Dual number and auto differentiation in C++

main.cpp

#include <iostream>

template<typename Scalar>
class DualNumber
{
public:
	inline DualNumber(const Scalar& realPart, const Scalar& dualPart = Scalar())
		:mReal(realPart)
		,mDual(dualPart)
	{
	}

	inline Scalar getReal() const {return mReal;}
	inline Scalar getDual() const {return mDual;}
private:
	Scalar mReal;
	Scalar mDual;
};

template<typename Scalar>
inline DualNumber<Scalar> makeDualNumber(const Scalar& realPart, const Scalar& dualPart = Scalar())
{
	return DualNumber<Scalar>(realPart, dualPart);
}

template<typename Scalar>
 inline DualNumber<Scalar> operator+(const DualNumber<Scalar>& lhs, const DualNumber<Scalar>& rhs)
{
	Scalar r1 = lhs.getReal();
	Scalar d1 = lhs.getDual();
	Scalar r2 = rhs.getReal();
	Scalar d2 = rhs.getDual();
	return makeDualNumber(r1 + r2, d1 + d2);
}

template<typename Scalar>
inline DualNumber<Scalar> operator*(const DualNumber<Scalar>& lhs, const DualNumber<Scalar>& rhs)
{
	Scalar r1 = lhs.getReal();
	Scalar d1 = lhs.getDual();
	Scalar r2 = rhs.getReal();
	Scalar d2 = rhs.getDual();
	return makeDualNumber(r1 * r2, r1 * d2 + r2 * d1);
}

template<typename Scalar>
inline DualNumber<Scalar> operator-(const DualNumber<Scalar>& lhs, const DualNumber<Scalar>& rhs)
{
	Scalar r1 = lhs.getReal();
	Scalar d1 = lhs.getDual();
	Scalar r2 = rhs.getReal();
	Scalar d2 = rhs.getDual();
	return makeDualNumber(r1 * r2, r1 * d2 + r2 * d1);
}

template<typename Scalar>
inline DualNumber<Scalar> operator/(const DualNumber<Scalar>& lhs, const DualNumber<Scalar>& rhs)
{
	Scalar r1 = lhs.getReal();
	Scalar d1 = lhs.getDual();
	Scalar r2 = rhs.getReal();
	Scalar d2 = rhs.getDual();
	return makeDualNumber(r1 / r2, (d1 * r2 - r1 * d2) / (r2 * r2));
}

template<typename Scalar>
inline DualNumber<Scalar> operator-(const DualNumber<Scalar>& operand)
{
	return makeDualNumber(- operand.getReal(), - operand.getDual())
}

template<typename Scalar>
Scalar testFunc(Scalar x)
{
	return (x * x) / (x + Scalar(1));
}

int main()
{
	std::cout<< testFunc<DualNumber<float>>(makeDualNumber(2.f, 1.f)).getDual() << std::endl;
	std::cin.get();
}

Hi, I found this useful.

I found a missing semi-colon here, and here, there is a >> after float that should be > >.