Efficiently Running SQL Queries on GPU
Dimitri Dojchinovski, Marjan Gušev, Vladimir Zdraveski · 2018
In the last decade parallel computation (CUDA in particular), has accelerated the world. Applications used in astronomy, biology, chemistry, physics, data mining, manufacturing, finance, machine-learning and other computational intense fields are increasingly using CUDA to deliver the benefits of GPU acceleration. Also, CUDA is used in boosting the performance of database-oriented problems, speeding up the SQL performance on a huge dataset resulting in acquiring the data in rapid speed. This paper focuses on reading data from a. csv file, implementing kernels and performing the basic SQL aggregation functions on the GPU, and comparing their performance with the MySQL counterpart, showing the worthiness and importance of the concept.