Jianting Zhang, Camille Kamga, Hongmian Gong, Le Gruenwald
2012 Proceedings of the ACM SIGKDD International Workshop on Urban Computing - UrbComp '12  
Volumes of urban sensing data captured by consumer electronic devices are growing exponentially and current disk-resident database systems are becoming increasingly incapable of handling such large-scale data efficiently. In this paper, we report our design and implementation of U 2 SOD-DB, a columnoriented, Graphics Processing Unit (GPU)-accelerated, inmemory data management system targeted at large-scale ubiquitous urban sensing origin-destination data. Experiment results show that U 2 SOD-DB
more » ... is capable of handling hundreds of millions of taxi-trip records with GPS recorded pickup and dropoff locations and times efficiently. Spatial and temporal aggregations on 150 million pickup locations and times in middle-town and downtown Manhattan areas in the New York City (NYC) can be completed in a fraction of a second. This is 10-30X faster than a serial CPU implementation due to GPU accelerations. Spatially joining the 150 million taxi pickup locations with 43 thousand polygons in identifying trip purposes has reduced the runtime from 30.5 hours to around 1,000 seconds and achieved a two orders (100X) speedup using a hybrid CPU-GPU approach.
doi:10.1145/2346496.2346522 dblp:conf/kdd/ZhangKGG12 fatcat:aqxz5s23jjblziwllw6oc64eje