Acceleration and execution of relational queries using general purpose graphics processing unit (GPGPU)

Database;GPU

Wu, Haicheng ; Yalamanchili, Sudhakar Electrical and Computer Engineering Kim, Hyesoon Wills, Linda Vuduc, Richard Pande, Santosh ; Yalamanchili, Sudhakar

University:Georgia Institute of Technology

Department:Electrical and Computer Engineering

关键词: Database;    GPU;   

Others  :  https://smartech.gatech.edu/bitstream/1853/54405/1/WU-DISSERTATION-2015.pdf

美国|英语

来源: SMARTech Repository

【 摘 要 】

This thesis first mapsthe relational computation onto Graphics Processing Units (GPU)s by designing aseries of tools and thenexplores the different opportunities of reducing the limitation brought by thememory hierarchy across the CPU and GPU system.First, a complete end-to-end compiler and runtime infrastructure, Red Fox, is proposed.Theevaluation on the full set ofindustry standard TPC-H queries on a single node GPUshows on average Red Fox is 11.20x faster compared with a commercial database system on a stateof art CPU machine.Second, a new compiler technique called kernel fusion is designed to fuse the code bodies of severalrelational operators to reduce data movement.Third, a multi-predicate join algorithm isdesigned for GPUs which can provide much better performance and be used withmore flexibility compared with kernel fusion.Fourth, the GPU optimized multi-predicate join is integrated into amulti-threaded CPU database runtime system that supports out-of-coredata set to solve real world problem.This thesis presents key insights, lessons learned, measurements from theimplementations, and opportunities for further improvements.