1: \begin{abstract}

2: We present a scalable iterative solver for high-order hybridized

3: discontinuous Galerkin (HDG) discretizations of linear partial

4: differential equations. It is an interplay between domain

5: decomposition methods and HDG discretizations, and hence inheriting

6: advances from both sides. In particular, the method can be viewed as a

7: Gauss-Seidel approach that requires only independent

8: element-by-element and face-by-face local solves in each iteration. As

9: such, it is well-suited for current and future computing systems with

10: massive concurrencies.  Unlike conventional Gauss-Seidel schemes which

11: are purely algebraic, the convergence of iHDG, thanks to the built-in

12: HDG numerical flux, does not depend on the ordering of unknowns.  We

13: rigorously show the convergence of the proposed method for the

14: transport equation, the linearized shallow water equation and the

15: convection-diffusion equation. For the transport equation, the method

16: is convergent regardless of mesh size $h$ and solution order $p$, and

17: furthermore the convergence rate is independent of the solution

18: order. For the linearized shallow water and the convection-diffusion

19: equations we show that the convergence is conditional on both $h$ and

20: $p$. Extensive steady and time-dependent numerical results for the 2D and

21: 3D transport equations, the linearized shallow water equation, and

22: the convection-diffusion equation are presented to verify the theoretical

23: findings.

24: 

25: \end{abstract}

26: