[1910.00935] DiffTaichi: Differentiable Programming for Physical Simulation
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs. arXiv Operational Status Get status notifications via email or slack
View PDF HTML (experimental) Abstract:We present DiffTaichi, a new differentiable programming language tailored for building high-performance differentiable physical simulators. Based on an imperative programming language, DiffTaichi generates gradients of simulation steps using source code transformations that preserve arithmetic intensity and parallelism. A light-weight tape is used to record the whole simulation program structure and replay the gradient kernels in a reversed order, for end-to-end backpropagation. We demonstrate the performance and productivity of our language in…
saved by
related reading
- Physics-Informed Diffusion Modelsarxiv.org
- Differentiable Programming from Scratchthenumb.at
- [2403.14606] The Elements of Differentiable Programmingarxiv.org
- [2010.04017] DiffTune: Optimizing CPU Simulator Parameters with Learned Differentiable Surrogatesarxiv.org
- Tiny TPUtinytpu.com
- 2404.17625arxiv.org
- PyTorch internals : ezyang's blogblog.ezyang.com
- Differentiable Path Tracing on the GPU/TPUblog.evjang.com
- Tristan's Site - Tristan Humethume.ca
- di.ku.dk/english/research/phd/phd-theses/2025/dissertation_Robert.pdfdi.ku.dk
- GitHub - DiffEqML/torchdyn: A PyTorch library entirely dedicated to neural differential equations, implicit models and related numerical methodsgithub.com
- Welcome … — Physics-based Deep Learningphysicsbaseddeeplearning.org