feat: set task split to hardware_concurrency()

src/simulation.cpp

@@ -166,7 +166,8 @@ static void PopulateChanges() {
 	static bool first_run{ true };
 #if SIM_MULTITHREADING
 	BlockUntilTasksDone();
-	constexpr size_t split{ 4 };
+	// for some reason three tasks per thread is faster than one task per thread, i don't get it either
+	size_t const split{ std::thread::hardware_concurrency() };
 	if (first_run) {
 		first_run = false;
 		born.resize(split);