An application-specific instruction set for accelerating set-oriented database primitives

Abstract

The key task of database systems is to efficiently manage large amounts of data. A high query throughput and a low query latency are essential for the success of a database system. Lately, research focused on exploiting hardware features like superscalar execution units, SIMD, or multiple cores to speed up processing. Apart from these software optimizations for given hardware, even tailor-made processing circuits running on FPGAs are built to run mostly stateless query plans with incredibly high throughput. A similar idea, which was already considered three decades ago, is to build tailor-made hardware like a database processor. Despite their superior performance, such application-specific processors were not considered to be beneficial because generalpurpose processors eventually always caught up so that the high development costs did not pay off. In this paper, we show that the development of a database processor is much more feasible nowadays through the availability of customizable processors. We illustrate exemplarily how to create an instruction set extension for set-oriented database primitives. The resulting application-specific processor provides not only a high performance but it also enables very energyefficient processing. Our processor requires in various configurations more than 960x less energy than a high-end x86 processor while providing the same performance.

Advances in Databases and Information Systems

“…Therefore, disk Input-Output (IO) was the dominating cost factor. Nowadays, it is possible to equip servers with several terabytes of main memory, which allows us to keep databases in main memory to avoid the IO bottleneck (Arnold et al 2014). Therefore, the performance of databases became limited by memory access and processing power (Breß et al 2014).…”

Tiny GPU Cluster for Big Spatial Data: A Preliminary Performance Evaluation

“…Using large quantities of small processors to achieve similar throughputs while reducing energy footprint is becoming an increasingly important topic in Big Data research. Works on using low-power ASICs [12] and FPGAs [13] and power-efficient GPUs [14], Intel MICs [15] and SoCs [16] to process Big Data have been reported in the past few years with exciting results. Several previous works on evaluating standalone ARM-based systems for relational data query processing [17] [16] have shown that while these low profile systems are excellent for power-efficient computing under low utilization, they may not necessarily lead to significant energy saving under high utilization.…”

Hardware accelerators for information retrieval and data mining

“…Software/hardware co-design for data sort is studied in detail in [14][15][16][17][18][19]. It is shown in [17] that the fastest known even-odd merge and bitonic merge circuits [20] are very resource consuming and can only be used effectively in existing FPGAs for sorting very small data sets.…”