warp.tile_cholesky_solve#

warp.tile_cholesky_solve(
L: Tile[Float, tuple[int, int]],
y: Tile[Float, tuple[int]],
) → Tile[Float, tuple[int]]#

  • Kernel

With L such that LL^T = A, solve for x in Ax = y

Note that computing the adjoint is not yet supported.

Supported datatypes are:

  • float32

  • float64

param L:

A square, lower triangular, matrix, such that LL^T = A

param y:

A 1D or 2D tile of length M

returns x:

A tile of the same shape as y such that LL^T x = y