search

neka-nat / cupoch

Robotics with GPU computing
MIT License
918 stars 107 forks source link

DBSCAN slow than cpu method #85

Closed Xuechun-Hu closed 2 years ago

Xuechun-Hu commented 3 years ago

Test using a small point cloud with 15million points, and it takes extremely long to process. The gpu is nvidia a600 with 48gb memory. Cuda version is 11.2. Is cuml used in the library? It suffers the batch shrinking problem

[2021-10-12 12:00:22.331] [debug] Read geometry::PointCloud: 15636915 vertices.

[2021-10-12 12:00:22.340] [debug] [RemoveNoneFinitePoints] 15636915 nan points have been removed. (15636915, 3) (15636915, 3) block_0 has 15636915 points | ID | GPU | MEM |

| 0 | 18% | 2% | [2021-10-12 12:00:22.798] [debug] Precompute Neighbours Precompute Neighbours[================> ] 40%

neka-nat commented 3 years ago

Hi, Thank you for the reporting! Can you share the code and data for this experiment?

Xuechun-Hu commented 3 years ago

Hi @neka-nat

''' import cupoch as cph

cph.utility.set_verbosity_level(cph.utility.Debug)

pcl = cph.io.read_point_cloud( "data/pointcloud/block_0.ply") pcl_name = 'block_0' cph_labels = pcl.cluster_dbscan(eps=1, min_points=100, print_progress=True).cpu()

'''

I'm trying to cluster a (15636915, 2) dataset, which looks like this [[-230.19247 43.109245] [-225.90079 41.327675] [-222.1525 33.68174 ] ... [-224.55055 22.347908] [-224.50694 22.773636] [-223.6284 24.611273]]

neka-nat commented 3 years ago

Thank you for sharing the code! Can you also share "block_0.ply" pointcloud data?

Xuechun-Hu commented 3 years ago

https://drive.google.com/drive/folders/1FXQ9zT3uqEblHsw5xaC-w_KyUaczEAI4?usp=sharing there are two pointcloud files ,block_0 is the relatively smaller one

neka-nat commented 3 years ago

↑Can you give me access to Google Drive?