Multi-stream 2D frequency table computation on dataflow architecture

Author

Skoda, P. ; Sruk, V. ; Medved Rogina, B.

Author_Institution

Ruder Boskovic Inst., Zagreb, Croatia

fYear

2015

fDate

25-29 May 2015

Firstpage

275

Lastpage

280

Abstract

Frequency table computation is a common procedure used in variety of machine learning algorithms. In this paper we present a parallelized kernel for computing frequency tables. The kernel is targeted for dataflow architecture implemented on field programmable gate array (FPGA). Its performance was evaluated against a parallelized software implementation running on a 6-core CPU. The kernel with six concurrent input data streams running on 300 MHz achieved speedup of up to 6.26×, compared to 6 threaded software implementation running on 3.2 GHz CPU.

Keywords

data flow computing; field programmable gate arrays; learning (artificial intelligence); 6-core CPU; FPGA; concurrent input data streams; dataflow architecture; field programmable gate array; machine learning algorithms; multistream 2D frequency table computation; parallelized kernel; parallelized software implementation; threaded software implementation; Clocks; Computer architecture; Field programmable gate arrays; Kernel; Radiation detectors; Random access memory; Throughput; dataflow; decision tree learning; field programmable gate arrays; frequency table;

fLanguage

English

Publisher

ieee

Conference_Titel

Information and Communication Technology, Electronics and Microelectronics (MIPRO), 2015 38th International Convention on

Conference_Location

Opatija

Type

conf

DOI

10.1109/MIPRO.2015.7160279

Filename

7160279