An FPGA-Based Hardware Accelerator for Energy-Efficient Bitmap Index Creation

TL;DR

This paper introduces FPGA-based hardware accelerators, BIC64K8 and BIC32K16, that significantly speed up bitmap index creation for online analytics while reducing energy consumption compared to CPU and GPU solutions.

Contribution

The paper presents two novel FPGA-based accelerators for bitmap index creation, achieving high throughput and low energy consumption, which are improvements over existing CPU and GPU methods.

Findings

Achieve 1.43 GB/s and 1.46 GB/s indexing throughput at 100 MHz.

Maintain stable throughput regardless of data set size.

Consume only 6.76% and 3.28% energy compared to CPU and GPU implementations.

Abstract

Bitmap index is recognized as a promising candidate for online analytics processing systems, because it effectively supports not only parallel processing but also complex and multi-dimensional queries. However, bitmap index creation is a time-consuming task. In this study, by taking full advantage of massive parallel computing of field-programmable gate array (FPGA), two hardware accelerators of bitmap index creation, namely BIC64K8 and BIC32K16, are originally proposed. Each of the accelerator contains two primary components, namely an enhanced content-addressable memory and a query logic array module, which allow BIC64K8 and BIC32K16 to index 65,536 8-bit words and 32,768 16-bit words in parallel, at every clock cycle. The experimental results on an Intel Arria V 5ASTFD5 FPGA prove that at 100 MHz, BIC64K8 and BIC32K16 achieve the approximate indexing throughput of 1.43 GB/s and 1.46 GB/s, respectively. The throughputs are also proven to be stable, regardless the size of the data sets. More significantly, BIC32K16 only consumes as low as 6.76% and 3.28% of energy compared to the central-processing-unit- and graphic-processing-unit-based designs, respectively.