update figures

Signed-off-by: Steven Hahn <[email protected]>

paper/paper.md

@@ -93,7 +93,6 @@ Many areas of science exhibit physical [^1] processes which are described by hig
 The Adaptive Sparse Grid Discretization (ASGarD) code is a framework specifically targeted at solving high-dimensional PDEs using a combination of a Discontinuous-Galerkin Finite Element solver implemented atop an adaptive sparse grid basis. The adaptivity aspect allows for the sparsity of the basis to be adapted to the properties of the problem of interest, which facilitates retaining the advantages of sparse grids in cases where the standard sparse grid selection rule is not the best match. A prototype of the non-adaptive sparse-grid implementation was used to produce the results of [@dazevedo2020] for 3D time-domain Maxwell's equations. The implementation utilizes both CPU and GPU resources, as well as being single and multi-node capable. Performance portability is achieved by casting the computational kernels as linear algebra operations and relying on vendor provided BLAS libraries. Several test problems are provided, including advection up to 6D with either explicit or implicit timestepping.
 
 ![Illustration of the curse of dimensionality in the context of solving a 6 dimensional PDE (e.g., those at the heart of magnetically confined fusion plasma physics) on modern supercomputers, and how the memory required to store the solution vector (solid black curves) and the matrix (magenta curves) in both naive and Sparse Grid based discretizations as the resolution of the simulation domain is varied. Memory limits of the Titan and Summit supercomputers at Oak Ridge National Laboratory, in addition to an approximate value for an ExaScale supercomputer, are overlaid for context](figures/sparse-vs-full.png)
-
 ![Potential memory savings of a Sparse Grid based solver represented as the ratio of the naive tensor product full-grid (FG) degrees of freedom (DoF) to the sparse-grid (SG) DoF.](figures/savings.png) 
 
 # Statement of Need