Inverse from Cholesky decomposition

Compute the inverse of a symmetric, positive definite matrix from its Cholesky decomposition. The input x should be an upper triangular matrix from chol().

Usage

chol2inv(x, size = NCOL(x), ...)

# Default S3 method
chol2inv(x, size = NCOL(x), ...)

# S3 method for class 'mlx'
chol2inv(x, size = NCOL(x), ..., device = NULL)

Arguments

x

An mlx matrix (2-dimensional array).

size

Ignored; included for compatibility with base R.

...

Additional arguments; ignored.

device

Execution target for APIs that expose a one-off device or stream override. Supply "gpu", "cpu", or an mlx_stream created via mlx_new_stream(). Ordinary array operations use the current mlx_device() instead.

Value

The inverse of the original matrix (before Cholesky decomposition).

Details

As of MLX 0.31.1, this operation only runs on CPU. Run it inside with_device() or local_device(), or pass device = "cpu".

See also

chol(), solve(), mlx_cholesky_inv()

Examples

A <- mlx_matrix(c(4, 1, 1, 3), 2, 2)
U <- chol(A, device = "cpu")
A_inv <- chol2inv(U, device = "cpu")
# Verify: A %*% A_inv should be identity
A %*% A_inv
#> mlx array [2 x 2]
#>   dtype: float32
#>   values:
#>      [,1] [,2]
#> [1,]    1    0
#> [2,]    0    1