Error in Singular Value Decomposition Example of Cuda Toolkit Documentation?

Accelerated Computing GPU-Accelerated Libraries
1

In the example for Singular Value Decomposition found under the following link https://docs.nvidia.com/cuda/cusolver/index.html#svd_examples, matrices U and VT are initialized in the following way:

const int m = 3;
    const int n = 2;
    const int lda = m;
/*       | 1 2  |
 *   A = | 4 5  |
 *       | 2 1  |
 */
    double A[lda*n] = { 1.0, 4.0, 2.0, 2.0, 5.0, 1.0};
    double U[lda*m]; // m-by-m unitary matrix 
    double VT[lda*n];  // n-by-n unitary matrix

Why is VT initialized as an array of length m*n if it is suppopsed to be an n-by-n matrix?

2

That is a bug. Thank you for bringing it to our attention.

VT should be n x n.