Task Scheduling for GPU Accelerated Hybrid OLAP Systems with Multi-core Support and Text-to-Integer Translation

Abstract

OLAP (On-Line Analytical Processing) is a powerful method for analyzing the excessive amount of data related to business intelligence applications. OLAP utilizes the efficient multidimensional data structure referred to as the OLAP cube to answer multi-faceted analytical queries. As queries become more complex and the dimensionality and size of the cube grows, the processing time required to aggregate queries increases.In this paper, we are proposing: (1) a parallel implementation of MOLAP cube using OpenMP; (2) a text-tointeger translation method to allow effective string processing on GPU; and (3) a new scheduling algorithm that support these new features.To be able to process string queries on the GPU, we are introducing a text-to-integer translation method which works with multiple dictionaries. The translation is necessary only for the GPU side of the system. To support the translation and parallel CPU implementation, a new scheduling algorithm is proposed. The scheduler divides multi-core processor(s) of a shared memory system into a processing partition and a preprocessing (or translation) partition.The performance of the new system is evaluated. The textto-integer translation adds a new vital functionality to our system; however it also slows down the GPU processing by 7% when compare to original implementation without string support.The performance measurements indicate that due to the parallel implementation, the processing rate of the CPU partition improves from 12 to 110 queries per second. Moreover, the CPU partition is now able to process OLAP cubes of size 32 GB at rate of 11 queries per second. The total performance of the entire hybrid system (CPU + GPU) increased from 102 to 228 queries per second.

Why it is time for a HyPE

“…Cost models are the base of such decisions, but estimating the cost of an operation for heterogeneous processors is a complex task. Mostly, an analytical cost model is built for the processing device and has to be calibrated according to the specific hardware in a system [18,30]. Some approaches use learning-based strategies, which map feature values (e.g., input data size, selectivity, data skew) to execution times [5,25].…”

“…Malik and others introduce a scheduling approach for hybrid OLAP systems [30]. They use a tailor-made calibrationbased analytical cost model to estimate execution times of CPU and GPU algorithms.…”

“…A new emerging trend applies General Purpose Graphical Processing (GPGPU) on database systems to create GPUaccelerated database management systems. In fact, a plenitude of research has investigated the acceleration of database systems using GPUs: relational operators [3,10,18], indexscan acceleration [4,28], kNN-search [12,39], online aggregation [29,30,38], sorting [14], spatial range queries [35], XML path filtering [31], duplicate detection [11], and MapReduce [17]. Due to the properties of GPUs, a GPU…”

A Framework for Developing Real-Time OLAP algorithm using Multi-core processing and GPU:Heterogeneous Computing

“…GPU in turns builds the cubes from traditional tables that are stored in GPU memory in addition to performing queries that are costly for CPU. GPU has six partitions that are used to process queries from fact tables, while CPU has two partitions, figure 2 - [4]. The same idea (exploiting GPU to serve OLAP purposes) has been addressed also by C. Weyerhaeuser et al [38] with Business Intelligence Accelerator (BIA: highly distributed analytical engine supports OLAP).…”

“…Hence, combining both CPU and GPU is a must in order to assign translation tasks to the CPU and processing tasks to the GPU. The proposed solution is essentially based on [4]. The algorithm will be based on TCP-DS benchmark by modifying the code of DsGen software which is open-source code.…”