Presentation
Locality-Aware Automatic Differentiation on the GPU for Mesh-Based Computations
SessionSpeedy 3D
DescriptionWe present a GPU system for automatic differentiation for functions defined on triangle meshes that exploits locality and sparsity to deliver efficient derivative computation. From simple user-defined energy terms, the system produces gradients, Jacobians, and sparse Hessians, accelerating simulation and optimization workflows and shifting the computational bottleneck away from differentiation.
Event Type
Technical Paper
TimeMonday, 20 July 20262:50pm - 3:00pm PDT
LocationRoom 403 B
Digital Library
PDF
Session Time & Location
Sunday, 19 July 20266:00pm - 8:45pm PDTHall K
Monday, 20 July 20262:00pm - 3:30pm PDTRoom 403 B
Geometry
Full Conference Supporter
Full Conference
