Merge pull request #70 from PeterTh/master

Added threadpool

Android.mk

@@ -45,6 +45,7 @@ LOCAL_SRC_FILES :=\
     profiler/profiler.cpp \
     thread/threadutil.cpp \
     thread/prioritizedworkqueue.cpp \
+    thread/threadpool.cpp \
     gfx_es2/glsl_program.cpp \
     gfx_es2/gl_state.cpp \
     gfx_es2/draw_buffer.cpp.arm \

native.vcxproj

@@ -245,6 +245,7 @@
     <ClInclude Include="profiler\profiler.h" />
     <ClInclude Include="thread\prioritizedworkqueue.h" />
     <ClInclude Include="thread\thread.h" />
+    <ClInclude Include="thread\threadpool.h" />
     <ClInclude Include="thread\threadutil.h" />
     <ClInclude Include="ui\screen.h" />
     <ClInclude Include="ui\ui.h" />
@@ -358,6 +359,7 @@
     <ClCompile Include="net\resolve.cpp" />
     <ClCompile Include="profiler\profiler.cpp" />
     <ClCompile Include="thread\prioritizedworkqueue.cpp" />
+    <ClCompile Include="thread\threadpool.cpp" />
     <ClCompile Include="thread\threadutil.cpp" />
     <ClCompile Include="ui\screen.cpp" />
     <ClCompile Include="ui\ui.cpp" />

native.vcxproj.filters

@@ -251,6 +251,9 @@
     <ClInclude Include="file\ini_file.h">
       <Filter>file</Filter>
     </ClInclude>
+    <ClInclude Include="thread\threadpool.h">
+      <Filter>thread</Filter>
+    </ClInclude>
   </ItemGroup>
   <ItemGroup>
     <ClCompile Include="gfx\gl_debug_log.cpp">
@@ -444,6 +447,9 @@
     <ClCompile Include="file\ini_file.cpp">
       <Filter>file</Filter>
     </ClCompile>
+    <ClCompile Include="thread\threadpool.cpp">
+      <Filter>thread</Filter>
+    </ClCompile>
   </ItemGroup>
   <ItemGroup>
     <Filter Include="gfx">

thread/thread.h

@@ -29,6 +29,7 @@
 // WIN32
 
 #define WIN32_LEAN_AND_MEAN
+#define NOMINMAX
 #include <Windows.h>
 
 #if defined(_MSC_VER) && defined(_MT)

thread/threadpool.cpp

@@ -0,0 +1,78 @@
+#include "threadpool.h"
+
+///////////////////////////// WorkerThread
+
+WorkerThread::WorkerThread() : active(true), started(false) {
+	thread = new std::thread(std::bind(&WorkerThread::WorkFunc, this));
+	doneMutex.lock();
+	while(!started) { };
+}
+
+WorkerThread::~WorkerThread() {
+	mutex.lock();
+	active = false;
+	signal.notify_one();
+	mutex.unlock();
+	thread->join();
+	delete thread;
+}
+
+void WorkerThread::Process(const std::function<void()>& work) {
+	mutex.lock();
+	work_ = work;
+	signal.notify_one();
+	mutex.unlock();
+}
+
+void WorkerThread::WaitForCompletion() {
+	done.wait(doneMutex);
+}
+
+void WorkerThread::WorkFunc() {
+	mutex.lock();
+	started = true;
+	while(active) {
+		signal.wait(mutex);
+		if(active) { 
+			work_();
+			doneMutex.lock();
+			done.notify_one();
+			doneMutex.unlock();
+		}
+	}
+}
+
+///////////////////////////// ThreadPool
+
+ThreadPool::ThreadPool(int numThreads) : numThreads(numThreads), workersStarted(false) {
+}
+
+void ThreadPool::StartWorkers() {
+	if(!workersStarted) {
+		for(int i=0; i<numThreads; ++i) {
+			workers.push_back(std::make_shared<WorkerThread>());
+		}
+		workersStarted = true;
+	}
+}
+
+void ThreadPool::ParallelLoop(std::function<void(int,int)> loop, int lower, int upper) {
+	mutex.lock();
+	StartWorkers();
+	int range = upper-lower;
+	if(range >= numThreads*2) { // don't parallelize tiny loops (this could be better, maybe add optional parameter that estimates work per iteration)
+		// could do slightly better load balancing for the generic case, 
+		// but doesn't matter since all our loops are power of 2
+		int chunk = range/numThreads; 
+		for(int s=lower, i=0; i<numThreads; s+=chunk, ++i) {
+			workers[i]->Process(std::bind(loop, s, std::min(s+chunk,upper)));
+		}
+		for(int i=0; i<numThreads; ++i) {
+			workers[i]->WaitForCompletion();
+		}
+	} else {
+		loop(lower, upper);
+	}
+	mutex.unlock();
+}
+

thread/threadpool.h

@@ -0,0 +1,57 @@
+#pragma once
+
+#include "thread.h"
+#include "base/mutex.h"
+
+#include <functional>
+#include <vector>
+
+// This is the simplest possible worker implementation I can think of
+// but entirely sufficient for the given purpose.
+// Only handles a single item of work at a time.
+class WorkerThread {
+public:
+	WorkerThread();
+	~WorkerThread();
+
+	// submit a new work item
+	void Process(const std::function<void()>& work);
+	// wait for a submitted work item to be completed
+	void WaitForCompletion();
+
+private:
+	std::thread *thread; // the worker thread
+	condition_variable signal; // used to signal new work
+	condition_variable done; // used to signal work completion
+	recursive_mutex mutex, doneMutex; // associated with each respective condition variable
+	volatile bool active, started;
+	std::function<void()> work_; // the work to be done by this thread
+
+	void WorkFunc();
+
+	WorkerThread(const WorkerThread& other); // prevent copies
+	void operator =(const WorkerThread &other);
+};
+
+// A thread pool manages a set of worker threads, and allows the execution of parallel loops on them
+// individual parallel loops are fully sequentialized to simplify synchronization, which should not 
+// be a problem as they should each use the entire system
+class ThreadPool {
+public:
+	ThreadPool(int numThreads);
+	// don't need a destructor, "workers" is cleared on delete, 
+	// leading to the stopping and joining of all worker threads (RAII and all that)
+
+	void ParallelLoop(std::function<void(int,int)> loop, int lower, int upper);
+
+private:
+	const int numThreads;
+	std::vector<std::shared_ptr<WorkerThread>> workers;
+	recursive_mutex mutex; // used to sequentialize loop execution
+
+	bool workersStarted;
+	void StartWorkers();
+	
+	ThreadPool(const ThreadPool& other); // prevent copies
+	void operator =(const WorkerThread &other);
+};