P-Cloth: Interactive Complex Cloth Simulation on Multi-GPU Systems using Dynamic Matrix Assembly and Pipelined Implicit Integrators

摘要

We present a novel parallel algorithm for cloth simulation that exploits multipleGPUs for fast computation and the handling of very high resolutionmeshes. To accelerate implicit integration, we describe new parallel algorithmsfor sparse matrix-vector multiplication (SpMV) and for dynamic matrixassembly on a multi-GPU workstation. Our algorithms use a novel workqueue generation scheme for a fat-tree GPU interconnect topology. Furthermore,we present a novel collision handling scheme that uses spatial hashingfor discrete and continuous collision detection along with a non-linear impact zone solver. Our parallel schemes can distribute the computationand storage overhead among multiple GPUs and enable us to perform almostinteractive simulation on complex cloth meshes, which can hardly behandled on a single GPU due to memory limitations. We have evaluatedthe performance with two multi-GPU workstations (with 4 and 8 GPUs, respectively)on cloth meshes with 0.5-1.65M triangles. Our approach canreliably handle the collisions and generate vivid wrinkles and folds at 2-5fps, which is significantly faster than prior cloth simulation systems. Weobserve almost linear speedups with respect to the number of GPUs.