Explore the use of an external Gaussian Quadrature package

[matt-pharr]

The Vacuum component and the in-progress Galerkin integrator for the Newcomb Euler-Lagrange equation both use gaussian quadrature to compute certain integrals. There are two main packages for this in Julia; GaussQuadrature.jl, and FastGaussQuadrature.jl. Each of these packages should be explored and their advantages/disadvantages considered to potentially exploit any speedup/simplification opportunities.

[jhalpern30]

A note on this - I currently have FastGaussQuadrature working in my 3D vacuum branch (#131). I think the main advantage is being able to cleanly do variable quadrature order at runtime. I don't think any speedup will be gained (at least in 2D vacuum) from using this package, since we just hardcode the coefficients for 8-point quadrature. I think the case where we would want to use this is if we decide that we'd like a variable quadrature order, which could be set in the input files. This would avoid having to hardcode a bunch of random Floats at the top of the file like we currently do for 8-point. Then, speedup could be obtained by showing that less than 8-point quadrature is sufficient. Another option is deriving some accuracy estimate from the numerical method and setting the quadrature order based on that so its not exposed to the user