# Sparse cholesky factorization with user permutation (Julia 1.1)

**URL:** <https://discourse.julialang.org/t/sparse-cholesky-factorization-with-user-permutation-julia-1-1/27080>\
**Category:** Numerics\
**Tags:** linearalgebra\
**Created:** [August 1, 2019, 3:41pm UTC](https://discourse.julialang.org/t/sparse-cholesky-factorization-with-user-permutation-julia-1-1/27080 "2019-08-01T15:41:44Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![JuhaHeiskala](https://avatars.discourse-cdn.com/v4/letter/j/ecc23a/32.png) [@JuhaHeiskala](https://discourse.julialang.org/u/JuhaHeiskala)\
**Post date:** [August 1, 2019, 3:41pm UTC](https://discourse.julialang.org/t/sparse-cholesky-factorization-with-user-permutation-julia-1-1/27080/1 "2019-08-01T15:41:44Z")

</div>

The sparse cholesky factorization does not generally keep the user permutation given as parameter.  
SuiteSparse performs post ordering which is not disabled (as far as I can tell) in the current Julia interface to SuiteSparse when a user permutation is given as parameter.  
My impression is that the sparse cholesky should not permute the factorization of A when a user permutation is given as 1:size(A,1).

“fact\_” method in SuiteSparse/cholmod.jl has a “postorder” keyword argument, but that is not set to false in “cholesky” method when cholesky is called with user permutation argument defined.

This can be easily tested by creating a random sparse positive definite matrix A and calling “cholesky” with user permutation set as 1:size(A,1). All cases I have tried for 10x10 matrices has permuted the factorization.

Juha

---

<div class="post-metadata">

**Author:** ![andreasnoack](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/andreasnoack/32/27_2.png) [@andreasnoack](https://discourse.julialang.org/u/andreasnoack)\
**Post date:** [August 1, 2019, 7:53pm UTC](https://discourse.julialang.org/t/sparse-cholesky-factorization-with-user-permutation-julia-1-1/27080/2 "2019-08-01T19:53:08Z")

</div>

Indeed, see [https://github.com/JuliaLang/julia/issues/32760](https://github.com/JuliaLang/julia/issues/32760)