matrix_multiply_scalar.cu File Reference

Implementation of the Matrix::multiply_scalar method for GPU-accelerated multiplication of a matrix by a scalar. More...

#include "matrix.h"
#include <cuda_runtime.h>
#include <stdexcept>
#include <cfloat>
#include <cmath>

Include dependency graph for matrix_multiply_scalar.cu:

Go to the source code of this file.

Functions
__global__ void	multiplyScalarKernel (double *data, double scalar, int size)
	CUDA kernel for multiplying matrix elements by a scalar. More...

Detailed Description

Implementation of the Matrix::multiply_scalar method for GPU-accelerated multiplication of a matrix by a scalar.

Definition in file matrix_multiply_scalar.cu.

Function Documentation

multiplyScalarKernel()

__global__ void multiplyScalarKernel	(	double *	data,
		double	scalar,
		int	size
	)

CUDA kernel for multiplying matrix elements by a scalar.

Parameters

data	Pointer to the matrix data.
scalar	The scalar to multiply by.
size	Total number of elements in the matrix.

Definition at line 18 of file matrix_multiply_scalar.cu.

                                                                            {
    // Calculate global thread index
    int idx = blockIdx.x * blockDim.x + threadIdx.x;
 
    // Check if thread is within matrix bounds
    if (idx < size) {
        // Perform multiplication
        double result = data[idx] * scalar;
 
        // Handle overflow
        if (!isfinite(result)) {
            result = (result > 0.0) ? DBL_MAX : -DBL_MAX;
        }
 
        // Store the result
        data[idx] = result;
    }
}