The above benchmarks were performed on an NVIDIA Jetson TX1 with a single thread. Compressed Directional Distance Transform (CDDT/PCDDT).└── lodepng # simple PNG loading/saving library ├── gflags # command line flag library from Google ├── distance_transform.h # for computing euclidean distance transform └── vendor # various dependencies, see in here for licenses │ ├── setup.py # compilation rules for Cython │ ├── RangeLibc.pyx # wrapper file for using RangeLib from Python ├── tmp/ # make this directory for saving fine-grained timing information ├── make_plots.py # turns fine-grained benchmark information into violin plots ├── main.cpp # example c++ usage and simple benchmarks │ └── RangeUtils.h # various utility functions │ ├── kernels.cu # cuda kernels for super fast 2D ray casting │ ├── CudaRangeLib.h # cuda function header file │ ├── RangeLib.h # main RangeLib source code │ ├── lru_cache.h # implementation of LRU_cache, optionally used ├── CMakeLists.txt # compilation rules - includes, etc │ └── bin # this is where compiled binaries will be placed
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |