my textbook classifies LU decomposition as a problem that can be easily parallilized (as in: a beginner to multi-threading should be able to implement it with ease, in several lines of code). Unfortunately, it doesn't go much more in detail about it.
So far, I've been using the Doolittle algorithm for LU decomposition (specifically for matrix determinant calculation), but I fail to see independant subtasks that could run in parallel, it seems to me that there's a lot of variable dependency going on here.
If I'm correct about the Doolittle algorithm being unsuited for simple parallelization, which algorithm could've been implied and what subtasks of it could run in parallel (f.e. can the processed matrix be divided into independant blocks)?
PS: I've googled "parallel LU decomposition" and the results were either research papers about very complex modern algorithms (not what I'm looking for) or uncommented mess of a code.
Thank you in advance.