Hello! I'm Huiyu Xie. My research interests are in numerical parallel algorithms, with a focus on nodal Discontinuous Galerkin methods on GPUs.
My current work is primarily focused on GPU programming for solving hyperbolic PDEs. I'm now working on open-source software projects, including Trixi.jl, TrixiCUDA.jl, and libparanumal.
Update on Mar 20, 2025: I will be giving a lightning talk about TrixiCUDA.jl at JuliaCon 2025 this summer at University of Pittsburgh/Carnegie Mellon.
Update on Mar 8, 2025: I will be visiting Virginia Tech at Tim Warburton's lab this summer to conduct research on GPU-based simulation.
Update on Jan 16, 2024: I will be mentoring a new CUDA project as part of Google Summer of Code 2025. Please reach out if you are interested.
I love open source development, and I'm working as:
Owner/Lead developer of Trixi-GPU, providing GPU acceleration support for solving hyperbolic PDEs, with a current focus on the semidiscretization process of the solvers.
Host of the organization's webpage trixi-gpu.github.io.
Member/Active maintainer of Trixi-Framework, providing low-precision (single-precision and half-precision) numerical type stability support for solving hyperbolic PDEs on CPUs.
Contributor of JuliaGPU, open source software GPU computing in Julia.
Contributor of SciML, open source software for scientific machine learning.
Contributor of RAPIDS, open source software for GPU-accelerated data science.
Reviewer of JuliaCon 2024, JuliaCon 2025.
Contributor of Google Summer of Code 2023.
More to explore...
Here is my recent CV (Last updated: August 5, 2025).
Regarding math, I love both analysis and proofs. I recommend Principles of Mathematical Analysis for its rigor and elegance. It's perfect for beginners in analysis and proofs.