Presentation / Installation


A Sparse Distributed Gigascale Resolution Material Point Method
SessionMost Def: Fast, Large and Learned Deformables
DescriptionWe present a four-layer distributed simulation system along with an MPM (Material Point Method) adaptation. Using hierarchical block-tile-cell sparse grids and an efficient dynamic load balance optimization, our system can be distributed to multiple MPI ranks with user-selected backend HPC programming models, including OpenMP and CUDA.
Authors
Yuxing Qiu
University of California Los Angeles
Samuel Temple Reeve
Oak Ridge National Laboratory
Minchen Li
University of California Los Angeles
Timestep Technologies
Yin Yang
The University of Utah
Timestep Technologies
Stuart Slattery
Oak Ridge National Laboratory
Chenfanfu Jiang
University of California Los Angeles
Timestep Technologies
Presenter
Yuxing Qiu
University of California, Los Angeles
Event Type
Technical Paper
TimeWednesday, 9 August 20232:22pm - 2:33pm PDT
LocationPetree Hall C
ACM Digital Library Technical Papers pdfs
Session Time & Location
Sunday, 6 August 20236pm - 8:30pm PDTWest Hall B
Wednesday, 9 August 20232pm - 3:30pm PDTPetree Hall C
Interest Areas
Research & Education
Recordings
Livestreamed
Recorded
Keywords
Animation/Simulation
Modeling
Registration Categories
FC
FCS
V
VS
EFC
Next PresentationNext Presentation
Second-order Stencil Descent for Interior-point Hyperelasticity