Dot product implementation in C++ - c++

Question

I am struggling to get my method to work. I am a beginner in C++. I am trying to find a dot product of two vectors. I am getting a few errors in the for loops. How can I fix them?
float dot(Matrixf const& vec1, Matrixf const& vec2) {
// error check
if (!vec1.isVector() || !vec2.isVector()) {
throw std::runtime_error("Unable to do dot product: not column vectors.");
}
if (vec1.nrows() != vec2.nrows()) {
throw std::runtime_error("Unable to do dot product: vector lengths not equal.");
}
/** implementing dot product *************************************/
float ret = 0;
for(unsigned i = 0; i < vec1.ncols(); ++i){
for(unsigned j =0; j< vec2.nrows(); ++j){
ret += vec1[i] * vec2[j];
}
}
return ret;
}
Matrixf class
#include "matrixf.h"
#include <iostream>
Matrixf::Matrixf(unsigned int rows, unsigned int cols) {
rows_ = rows;
cols_ = cols;
data_ = new float[rows_ * cols_];
// set all initial values to zero
for (unsigned int r = 0; r < rows_; ++r) {
for (unsigned int c = 0; c < cols_; ++c) {
data_[r * cols_ + c] = 0;
}
}
}
Matrixf::~Matrixf() {
delete data_;
}
Matrixf::Matrixf(Matrixf const& other) {
rows_ = other.rows_;
cols_ = other.cols_;
data_ = new float[rows_ * cols_];
for (unsigned int i = 0; i < rows_ * cols_; ++i) {
data_[i] = other.data_[i];
}
}
Matrixf& Matrixf::operator=(Matrixf const& other) {
// handles self assignment
if (this == &other) {
return *this;
}
delete data_;
rows_ = other.rows_;
cols_ = other.cols_;
data_ = new float[rows_ * cols_];
for (unsigned int i = 0; i < rows_ * cols_; ++i) {
data_[i] = other.data_[i];
}
return *this;
}
float Matrixf::get(unsigned int row, unsigned int col) const {
#ifndef NDEBUG
if (row >= rows_ || col >= cols_) {
throw std::runtime_error("Matrix index out of bounds.");
}
#endif
return data_[row * cols_ + col];
}
void Matrixf::set(unsigned int row, unsigned int col, float val) {
#ifndef NDEBUG
if (row >= rows_ || col >= cols_) {
throw std::runtime_error("Matrix index out of bounds.");
}
#endif
data_[row * cols_ + col] = val;
}
float& Matrixf::operator()(unsigned int row, unsigned int col) {
return data_[row * cols_ + col];
}
float Matrixf::operator()(unsigned int row, unsigned int col) const {
return data_[row * cols_ + col];
}
unsigned int Matrixf::nrows() const {
return rows_;
}
unsigned int Matrixf::ncols() const {
return cols_;
}
bool Matrixf::isVector() const {
return (cols_ == 1);
}
Matrixf Matrixf::eye(unsigned int size) {
Matrixf e(size, size);
for (unsigned int i = 0; i < size; ++i) {
e.set(i, i, 1);
}
return e;
}
std::ostream& operator << (std::ostream& os, Matrixf const& matrix) {
for (unsigned int r = 0; r < matrix.nrows(); ++r) {
for (unsigned int c = 0; c < matrix.ncols(); ++c) {
os << matrix.data_[r * matrix.cols_ + c] << " ";
}
os << "\n";
}
return os;
}

Answer 1

I think you just want one loop:
for(unsigned i = 0; i < vec1.ncols(); ++i){
ret += vec1[i] * vec2[i];
}
I also notice that you compare
vec1.nrows() != vec2.nrows()
but you use ncols() in the loop. Which one do you want?

Answer 2

I see from your other question that you write a ray tracer.
In ray tracers, it is common and virtually always the case that you have separate data structures for vectors and matrices, because they are almost always used differently, and specializations in programming almost always lead to faster code.
If you then define your dot product for only vectors, the dot product code will become simple.