Category Simulation & Modeling Projects

Reading Time: 9 minutesA practical guide to choosing the right open-source license for your research code — MIT, BSD, Apache 2.0, and GPL explained in plain language. Includes funder mandates and institutional IP considerations.

Reading Time: 10 minutesA practical comparison of Gmsh, Netgen, and Trimesh for Python-based mesh generation. Learn when to use each tool for volume and surface meshing in scientific simulation.

Reading Time: 9 minutesWhat is a Research Software Engineer career? Entry paths, salary ranges, skill frameworks, GenAI impact, and how to break into RSE roles across academia, national labs, and industry.

Reading Time: 11 minutesLearn practical error control strategies for PDE solvers — adaptive mesh refinement, tolerance tuning, goal-oriented methods, and when to use each approach.

Reading Time: 13 minutesLearn why anisotropic mesh adaptation reduces degrees of freedom by 4–30× for advection-dominated flows. Covers metric tensors, error estimators, and practical implementation.

Reading Time: 9 minutesKey Takeaways Seven major funding models exist for open source scientific software, ranging from multi-donor philanthropy (EOSS/OS4S, ~$58M) to government grants (NASA OSTFL, NSF PESOSE, UKRI RSMF, DFG) and EU framework programs (€93.5B Horizon Europe). The “maintenance trap” is the root cause of abandoned research software: academic career systems reward publications, not code maintenance, creating […]

Reading Time: 12 minutesLearn how to make simulation results reproducible using Jupyter notebooks, version control, environment management, Zenodo archiving, and FAIR principles. A practical guide for computational researchers.

Reading Time: 9 minutesLearn when to use spectral methods, discontinuous Galerkin, spectral elements, and hp-FEM. Practical comparison of high-order methods vs traditional FDM, FEM, FVM.

Reading Time: 12 minutesAdvection-dominated partial differential equations are difficult because information moves mainly along a flow direction while physical diffusion provides little smoothing. A numerical method that works well for a diffusion-dominated problem may produce oscillations, excessive smearing, or nonphysical values when advection becomes stronger. The correct response is not simply to select the most diffusive scheme available. […]

Reading Time: 11 minutesLearn to write custom GPU kernels for physics simulation with NVIDIA Warp, CUDA Python, and HIP/ROCm. Covers performance patterns, differentiable physics, and production deployments.