Carmack Fast Inverse Square Root

fastSqrt.cpp

#include <iostream>
#include <cmath>
#include <iomanip>
#include <atomic>
#include <mutex>
#include <thread>
#include <condition_variable>
#include <sstream>
#include <vector>
#include <fstream>

static int threadCount = 0;
std::mutex m;

float Q_rsqrt(float number) {
	long i;
	float x2, y;
	const float threehalfs = 1.5F;

	x2 = number * 0.5F;
	y = number;
	i = *(long *)&y;                       // evil floating point bit level hacking
	i = 0x5f3759df - (i >> 1);               // what the fuck? 
	y = *(float *)&i;
	y = y * (threehalfs - (x2 * y * y));   // 1st iteration
										   //	y  = y * ( threehalfs - ( x2 * y * y ) );   // 2nd iteration, this can be removed

	return y;
}

void loopThread(unsigned long low, unsigned long high) {
	unsigned long i = low;
	std::stringstream s;
	s << "csv/" << low << "-" << high << ".csv";

	std::ofstream file;
	file.open(s.str());

	do {
		float* f = (float*)&i;
		file << std::hex << i << "," << *f << "," << (1.0f / sqrtf(*f)) << "," << Q_rsqrt(*f) << "\n";
		++i;
	} while (i != high);

	m.lock();
	--threadCount;
	m.unlock();

	file.close();
}

int main() {
	std::vector<std::thread> threads;
	unsigned long i = 0;
	unsigned long amount = 256 * 256;
	do {
		threads.emplace_back(std::thread(loopThread, i, i + amount - 1));
		std::cout << "Starting: " << i << "-" << i + amount - 1 << "\n";

		m.lock();
		++threadCount;
		m.unlock();

    // I have no idea what I'm doing.
		while (threadCount >= std::thread::hardware_concurrency()) {}

		i += amount;
	} while (i != 0);

	for (auto& t : threads) {
		t.join();
	}
	return 0;
}