Skip to content


Switch branches/tags

Name already in use

A tag already exists with the provided branch name. Many Git commands accept both tag and branch names, so creating this branch may cause unexpected behavior. Are you sure you want to create this branch?

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time


Test status License: MIT

ultraSEM is a software system for solving second-order linear partial differential equations (PDEs) on unstructured meshes in two dimension. The software is an implementation of the ultraspherical spectral element method [1], which allows hp-adaptivity to be efficiently performed in the high-p regime (where h is the mesh size and p is the polynomial order).


To install ultraSEM, simply add the top-level directory of this repository to your MATLAB path. This can be done via the pathtool command or by executing the command addpath(ultraSEMroot). To keep ultraSEM in your path for future MATLAB sessions, execute the savepath command.

Getting started

Solving PDEs with ultraSEM is easy. Suppose we have the PDE Lu = f, where L is a partial differential operator (PDO) and f is a function. Given a domain dom and boundary conditions bc, the solution u can be approximated using degree-p polynomials on each element of dom by executing

S = ultraSEM(dom, L, f, p);
u = S \ bc;

The ultraSEM object S is a direct solver for the PDE Lu = f, which may be invoked for the given boundary conditions via the \ command. The solution is returned as an ultraSEM.Sol object, which overloads a host of functions for plotting (e.g., plot, contour) and evaluation (e.g., feval, norm).

Constructing a domain

A domain is represented in the ultraSEM system as an ultraSEM.Domain object. Convenient functions for constructing rectangles, quadrilaterals, triangles, and polygons are available via the commands ultraSEM.rectangle, ultraSEM.quad, ultraSEM.triangle, and ultraSEM.polygon, respectively. Elements can be combined to form larger domains by merging them with the & operator. More general meshes can be constructed using the refine(dom) method, which performs uniform h-refinement on a given domain dom, or the refinePoint(dom, [x,y]) method, which performs adaptive h-refinement on dom around the point (x, y).

Specifying a PDE

A PDO is specified by its coefficients for each derivative in the form {{dxx, dxy, dyy}, {dx, dy}, b}, where each term dxx, dxy, dyy, dx, dy, and b can be a scalar (constant coefficient) or function handle (variable coefficient). The following syntax variations are permitted:

  • {{dxx, dyy}, __, __}, in which case dxy = 0.
  • {dxx, __, __}, in which case dyy = dxx and dxy = 0.
  • {__, dx, __}, in which case dy = dx.

The righthand side f and boundary conditions bc may be scalars or function handles.


Let's solve a simple problem with ultraSEM. First, construct a pentagonal domain:

dom = ultraSEM.polygon(5);

Then, refine the elements of the domain to make a finer mesh:

dom = refine(dom);

Let's solve the constant-coefficient Helmholtz equation lap(u) + 1000*u = -1 on this domain with zero Dirichlet boundary conditions:

L = {1, 0, 1000};
f = -1;
bc = 0;

Construct an ultraSEM object for this problem using 20th degree polynomials on each element of the mesh:

p = 20; 
S = ultraSEM(dom, L, f, p);

Solve the problem using the \ command:

u = S \ bc;


[1] Daniel Fortunato, Nicholas Hale, and Alex Townsend, The ultraspherical spectral element method, (2020).


The ultraspherical spectral element method







No releases published


No packages published