# Red-black Gauss-Seidel **URL:** \ **Category:** Users\ **Created:** [October 22, 2025, 4:05pm UTC](https://chapel.discourse.group/t/red-black-gauss-seidel/45645 "2025-10-22T16:05:20Z")\ **Posts on this page:** 1\ **Showing post:** 7
**Author:** ![annacfs](https://yyz2.discourse-cdn.com/free1/user_avatar/chapel.discourse.group/annacfs/32/361_2.png) [@annacfs](https://chapel.discourse.group/u/annacfs)\ **Post date:** [October 27, 2025, 10:31am UTC](https://chapel.discourse.group/t/red-black-gauss-seidel/45645/7 "2025-10-27T10:31:54Z")
> [@Code optimization](https://chapel.discourse.group/t/code-optimization/17990/24): > > Hello @damianmoz and @annacfs, wrt parallelizing the SOR method, one possible option is to use a "checker-board" approach as described on this page: [CS267: Notes for Lectures 15 and 16, Mar 5 and 7, 1996](http://people.eecs.berkeley.edu/~demmel/cs267/lecture24/lecture24.html). The basic idea is to split your array into two groups of non-adjacent elements (in 2D, this ends up looking like the squares of a checkerboard). Then, the new values of each group can be computed in parallel, because they don't rely on the values of the neighboring elements. I've put together… --- _[View the full topic](https://chapel.discourse.group/t/red-black-gauss-seidel/45645)._