You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Related to JuliaApproximation/SemiclassicalOrthogonalPolynomials.jl#68, I have a workaround for the problem described therein but it relies on being able to tell when a matrix is not positive definite, the standard test for that is to just run Cholesky. The plan is to try cholesky raw, if it fails we can perturb a bit.
But:
When an infinite dimensional Cholesky decomposition fails (due to positive definiteness) it crashes in a strange uncatchable way. Here is an example:
It's strange because it looks to me like it doesn't realize an error occured and instead it loads the error into the cholesky factor. The finite dimensional case works.
The text was updated successfully, but these errors were encountered:
Related to JuliaApproximation/SemiclassicalOrthogonalPolynomials.jl#68, I have a workaround for the problem described therein but it relies on being able to tell when a matrix is not positive definite, the standard test for that is to just run Cholesky. The plan is to try cholesky raw, if it fails we can perturb a bit.
But:
When an infinite dimensional Cholesky decomposition fails (due to positive definiteness) it crashes in a strange uncatchable way. Here is an example:
It's strange because it looks to me like it doesn't realize an error occured and instead it loads the error into the cholesky factor. The finite dimensional case works.
The text was updated successfully, but these errors were encountered: