Concurrent query processing in a GPU-based database system

Abstract

The unrivaled computing capabilities of modern GPUs meet the demand of processing massive amounts of data seen in many application domains. While traditional HPC systems support applications as standalone entities that occupy entire GPUs, there are GPU-based DBMSs where multiple tasks are meant to be run at the same time in the same device. To that end, system-level resource management mechanisms are needed to fully unleash the computing power of GPUs in large data processing, and there were some researches focusing on it. In our previous work, we explored the single compute-bound kernel modeling on GPUs under NVidia’s CUDA framework and provided an in-depth anatomy of the NVidia’s concurrent kernel execution mechanism (CUDA stream). This paper focuses on resource allocation of multiple GPU applications towards optimization of system throughput in the context of systems. Comparing to earlier studies of enabling concurrent tasks support on GPU such as MultiQx-GPU, we use a different approach that is to control the launching parameters of multiple GPU kernels as provided by compile-time performance modeling as a kernel-level optimization and also a more general pre-processing model with batch-level control to enhance performance. Specifically, we construct a variation of multi-dimensional knapsack model to maximize concurrency in a multi-kernel environment. We present an in-depth analysis of our model and develop an algorithm based on dynamic programming technique to solve the model. We prove the algorithm can find optimal solutions (in terms of thread concurrency) to the problem and bears pseudopolynomial complexity on both time and space. Such results are verified by extensive experiments running on our microbenchmark that consists of real-world GPU queries. Furthermore, solutions identified by our method also significantly reduce the total running time of the workload, as compared to sequential and MultiQx-GPU executions.

Workload-Driven Analysis on the Performance Characteristics of GPU-Accelerated DBMSes

“…However, our research characterizes the performance of various query properties and identifies several bottlenecks in several state-of-the-art GPU DBMSes. Several studies [5], [6] have investigated concurrent query processing in a GPU DBMS, but their scope was limited to a single system rather than multiple disparate systems. Shanbhag et al [7] compared the performance gains of running queries on bandwidth ratios of CPUs and GPUs.…”

Design of Polyphase Channelization Algorithm Based on CUDA Stream Architecture

“…The multi-stream model on GPU platform includes OpenCL and CUDA, and the multistream model on Intel-MIC platform is h-stream. Current multi-stream model has been applied in many fields [7][8][9], but these researches are mostly aimed at the optimization of a specific scenario and programming model. Reference [7] proposed a performance model and an optimization strategy based on the intel-MIC platform h-stream model single application scenario.…”

“…Current multi-stream model has been applied in many fields [7][8][9], but these researches are mostly aimed at the optimization of a specific scenario and programming model. Reference [7] proposed a performance model and an optimization strategy based on the intel-MIC platform h-stream model single application scenario. Reference [8] conducted an in-depth study on CUDA multi-stream concurrency model, which conducted an in-depth analysis on the performance of CUDA platform single kernel and multi kernel streaming model.…”