search

torch-points3d / torch-points-kernels

Pytorch kernels for spatial operations on point clouds
MIT License
96 stars 25 forks source link

Add the implementation of Gridding reverse (arXiv 2006.03761) #58

Closed hzxie closed 1 year ago

hzxie commented 3 years ago

Hi, @nicolas-chaulet

The compiler says:

/usr/bin/ld: build/temp.linux-x86_64-3.8/cuda/src/gridding_gpu.o: in function `compute_index(int, int, int, int, int)':
tmpxft_003a9dfe_00000000-5_gridding_gpu.cudafe1.cpp:(.text+0x180): multiple definition of `compute_index(int, int, int, int, int)'; build/temp.linux-x86_64-3.8/cuda/src/gridding_reverse_gpu.o:tmpxft_003a9d70_00000000-5_gridding_reverse_gpu.cudafe1.cpp:(.text+0x180): first defined here

However, other functions (e.g., compute_weights, get_n_threads) do not have the same issue. I tried to put it into a header file (*.h) and include it in gridding_gpu.cu and gridding_reverse_gpu.cu. But the problem still exists. Can you help me to solve this problem?

nicolas-chaulet commented 3 years ago

Hi, I don't have a GPU ready at the moment... I might have time next week.