Categoria: SysAdmin/DevOps

Test article generated with aMELia Qt6. Several bug fixes, this is generated without prior model knowledge or external sources. A true local-first AI tool. The app was fed with Deitel C++ book and several advanced math PDFs from M.I.T.

1. Why “Advanced” C++?

The goal of this tutorial is not to teach you the syntax of C++ – that is already covered in the book’s early chapters – but to show how the language can be used to solve real mathematical and logical problems with correctness, safety, and performance.
We’ll build on the excerpts from C++_DEITEL.pdf and add:

• Deep dives into the concepts that matter for numerical work (overflow, precision, bounds‑checking).
• Hands‑on code that you can copy‑paste, compile, and run.
• Mini‑projects that tie together several advanced features.
• Exercises that reinforce the material and give you a chance to experiment.

Tip – Keep a small “sandbox” project folder (e.g., advanced_cpp/) and add a new .cpp file for each section. Compile with g++ -std=c++20 -Wall -Wextra -O2 -o <exe> <file>.cpp.

2. Foundations Re‑visited (for the mathematically‑oriented learner)

Exercise 1 – Write a function that returns the n‑th Fibonacci number using both an iterative loop and a recursive call. Compare the runtime for n = 40.

3. Numerical Algorithms – From Theory to Code

3.1 Newton–Raphson for Square Roots

The book mentions the importance of checking for overflow before performing arithmetic. In the Newton method we repeatedly compute x / guess, so we must guard against division by zero and ensure the intermediate values stay within double limits.

#include <cmath>
#include <stdexcept>

double sqrt_newton(double x, double tolerance = 1e-12) {
    if (x < 0) throw std::domain_error("Negative input");
    if (x == 0) return 0;
    double guess = x / 2.0;
    while (std::abs(guess * guess - x) > tolerance) {
        guess = (guess + x / guess) / 2.0;
    }
    return guess;
}

Explanation – The iteration g_{k+1} = (g_k + x/g_k)/2 converges quadratically.
Overflow check – x / guess can overflow if x is huge and guess is tiny; the guard x < 0 and the initial guess = x/2 keep us in a safe range.

Exercise 2 – Compare sqrt_newton with std::sqrt for x = 1e308. Which one is faster? Which one is more accurate?

3.2 Gaussian Elimination (Bare‑Bones)

The book’s discussion of sizeof and built‑in arrays reminds us that raw arrays lack bounds checking. We’ll use std::vector<std::vector<double>> to hold the matrix, which gives us dynamic sizing and automatic bounds checking.

#include <vector>
#include <stdexcept>

void gaussian_elimination(std::vector<std::vector<double>>& A,
                          std::vector<double>& b) {
    const size_t n = A.size();
    for (size_t k = 0; k < n; ++k) {
        // Simple pivoting: find row with max |A[i][k]|
        size_t pivot = k;
        for (size_t i = k + 1; i < n; ++i)
            if (std::abs(A[i][k]) > std::abs(A[pivot][k]))
                pivot = i;
        if (A[pivot][k] == 0) throw std::runtime_error("Singular matrix");

        std::swap(A[k], A[pivot]);
        std::swap(b[k], b[pivot]);

        for (size_t i = k + 1; i < n; ++i) {
            double factor = A[i][k] / A[k][k];
            for (size_t j = k; j < n; ++j) A[i][j] -= factor * A[k][j];
            b[i] -= factor * b[k];
        }
    }
    // Back substitution omitted for brevity
}

Why this is advanced – We’re manipulating raw memory (A[i][j]) but still using the safety of std::vector. The algorithm is O(n³) and is the backbone of many scientific libraries.

Exercise 3 – Solve the system
[
\begin{cases}
2x + 3y = 8\
5x – y = 2
\end{cases}
]
using gaussian_elimination. Verify the solution.

4. Symbolic Manipulation – Building an Expression Tree

The book’s discussion of static_cast<double> and setprecision shows how we can control numeric output. For symbolic work we need a representation of expressions that can be evaluated, differentiated, or simplified.

#include <memory>
#include <unordered_map>
#include <string>
#include <stdexcept>

enum class Op { Add, Sub, Mul, Div };

struct Expr {
    virtual double evaluate(const std::unordered_map<std::string, double>& vars) const = 0;
    virtual ~Expr() = default;
};

struct Var : Expr {
    std::string name;
    double evaluate(const std::unordered_map<std::string, double>& vars) const override {
        auto it = vars.find(name);
        if (it == vars.end()) throw std::runtime_error("Undefined variable");
        return it->second;
    }
};

struct Const : Expr {
    double value;
    double evaluate(const std::unordered_map<std::string, double>&) const override {
        return value;
    }
};

struct Binary : Expr {
    Op op;
    std::unique_ptr<Expr> left, right;
    double evaluate(const std::unordered_map<std::string, double>& vars) const override {
        double l = left->evaluate(vars);
        double r = right->evaluate(vars);
        switch (op) {
            case Op::Add: return l + r;
            case Op::Sub: return l - r;
            case Op::Mul: return l * r;
            case Op::Div: return l / r;
        }
        throw std::logic_error("Unknown operator");
    }
};

How to use it – Build an expression like x^2 + 3x + 5 by composing Binary nodes.
Why it’s advanced – The tree is a data structure that can be traversed recursively, enabling symbolic differentiation or simplification.

Exercise 4 – Write a function double derivative(const Expr&, const std::string& var) that returns the derivative of an expression with respect to var. (Hint: use the power rule for Binary::Mul where one child is a Const and the other a Var.)

5. Logic & Bit‑Level Operations

5.1 Simple Prime Test (Trial Division)

#include <cstdint>
#include <cmath>

bool is_prime(uint64_t n) {
    if (n < 2) return false;
    if (n % 2 == 0) return n == 2;
    for (uint64_t i = 3; i * i <= n; i += 2)
        if (n % i == 0) return false;
    return true;
}

Why this is useful – Many cryptographic algorithms rely on prime numbers. The function is O(√n) and safe for 64‑bit integers.

Exercise 5 – Use is_prime to generate the first 100 primes and store them in a std::vector<uint64_t>.

5.2 Compile‑time Logical Gates

template<bool A, bool B>
struct And { static constexpr bool value = A && B; };

template<bool A, bool B>
struct Or  { static constexpr bool value = A || B; };

template<bool A>
struct Not { static constexpr bool value = !A; };

Use case – Metaprogramming decisions based on compile‑time constants.
Example – static_assert(And<true, false>::value == false, "");

Exercise 6 – Implement a constexpr If<Cond, Then, Else> that selects one of two types based on a boolean condition.

6. Advanced Template Metaprogramming

6.1 Compile‑time Fibonacci

template<int N>
struct Fib {
    static constexpr int value = Fib<N-1>::value + Fib<N-2>::value;
};

template<>
struct Fib<0> { static constexpr int value = 0; };

template<>
struct Fib<1> { static constexpr int value = 1; };

Why it matters – Demonstrates recursion at compile time, which can be used for generating lookup tables or static assertions.

Exercise 7 – Compute Fib<30>::value at compile time and print it in main().

6.2 Type‑Safe Matrix Class

#include <array>
#include <cstddef>

template<typename T, std::size_t R, std::size_t C>
class Matrix {
    std::array<std::array<T, C>, R> data{};
public:
    T& operator()(std::size_t r, std::size_t c) { return data[r][c]; }
    const T& operator()(std::size_t r, std::size_t c) const { return data[r][c]; }

    template<std::size_t K>
    Matrix<T, R, K> operator*(const Matrix<T, C, K>& other) const {
        Matrix<T, R, K> result{};
        for (std::size_t i = 0; i < R; ++i)
            for (std::size_t j = 0; j < K; ++j)
                for (std::size_t k = 0; k < C; ++k)
                    result(i, j) += (*this)(i, k) * other(k, j);
        return result;
    }
};

Compile‑time safety – The dimensions are part of the type, so mismatched multiplications are caught by the compiler.

Exercise 8 – Create a Matrix<double, 3, 3> representing a rotation matrix and multiply it by a Matrix<double, 3, 1> representing a point.

7. Mini‑Projects (KB‑Prioritized)

Project Tip – Start with a specification (what the program should do), then write a test harness that verifies each small piece before integrating.

8. Testing & Validation

1. Unit tests – Use a lightweight framework like Catch2 or Google Test.
2. Property‑based tests – For numeric functions, test properties such as f(x) * f(1/x) == 1 for sqrt_newton.
3. Performance benchmarks – Measure the runtime of sqrt_newton vs std::sqrt for large inputs.
4. Static analysis – Run clang-tidy or cppcheck to catch potential overflows or undefined behavior.

Exercise 9 – Write a Catch2 test that verifies is_prime returns true for all primes below 1000.

9. Common Pitfalls & How to Avoid Them

10. Final Thoughts

• Precision first – Always choose the numeric type that matches the required precision.
• Safety first – Use constexpr, static_assert, and bounds‑checked containers to catch errors at compile time.
• Performance second – Profile before optimizing; use -O2 or -O3 and measure.
• Iterate – Start simple, test, then add complexity.

With the material above, you now have a roadmap to move from textbook examples to production‑ready, mathematically sound C++ code. Happy coding!