Menu

Post image 1
Post image 2
Post image 3
Post image 4
Post image 5
1 / 5
0

Sparse Cholesky Elimination Tree

Math, Numerics, and Software·Math, Numerics, and Software·23 days ago
#GQjDewCi
Reading 0:00
15s threshold

April 09, 2026 Here I derive the elimination tree for the (right-looking) sparse Cholesky algorithm for computing A = LL^T for lower triangular L and sparse matrices A . This tree forms the foundation for most sparse factorization software, even when the underlying assumptions of Cholesky (symmetric and definite) do not apply. Ultimately this tree tells us two things: 1. where nonzeros appear in the matrix L even if not present in the original A (i.e. “fill-in”) and 2. the task dependency graph of our resulting factorization. Traditionally this concept is usually presented in the context sparse triangular solves which is then used as a building-block to a Cholesky factorization. I wanted to instead work directly from a Cholesky factorization, which is what I do below.…

Continue reading — create a free account

Join HashtagPLUS to read full articles, follow hashtags, vote, and join the conversation.

Read More