This source code implements the AWP-CH algorithm for parallel computing exact geodesic distances on manifold triangle meshes without boundaries.'
-m [mesh file]: input model file
-s [source index]: index of source
-o [output mesh file with geodesic distances information]: (only support .obj files)
AWP.exe -m bunny_nf144k.m -s 13340 -o bunny_info.obj
As research code, it does not handle skinny or degenerate triangles. The code has been tested on the following GPUs
- GTX 970 with 1664 CUDA cores and 2.44 Tflops;
- GTX Titan X (Maxwell) with 3072 CUDA cores and 7.0 T358 flops; and
- GTX Titan Xp with 3840 CUDA cores and 12 Tflops.
If using the code, please cite the paper
Ying et al., Parallelizing Discrete Geodesic Algorithms with Perfect Efficiency, Computer-Aided Design, 2019.