Fine-Grained Tuple Transfer for Pipelined Query Execution on CPU-GPU Coprocessor

To leverage the massively parallel capability of GPU for query execution, GPU databases have been studied for over a decade. Recently, researchers proposed to execute queries with both CPU and GPU in a pipelined approach. In the pipelined query execution, the cross-processor tuple transfer plays a crucial role for the overall query execution performance. The state-of-the-art solution achieves cross-processor tuple transfer using a queue-like data structure. However, it is coarse-grained due to the use of a single spin lock to achieve thread-safety. This design causes performance issues as it prevents the threads from accessing the queue simultaneously. In this paper, we propose a fine-grained tuple transfer mechanism. It employs decoupled enqueue/dequeue to enable two threads on different processors to access the queue at the same time. Moreover, this mechanism explores subqueue-based locking to enable the threads on the same processor to access the queue at the same time. In particular, we implement a prototype system, namely π QC, which adopts fine-grained tuple transfer. Our experiments show that π QC achieves an order of magnitude better performance than existing GPU databases such as HeavyDB.