文件列表:

LICENSE.txt (2226, 2015-10-20)
bf_knn_device.cu (9200, 2015-10-20)
bf_knn_device.cuh (3602, 2015-10-20)
bf_knn_host.cu (12109, 2015-10-20)
bf_knn_host.h (2905, 2015-10-20)
demo.cc (6117, 2015-10-20)
util.h (4888, 2015-10-20)

Brute-Force k-Nearest Neighbors Search on the GPU ================================================= **bf-knn** implements a brute-force approach for finding k-nearest neighbors on the GPU for many queries in parallel. It takes advantage of recent advances in fundamental GPU computing primitives. The squared Euclidean distances between queries and references are calculated by a CUDA kernel modified from a matrix multiplication subroutine in the [MAGMA](http://icl.cs.utk.edu/magma/) library. The nearest neighbors selection is accomplished by a truncated merge sort built on top of sorting and merging functions in the [Modern GPU](http://nvlabs.github.io/moderngpu/) library. Compared to state-of-the-art approaches, **bf-knn** is faster and handles larger inputs. Shengren Li and Nina Amenta. [**Brute-Force k-Nearest Neighbors Search on the GPU**](http://dx.doi.org/10.1007/978-3-319-25087-8_25). Similarity Search and Applications 2015. Instructions to download **bf-knn** and compile the demo: ``` git clone git@github.com:NVlabs/moderngpu.git git clone git@github.com:geomlab-ucd/bf-knn.git cd bf-knn nvcc -arch=sm_21 -I ../moderngpu/include/ bf_knn_device.cu bf_knn_host.cu demo.cc -o bf-knn-demo ``` [View License](LICENSE.txt)

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：