SVD Calculator

Compute the full singular value decomposition of a real matrix and get orthogonal factors U and V and the rectangular diagonal matrix Σ.

Result
—

How to calculate

Rotate the columns to make them orthogonal. Column norms give the singular values, and normalizing the nonzero columns gives one basis. Accumulate the rotations for the other basis and complete both to square orthogonal matrices.

Worked example

Use Matrix A 3, 0 / 0, 2 / 0, 0. The result is A = UΣVᵀ.

Input and result notes

Use a real matrix up to 8 × 8. Numerical factors are rounded. Very small singular values must be resolvable at the supported precision. Use one row per line and separate entries with commas or spaces. Integers, decimals, and compact fractions such as 3/5 are supported. The table gives V transposed. Singular vectors can differ by signs or by a choice of basis in repeated singular subspaces.

LAPACK one-sided Jacobi singular value decomposition

Related calculators