Optimized performance.

Makefile

@@ -1,6 +1,6 @@
 CXX = g++
 
-CXXFLAGS  = -std=c++14 -Wall -Wextra -Wno-unused-parameter -march=native -Ofast -fopenmp -flto
+CXXFLAGS  = -std=c++20 -Wall -Wextra -Wno-unused-parameter -march=native -Ofast -flto -fopenmp
 
 DEPS = util.h vec3.h color.h ray.h camera.h hittable.h hittable_list.h sphere.h material.h lodepng.h
 OBJ = wtracer.o material.o vec3.o lodepng.o
@@ -20,5 +20,5 @@ $(TARGET): $(OBJ)
 	$(CXX) $(CXXFLAGS) -o $@ $^
 
 clean:
-	$(RM) $(TARGET) $(OBJ) image.ppm
+	$(RM) $(TARGET) $(OBJ) image.png
 

wtracer.cpp

@@ -1,10 +1,15 @@
 #include <iostream>
 #include <iomanip>
 #include <memory>
+#include <cstdio>
 
 #define LODEPNG_NO_COMPILE_DECODER
 #include "lodepng.h"
 
+#ifdef _OPENMP
+#include <omp.h>
+#endif
+
 #include "color.h"
 #include "vec3.h"
 #include "ray.h"
@@ -78,11 +83,52 @@ Hittable_list setup_random_scene() {
     return world;
 }
 
+struct render_params {
+    int image_width;
+    int image_height;
+    int samples_per_pixel;
+    int max_depth;
+    Hittable_list& world;
+    Camera& cam;
+};
+
+void split_line(const int& line_length, int& start, int& end) {
+#ifdef _OPENMP
+    const int line_length_thread = static_cast<int>(line_length/omp_get_num_threads());
+    start = omp_get_thread_num() * line_length_thread;
+    end = std::min((omp_get_thread_num()+1) * line_length_thread, line_length);
+#else
+    start = 0;
+    end = line_length;
+#endif
+}
+
+void render(const render_params& params, std::vector<Color>& image) {
+    int line_start, line_end;
+    split_line(params.image_width, line_start, line_end);
+    for (int i = params.image_height-1; i>=0; --i) {
+        #ifdef _OPENMP
+        if (omp_get_thread_num() == 0)
+        #endif
+        std::cerr << "\rScanline remaining: " << std::setw(4) << i << std::flush;
+        for (int j = line_start; j<line_end; ++j) {
+            Color pixel_color(0, 0, 0);
+            for (int s = 0; s<params.samples_per_pixel; ++s) {
+                auto u = double(j + random_double(-0.5, 0.5)) / (params.image_width - 1);
+                auto v = double(i + random_double(-0.5, 0.5)) / (params.image_height - 1);
+                Ray r = params.cam.get_ray(u, v);
+                pixel_color += ray_color(r, params.world, params.max_depth);
+            }
+            image[i*params.image_width+j] = pixel_color;
+        }
+    }
+}
+
 int main() {
     const auto aspect_ratio = 16.0 / 9.0;
     //const int image_width = 1280;
-    const int image_width = 768;
+    //const int image_width = 768;
-    //const int image_width = 384;
+    const int image_width = 384;
     const int image_height = static_cast<int>(image_width / aspect_ratio);
     //const int samples_per_pixel = 1000;
     const int samples_per_pixel = 400;
@@ -98,21 +144,20 @@ int main() {
 
     Camera cam(lookfrom, lookat, vup, 20, aspect_ratio, aperture, dist_to_focus);
 
+    render_params params = {
+        image_width,
+        image_height,
+        samples_per_pixel,
+        max_depth,
+        world,
+        cam,
+    };
+
     std::vector<Color> image(image_width * image_height);
 
-    for (int i = image_height-1; i>=0; --i) {
+    #pragma omp parallel
-        std::cerr << "\rScanline remaining: " << std::setw(4) << i << std::flush;
+    {
-        #pragma omp parallel for
+        render(params, image);
-        for (int j = 0; j<image_width; ++j) {
-            Color pixel_color(0, 0, 0);
-            for (int s = 0; s<samples_per_pixel; ++s) {
-                auto u = double(j + random_double(-0.5, 0.5)) / (image_width - 1);
-                auto v = double(i + random_double(-0.5, 0.5)) / (image_height - 1);
-                Ray r = cam.get_ray(u, v);
-                pixel_color += ray_color(r, world, max_depth);
-            }
-            image.at(i*image_width+j) = pixel_color;
-        }
     }
 
     std::cerr << "\nAssembling image.\n";
@@ -120,7 +165,7 @@ int main() {
     std::vector<unsigned char> img_lode(image_width * image_height * 3);
     for (int i = 0; i<image_height; ++i) {
         for (int j = 0; j<image_width; ++j) {
-            write_color_vec(img_lode, i*image_width*3+j*3, image.at((image_height-1-i)*image_width+j), samples_per_pixel);
+            write_color_vec(img_lode, i*image_width*3+j*3, image[(image_height-1-i)*image_width+j], samples_per_pixel);
         }
     }
     std::cerr << "Writing file \"image.png\".\n";