed); return 1; }
+double sample_few(uint64_t* seed) { return sample_to(1, 3, seed); }
+double sample_many(uint64_t* seed) { return sample_to(2, 10, seed); }
 
-double sample_few(uint64_t* seed)
-{
-    return sample_to(1, 3, seed);
-}
-
-double sample_many(uint64_t* seed)
-{
-    return sample_to(2, 10, seed);
-}
-
-int main()
-{
-    // set randomness seed
-    uint64_t* seed = malloc(sizeof(uint64_t));
-    *seed = 1000; // xorshift can't start with 0
+double sample_model(uint64_t* seed){
 
     double p_a = 0.8;
     double p_b = 0.5;
@@ -38,8 +18,17 @@ int main()
     int n_dists = 4;
     double weights[] = { 1 - p_c, p_c / 2, p_c / 4, p_c / 4 };
     double (*samplers[])(uint64_t*) = { sample_0, sample_1, sample_few, sample_many };
+    double result = sample_mixture(samplers, weights, n_dists, seed);
 
-    double result_one = sample_mixture(samplers, weights, n_dists, seed);
-    printf("result_one: %f\n", result_one);
+    return result;
+}
+
+int main()
+{
+    // set randomness seed
+    uint64_t* seed = malloc(sizeof(uint64_t));
+    *seed = 1000; // xorshift can't start with 0
+
+    printf("result_one: %f\n", sample_model(seed));
     free(seed);
 }

examples/core/02_time_to_botec/example.c

@@ -2,37 +2,35 @@
 #include <stdio.h>
 #include <stdlib.h>
 
+double sample_0(uint64_t* seed) { UNUSED(seed); return 0; }
+double sample_1(uint64_t* seed) { UNUSED(seed); return 1; }
+double sample_few(uint64_t* seed) { return sample_to(1, 3, seed); }
+double sample_many(uint64_t* seed) { return sample_to(2, 10, seed); }
+
+double sample_model(uint64_t* seed){
+
+    double p_a = 0.8;
+    double p_b = 0.5;
+    double p_c = p_a * p_b;
+
+    int n_dists = 4;
+    double weights[] = { 1 - p_c, p_c / 2, p_c / 4, p_c / 4 };
+    double (*samplers[])(uint64_t*) = { sample_0, sample_1, sample_few, sample_many };
+    double result = sample_mixture(samplers, weights, n_dists, seed);
+
+    return result;
+}
+
 int main()
 {
     // set randomness seed
     uint64_t* seed = malloc(sizeof(uint64_t));
     *seed = 1000; // xorshift can't start with 0
 
-    double p_a = 0.8;
-    double p_b = 0.5;
-    double p_c = p_a * p_b;
-
-    double sample_0(uint64_t * seed)
-    {
-        UNUSED(seed);
-        return 0;
-    }
-    double sample_1(uint64_t * seed)
-    {
-        UNUSED(seed);
-        return 1;
-    }
-    double sample_few(uint64_t * seed) { return sample_to(1, 3, seed); }
-    double sample_many(uint64_t * seed) { return sample_to(2, 10, seed); }
-
-    int n_dists = 4;
-    double weights[] = { 1 - p_c, p_c / 2, p_c / 4, p_c / 4 };
-    double (*samplers[])(uint64_t*) = { sample_0, sample_1, sample_few, sample_many };
-
     int n_samples = 1000000;
     double* result_many = (double*)malloc((size_t)n_samples * sizeof(double));
     for (int i = 0; i < n_samples; i++) {
-        result_many[i] = sample_mixture(samplers, weights, n_dists, seed);
+        result_many[i] = sample_model(seed);
     }
     printf("Mean: %f\n", array_mean(result_many, n_samples));
 

examples/core/03_gcc_nested_function/example.c

@@ -2,39 +2,36 @@
 #include <stdio.h>
 #include <stdlib.h>
 
+double sample_model(uint64_t* seed){
+
+    double sample_0(uint64_t* seed) { UNUSED(seed); return 0; }
+    // Using a gcc extension, you can define a function inside another function
+    double sample_1(uint64_t* seed) { UNUSED(seed); return 1; }
+    double sample_few(uint64_t* seed) { return sample_to(1, 3, seed); }
+    double sample_many(uint64_t* seed) { return sample_to(2, 10, seed); }
+
+    double p_a = 0.8;
+    double p_b = 0.5;
+    double p_c = p_a * p_b;
+
+    int n_dists = 4;
+    double weights[] = { 1 - p_c, p_c / 2, p_c / 4, p_c / 4 };
+    double (*samplers[])(uint64_t*) = { sample_0, sample_1, sample_few, sample_many };
+    double result = sample_mixture(samplers, weights, n_dists, seed);
+
+    return result;
+}
+
 int main()
 {
     // set randomness seed
     uint64_t* seed = malloc(sizeof(uint64_t));
     *seed = 1000; // xorshift can't start with 0
 
-    double p_a = 0.8;
-    double p_b = 0.5;
-    double p_c = p_a * p_b;
-
-    int n_dists = 4;
-
-    // These are nested functions. They will not compile without gcc.
-    double sample_0(uint64_t * seed)
-    {
-        UNUSED(seed);
-        return 0;
-    }
-    double sample_1(uint64_t * seed)
-    {
-        UNUSED(seed);
-        return 1;
-    }
-    double sample_few(uint64_t * seed) { return sample_to(1, 3, seed); }
-    double sample_many(uint64_t * seed) { return sample_to(2, 10, seed); }
-
-    double (*samplers[])(uint64_t*) = { sample_0, sample_1, sample_few, sample_many };
-    double weights[] = { 1 - p_c, p_c / 2, p_c / 4, p_c / 4 };
-
     int n_samples = 1000000;
     double* result_many = (double*)malloc((size_t)n_samples * sizeof(double));
     for (int i = 0; i < n_samples; i++) {
-        result_many[i] = sample_mixture(samplers, weights, n_dists, seed);
+        result_many[i] = sample_model(seed);
     }
 
     printf("result_many: [");
@@ -42,5 +39,6 @@ int main()
         printf("%.2f, ", result_many[i]);
     }
     printf("]\n");
+
     free(seed);
 }

examples/core/04_gamma_beta/example.c

@@ -2,8 +2,6 @@
 #include <stdio.h>
 #include <stdlib.h>
 
-// Estimate functions
-
 int main()
 {
     // set randomness seed
@@ -11,33 +9,21 @@ int main()
     *seed = 1000; // xorshift can't start with 0
 
     int n = 1000 * 1000;
-    /*
+    double* gamma_array = malloc(sizeof(double) * (size_t)n);
     for (int i = 0; i < n; i++) {
-        double gamma_0 = sample_gamma(0.0, seed);
-        // printf("sample_gamma(0.0): %f\n", gamma_0);
+        gamma_array[i] = sample_gamma(1.0, seed);
     }
-		printf("\n");
-		*/
-
-    double* gamma_1_array = malloc(sizeof(double) * (size_t)n);
-    for (int i = 0; i < n; i++) {
-        double gamma_1 = sample_gamma(1.0, seed);
-        // printf("sample_gamma(1.0): %f\n", gamma_1);
-        gamma_1_array[i] = gamma_1;
-    }
-    printf("gamma(1) summary statistics = mean: %f, std: %f\n", array_mean(gamma_1_array, n), array_std(gamma_1_array, n));
-    free(gamma_1_array);
+    printf("gamma(1) summary statistics = mean: %f, std: %f\n", array_mean(gamma_array, n), array_std(gamma_array, n));
     printf("\n");
 
-    double* beta_1_2_array = malloc(sizeof(double) * (size_t)n);
+    double* beta_array = malloc(sizeof(double) * (size_t)n);
     for (int i = 0; i < n; i++) {
-        double beta_1_2 = sample_beta(1, 2.0, seed);
-        // printf("sample_beta(1.0, 2.0): %f\n", beta_1_2);
-        beta_1_2_array[i] = beta_1_2;
+        beta_array[i] = sample_beta(1, 2.0, seed);
     }
-    printf("beta(1,2) summary statistics: mean: %f, std: %f\n", array_mean(beta_1_2_array, n), array_std(beta_1_2_array, n));
-    free(beta_1_2_array);
+    printf("beta(1,2) summary statistics: mean: %f, std: %f\n", array_mean(beta_array, n), array_std(beta_array, n));
     printf("\n");
 
+    free(gamma_array);
+    free(beta_array);
     free(seed);
 }

examples/core/06_dissolving_fermi_paradox/example.c

@@ -4,16 +4,9 @@
 #include <stdio.h>
 #include <stdlib.h>
 
-int main()
-{
-    // Replicate <https://arxiv.org/pdf/1806.02404.pdf>, and in particular the red line in page 11.
-
-    // set randomness seed
-    uint64_t* seed = malloc(sizeof(uint64_t));
-    *seed = 1001; // xorshift can't start with a seed of 0
-
 double sample_fermi_logspace(uint64_t * seed)
 {
+    // Replicate <https://arxiv.org/pdf/1806.02404.pdf>, and in particular the red line in page 11.
     // You can see a simple version of this function in naive.c in this same folder
     double log_rate_of_star_formation = sample_uniform(log(1), log(100), seed);
     double log_fraction_of_stars_with_planets = sample_uniform(log(0.1), log(1), seed);
@@ -85,6 +78,14 @@ int main()
     // log_n > 0 => n > 1
 }
 
+
+int main()
+{
+
+    // set randomness seed
+    uint64_t* seed = malloc(sizeof(uint64_t));
+    *seed = 1001; // xorshift can't start with a seed of 0
+
     double logspace_fermi_proportion = 0;
     int n_samples = 1000 * 1000;
     for (int i = 0; i < n_samples; i++) {

examples/more/00_example_template/example.c

@@ -3,6 +3,10 @@
 #include <stdio.h>
 #include <stdlib.h>
 
+double sample_model(uint64_t* seed){
+    return sample_to(1, 10, seed);
+}
+
 int main()
 {
     // set randomness seed

examples/more/11_billion_lognormals_paralell/example.c

@@ -3,6 +3,10 @@
 #include <stdio.h>
 #include <stdlib.h>
 
+double sample_model(uint64_t * seed)
+{
+    return sample_lognormal(0, 10, seed);
+}
 // Estimate functions
 int main()
 {
@@ -13,13 +17,9 @@ int main()
 
     int n_samples = 1000 * 1000 * 1000;
     int n_threads = 16;
-    double sampler(uint64_t * seed)
-    {
-        return sample_lognormal(0, 10, seed);
-    }
     double* results = malloc((size_t)n_samples * sizeof(double));
 
-    sampler_parallel(sampler, results, n_threads, n_samples);
+    sampler_parallel(sample_model, results, n_threads, n_samples);
     double avg = array_sum(results, n_samples) / n_samples;
     printf("Average of 1B lognormal(0,10): %f\n", avg);
 

examples/more/12_time_to_botec_parallel/example.c

@@ -3,7 +3,7 @@
 #include <stdio.h>
 #include <stdlib.h>
 
-int main()
+double sampler_result(uint64_t * seed)
 {
     double p_a = 0.8;
     double p_b = 0.5;
@@ -17,11 +17,12 @@ int main()
     int n_dists = 4;
     double weights[] = { 1 - p_c, p_c / 2, p_c / 4, p_c / 4 };
     double (*samplers[])(uint64_t*) = { sample_0, sample_1, sample_few, sample_many };
-    double sampler_result(uint64_t * seed)
-    {
     return sample_mixture(samplers, weights, n_dists, seed);
 }
 
+int main()
+{
+
     int n_samples = 1000 * 1000, n_threads = 16;
     double* results = malloc((size_t)n_samples * sizeof(double));
     sampler_parallel(sampler_result, results, n_threads, n_samples);

squiggle_more.c

@@ -63,6 +63,8 @@ void sampler_parallel(double (*sampler)(uint64_t* seed), double* results, int n_
     {
 #pragma omp for
         for (i = 0; i < n_threads; i++) {
+            // It's possible I don't need the for, and could instead call omp
+            // in some different way and get  the thread number with omp_get_thread_num()
             int lower_bound_inclusive = i * quotient;
             int upper_bound_not_inclusive = ((i + 1) * quotient); // note the < in the for loop below,