Skip to content

Introduce CHOLMOD integration and covariance representations - #576

Open
peddie wants to merge 1 commit into
masterfrom
peddie/cholmod
Open

Introduce CHOLMOD integration and covariance representations#576
peddie wants to merge 1 commit into
masterfrom
peddie/cholmod

Conversation

@peddie

@peddie peddie commented Apr 23, 2026

Copy link
Copy Markdown
Contributor

Previously we offered only Eigen's dense LDLT decomposition as a backend for vanilla GP models. In the case of compact spatial kernels or other structured effects, the prior covariance may be sparse. To improve fitting performance on such problems, this commit introduces

  1. support for direct CHOLMOD solvers

  2. a CovarianceRepresentation based on CHOLMOD decompositions

This patch is going to fail CI until the bazel registry is updated.

Previously we offered only Eigen's dense LDLT decomposition as a
backend for vanilla GP models.  In the case of compact spatial kernels
or other structured effects, the prior covariance may be sparse.  To
improve fitting performance on such problems, this commit introduces

 1) support for direct CHOLMOD solvers

 2) a CovarianceRepresentation based on CHOLMOD decompositions
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant