feat: implement multithreading for collision resolution and ball updates

src/Ball.cpp

@@ -125,6 +125,9 @@ void Ball::handleWallCollision(const sf::Vector2f& windowSize) {
 }
 
 void Ball::resolveCollision(Ball& other) {
+    std::lock_guard<std::mutex> lockA(m_mutex);
+    std::lock_guard<std::mutex> lockB(other.m_mutex);
+
     sf::Vector2f posA = getPosition();
     sf::Vector2f posB = other.getPosition();
     sf::Vector2f delta = posB - posA;
@@ -143,7 +146,7 @@ void Ball::resolveCollision(Ball& other) {
     float vA_n = vA.x * normal.x + vA.y * normal.y;
     float vB_n = vB.x * normal.x + vB.y * normal.y;
 
-    float restitution = Constants::RESTITUTION;
+    float restitution = 0.95f;
     float vA_n_new = vB_n * restitution;
     float vB_n_new = vA_n * restitution;
 

src/Ball.hpp

@@ -1,5 +1,6 @@
 #pragma once
 #include <SFML/Graphics/CircleShape.hpp>
+#include <mutex>
 
 #include "PhysicalObject.hpp"
 
@@ -15,6 +16,8 @@ class Ball : public PhysicalObject {
     void handleWallCollision(const sf::Vector2f& windowSize);
     void updateColor();
 
+    mutable std::mutex m_mutex;
+
    public:
     Ball(float radius, const sf::Vector2f& pos, const sf::Vector2f& vel,
          const sf::Color& color);

src/Constants.hpp

@@ -8,6 +8,6 @@ constexpr float RESTITUTION = 0.8f;
 constexpr float FRICTION = 0.9f;
 constexpr float IMPULSE = 2000.f;
 constexpr float REST_PIXEL_VELOCITY = 2.0f;
-constexpr float BALL_RADIUS = 15.f;
+constexpr float BALL_RADIUS = 6.f;
-constexpr unsigned BALL_QUANTITY = 500;
+constexpr unsigned BALL_QUANTITY = 1000;
 }  // namespace Constants

src/DebugOverlay.cpp

@@ -9,8 +9,9 @@ DebugOverlay::DebugOverlay(const std::string& fontPath)
     m_text.setPosition(sf::Vector2f(5.f, 5.f));
 }
 
-void DebugOverlay::update(int drawCalls, float timescale,
+void DebugOverlay::update(int drawCalls, float timeScale,
-                          const sf::RenderWindow& window) {
+                          sf::RenderWindow& window, size_t threadCount,
+                          size_t ballsPerThread) {
     float elapsed = m_fpsClock.restart().asSeconds();
     if (elapsed > 0.f) m_fps = static_cast<int>(1.f / elapsed);
 
@@ -21,8 +22,9 @@ void DebugOverlay::update(int drawCalls, float timescale,
     oss << "Framerate: " << m_fps << " FPS\n";
     oss << "Frametime: " << (elapsed * 1000.f) << " ms\n";
     oss << "Mouse: " << mousePos.x << ", " << mousePos.y << "\n";
-    oss << "Time scale: " << timescale << "\n";
+    oss << "Time scale: " << timeScale << "\n";
-
+    oss << "\nThreads: " << threadCount;
+    oss << "\nBalls per thread: " << ballsPerThread;
     m_text.setString(oss.str());
 }
 

src/DebugOverlay.hpp

@@ -9,7 +9,8 @@ class DebugOverlay {
    public:
     DebugOverlay(const std::string& fontPath);
 
-    void update(int drawCalls, float timescale, const sf::RenderWindow& window);
+    void update(int drawCalls, float timeScale, sf::RenderWindow& window,
+                size_t threadCount, size_t ballsPerThread);
     void draw(sf::RenderWindow& window);
 
    private:

src/Game.cpp

@@ -3,6 +3,7 @@
 #include <SFML/Window/Event.hpp>
 #include <algorithm>
 #include <cmath>
+#include <thread>
 #include <tuple>
 #include <unordered_map>
 
@@ -17,7 +18,10 @@ Game::Game()
           std::bind(&Game::processMousePressed, this, std::placeholders::_1)),
       m_debugLines(sf::PrimitiveType::Lines),
       m_drawCallCount(0),
-      m_debugOverlay("assets/consolas.ttf") {
+      m_debugOverlay("assets/consolas.ttf"),
+      m_threadPool(std::max(1u, std::thread::hardware_concurrency() > 2
+                                    ? std::thread::hardware_concurrency() - 2
+                                    : 1u)) {
     m_window.create(sf::VideoMode({Constants::WIDTH, Constants::HEIGHT}),
                     "SFML Playground");
     m_window.setVerticalSyncEnabled(true);
@@ -78,46 +82,92 @@ Game::Grid Game::buildSpatialGrid() {
     return grid;
 }
 
-void Game::resolveSpatialCollisions(const Grid &grid) {
+void Game::resolveSpatialCollisionsParallel(const Grid &grid) {
     static const Cell forwardNeighbors[] = {
         {0, 0}, {1, 0}, {1, 1}, {0, 1}, {-1, 1}};
+
+    std::vector<std::pair<Ball *, Ball *>> collisionPairs;
+    std::mutex collisionsMutex;
+
     for (const auto &[cell, cellBalls] : grid) {
         for (const auto &offset : forwardNeighbors) {
             Cell neighborCell = {cell.first + offset.first,
                                  cell.second + offset.second};
             auto neighborIt = grid.find(neighborCell);
             if (neighborIt == grid.end()) continue;
+
             if (neighborCell == cell) {
                 for (size_t i = 0; i < cellBalls.size(); ++i) {
                     for (size_t j = i + 1; j < cellBalls.size(); ++j) {
-                        cellBalls[i]->resolveCollision(*cellBalls[j]);
+                        std::lock_guard<std::mutex> lock(collisionsMutex);
+                        collisionPairs.emplace_back(cellBalls[i], cellBalls[j]);
                     }
                 }
             } else {
                 for (Ball *ballA : cellBalls) {
                     for (Ball *ballB : neighborIt->second) {
-                        ballA->resolveCollision(*ballB);
+                        std::lock_guard<std::mutex> lock(collisionsMutex);
+                        collisionPairs.emplace_back(ballA, ballB);
                     }
                 }
             }
         }
     }
+
+    const size_t chunkSize = std::max(
+        size_t(1), collisionPairs.size() / std::thread::hardware_concurrency());
+    std::vector<std::future<void>> futures;
+
+    for (size_t i = 0; i < collisionPairs.size(); i += chunkSize) {
+        size_t end = std::min(i + chunkSize, collisionPairs.size());
+
+        futures.push_back(m_threadPool.enqueue([&collisionPairs, i, end]() {
+            for (size_t j = i; j < end; ++j) {
+                auto &[ballA, ballB] = collisionPairs[j];
+                ballA->resolveCollision(*ballB);
+            }
+        }));
+    }
+
+    for (auto &future : futures) {
+        future.get();
+    }
 }
 
 void Game::update() {
     float dt = m_clock.restart().asSeconds() * m_timeScale;
     if (dt > 0.1f) dt = 0.1f;
 
-    for (auto &object : m_objects) {
+    updateBallsParallel(dt);
-        if (auto *ball = dynamic_cast<Ball *>(object.get())) {
-            ball->update(dt, m_windowSize);
-        } else {
-            object->update(dt, m_windowSize);
-        }
-    }
 
     auto grid = buildSpatialGrid();
-    resolveSpatialCollisions(grid);
+
+    resolveSpatialCollisionsParallel(grid);
+}
+
+void Game::updateBallsParallel(float dt) {
+    const size_t chunkSize = std::max(
+        size_t(1), m_objects.size() / std::thread::hardware_concurrency());
+
+    std::vector<std::future<void>> futures;
+
+    for (size_t i = 0; i < m_objects.size(); i += chunkSize) {
+        size_t end = std::min(i + chunkSize, m_objects.size());
+
+        futures.push_back(m_threadPool.enqueue([&, i, end, dt]() {
+            for (size_t j = i; j < end; ++j) {
+                if (auto *ball = dynamic_cast<Ball *>(m_objects[j].get())) {
+                    ball->update(dt, m_windowSize);
+                } else {
+                    m_objects[j]->update(dt, m_windowSize);
+                }
+            }
+        }));
+    }
+
+    for (auto &future : futures) {
+        future.get();
+    }
 }
 
 void Game::render() {
@@ -128,8 +178,12 @@ void Game::render() {
 
     m_batchRenderer.clear();
 
+    std::vector<const Ball *> balls;
+    balls.reserve(m_objects.size());
+
     for (const auto &object : m_objects) {
         if (auto *ball = dynamic_cast<Ball *>(object.get())) {
+            balls.push_back(ball);
             m_batchRenderer.addBall(*ball);
         } else {
             object->draw(m_window);
@@ -148,7 +202,10 @@ void Game::render() {
 
     if (m_toggleDebug) {
         DebugDraw::drawBatchedLines(m_window, m_debugLines);
-        m_debugOverlay.update(m_drawCallCount + 2, m_timeScale, m_window);
+        size_t threadCount = m_threadPool.getThreadCount();
+        size_t ballsPerThread = m_objects.size() / threadCount;
+        m_debugOverlay.update(m_drawCallCount + 2, m_timeScale, m_window,
+                              threadCount, ballsPerThread);
         m_debugOverlay.draw(m_window);
     }
 

src/Game.hpp

@@ -15,6 +15,7 @@
 #include "DebugOverlay.hpp"
 #include "InputManager.hpp"
 #include "PhysicalObject.hpp"
+#include "ThreadPool.hpp"
 
 class Ball;
 class PhysicalObject;
@@ -32,12 +33,7 @@ class Game {
     DebugOverlay m_debugOverlay;
     bool m_toggleDebug = true;
     BatchRenderer m_batchRenderer;
-
+    ThreadPool m_threadPool;
-    void processKeyPressed(const sf::Event::KeyPressed& keyPressed);
-    void processMousePressed(const sf::Event::MouseButtonPressed& mousePressed);
-    void handleMouseClick(const sf::Vector2i& mousePos);
-    void update();
-    void render();
 
     struct CellHash {
         std::size_t operator()(const std::pair<int, int>& k) const {
@@ -45,13 +41,21 @@ class Game {
                    static_cast<std::size_t>(k.second) * 19349663;
         }
     };
-
     using Cell = std::pair<int, int>;
     using Grid = std::unordered_map<Cell, std::vector<Ball*>, CellHash>;
+
+    void processKeyPressed(const sf::Event::KeyPressed& keyPressed);
+    void processMousePressed(const sf::Event::MouseButtonPressed& mousePressed);
+    void handleMouseClick(const sf::Vector2i& mousePos);
+    void update();
+    void render();
+    void updateBallsParallel(float dt);
+    void resolveSpatialCollisionsParallel(const Grid& grid);
+
     Grid buildSpatialGrid();
-    void resolveSpatialCollisions(const Grid& grid);
 
    public:
     Game();
     void run();
+    size_t getThreadCount() const { return m_threadPool.getThreadCount(); }
 };

src/ThreadPool.hpp

@@ -0,0 +1,83 @@
+#pragma once
+#include <condition_variable>
+#include <functional>
+#include <future>
+#include <mutex>
+#include <queue>
+#include <thread>
+#include <vector>
+
+class ThreadPool {
+   public:
+    explicit ThreadPool(size_t numThreads);
+    ~ThreadPool();
+    size_t getThreadCount() const { return workers.size(); }
+
+    template <class F, class... Args>
+    auto enqueue(F&& f, Args&&... args)
+        -> std::future<typename std::invoke_result<F, Args...>::type>;
+
+   private:
+    std::vector<std::thread> workers;
+
+    std::queue<std::function<void()>> tasks;
+
+    std::mutex queueMutex;
+    std::condition_variable condition;
+    bool stop;
+};
+
+inline ThreadPool::ThreadPool(size_t numThreads) : stop(false) {
+    for (size_t i = 0; i < numThreads; ++i) {
+        workers.emplace_back([this] {
+            while (true) {
+                std::function<void()> task;
+
+                {
+                    std::unique_lock<std::mutex> lock(this->queueMutex);
+                    this->condition.wait(lock, [this] {
+                        return this->stop || !this->tasks.empty();
+                    });
+
+                    if (this->stop && this->tasks.empty()) return;
+
+                    task = std::move(this->tasks.front());
+                    this->tasks.pop();
+                }
+
+                task();
+            }
+        });
+    }
+}
+
+inline ThreadPool::~ThreadPool() {
+    {
+        std::unique_lock<std::mutex> lock(queueMutex);
+        stop = true;
+    }
+    condition.notify_all();
+
+    for (auto& worker : workers) worker.join();
+}
+
+template <class F, class... Args>
+inline auto ThreadPool::enqueue(F&& f, Args&&... args)
+    -> std::future<typename std::invoke_result<F, Args...>::type> {
+    using return_type = typename std::invoke_result<F, Args...>::type;
+
+    auto task = std::make_shared<std::packaged_task<return_type()>>(
+        std::bind(std::forward<F>(f), std::forward<Args>(args)...));
+
+    std::future<return_type> res = task->get_future();
+
+    {
+        std::unique_lock<std::mutex> lock(queueMutex);
+        if (stop) throw std::runtime_error("enqueue on stopped ThreadPool");
+
+        tasks.emplace([task]() { (*task)(); });
+    }
+
+    condition.notify_one();
+    return res;
+}