Chang Yu - Research Page

About Me

I'm currently a four-year undergraduate student in Software Engineering at the University of Electronic Science and Technology of China (UESTC), and a Research & Development intern at Research & Simulation Team of Taichi Graphics, where I was fortunate enough to be supervised by Dr. Tiantian Liu.

My research interests lie in computer graphics and computational physics. In particular, I devote myself to physically-based simulation and high-performance computing on modern graphics architecture. I am working on the strong coupling of numerical optimization and computational parallelization to effectively and accurately model and simulate real-world physical phenomena like hair, cloth, fluid, and their interactions.

CV | Github | Email: [email protected]

Education

B.Eng. in Software Engineering

University of Electronic Science and Technology of China @ Chengdu, China — Aug 2019 – Jul 2023 (Expected)

Publications

MeshTaichi: A Compiler for Efficient Mesh-based Operations
Chang Yu*, Yi Xu*, Ye Kuang, Yuanming Hu, Tiantian Liu (* joint first authors)
ACM Transactions on Graphics [Proceedings of SIGGRAPH Asia], 2022 [Paper] [Project Page] [Video]
Meshes are an indispensable representation in many graphics applications because they provide conformal spatial discretizations. However, mesh-based operations are often slow due to unstructured memory access patterns. We propose MeshTaichi, a novel mesh compiler that provides an intuitive programming model for efficient mesh-based operations. Our programming model hides the complex indexing system from users and allows users to write mesh-based operations using reference-style neighborhood queries. Our compiler achieves its high performance by exploiting data locality. We partition input meshes and prepare the wanted relations by inspecting users’ code during compile time. During run time, we further utilize on-chip memory (shared memory on GPU and L1 cache on CPU) to access the wanted attributes of mesh elements efficiently. Our compiler decouples low-level optimization options with computations, so that users can explore different localized data attributes and different memory orderings without changing their computation code. As a result, users can write concise code using our programming model to generate efficient mesh-based computations on both CPU and GPU backends. We test MeshTaichi on a variety of physically-based simulation and geometry processing applications with both triangle and tetrahedron meshes. MeshTaichi achieves a consistent speedup ranging from 1.4 times to 6 times, compared to state-of-the-art mesh data structures and compilers.

Real-time Physics Engine Based on MPM & PBD
Yilong Wu*, Chang Yu* (* joint first authors)
2020 International Conference on Virtual Reality and Visualization (ICVRV) [Paper] [Video]
We have proposed a real-time physics engine that can handle various materials and phenomena, including deformable objects, granular objects and water. Furthermore, based on the latest MLS-MPM method proposed by academia, our engine can handle the coupling of water and sand.

Experiences

Mar 2021 - Present Research & Development Intern , Taichi Graphics, China Development of the Taichi programming language, simulation research based on Taichi Lang

Research, development and deployment of MeshTaichi extension
A research program. Developed and deployed a novel mesh compiler based on Taichi that provides an intuitive programming model for efficient mesh-based operations. Devised the programming interface and backend code generation. Implemented XPBD, Projective Dynamics, and Lagrangian-force MPM for experiments. Submitted a patent. Work published in SIGGRAPH Asia 2022.

Development of prototype of Taichi's Vulkan GPU backend
Developed and deployed an IR builder to convert CHI IR (Taichi's intermediate representation) to SPIR-V, which could be executed by Vulkan SDK, as a solution for non-CUDA environment and AOT compilation. Technical talks presented on TaichiCon01/02.

Teaching
Teaching Assistant of Taichi Graphics Course S1 (in Chinese), Fall 2021

Jun 2020 - Mar 2021 Undergraduate Studio Co-leader , UESTC Unreal Studio

Taichi Implementation of Power Particle-in-Cell transfer scheme extended upon a hybrid Eulerian-Lagrangian fluid solver. [Project Page]

Implemented an implicit material point method solver with an fixed-corotated elastic model in Taichi Lang, using the Newton-Raphson method and matrix-free preconditioned conjugate gradient method to solve non-linear equations. [Project Page]

Implemented an explicit material point method simulator for simulating porous sand and water mixtures using APIC and MLS-MPM transfer schemes. GAMES 201 final project. [Project Page]

Designed and developed a multi-threading CPU ray-tracing renderer equipped with hand-writing SIMD optimization. Included many algorithms in light transport simulation and appearance modeling. [Project Page]

Implemented an Euler fluid solver in Taichi Lang to simulate magnetic substance based on multigrid preconditioned conjugate gradient Poisson solver and level-set method. [Project Page]