A performance evaluation of CUBE: one-dimensional 512 FPGA cluster

Authors:
Masato Yoshimi;Yuri Nishikawa;Mitsunori Miki;Tomoyuki Hiroyasu;Hideharu Amano;Oskar Mencer
Affiliations:
Faculty of Science and Engineering, Doshisha University, Japan;Graduate School of Science and Technology, Keio University, Japan;Faculty of Science and Engineering, Doshisha University, Japan;Faculty of Science and Engineering, Doshisha University, Japan;Graduate School of Science and Technology, Keio University, Japan;Department of Computing, Imperial College London, UK
Venue:
ARC'10 Proceedings of the 6th international conference on Reconfigurable Computing: architectures, Tools and Applications
Year:
2010

Citing 2
Cited 2

Splash 2

SPAA '92 Proceedings of the fourth annual ACM symposium on Parallel algorithms and architectures
BEE2: A High-End Reconfigurable Computing System

IEEE Design & Test

Implementation and evaluation of an arithmetic pipeline on FLOPS-2D: multi-FPGA system

ACM SIGARCH Computer Architecture News
Accelerating floating-point fitness functions in evolutionary algorithms: a FPGA-CPU-GPU performance comparison

Genetic Programming and Evolvable Machines

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper reports an evaluation of CUBE, which is a multi-FPGA system which can connect 512 FPGAs in a form of a simple one dimensional array. As the system well suits as stream-oriented application platforms, we evaluated its performance by implementing edit distance computation algorithm, which is a typical streaming algorithm. Performances are compared with Cell/B.E., NVIDIA’s GeForce GTX280 and a general multi-core microprocessor. The report also analyzes pipeline utilization, and discusses performance efficiency, logic consumption and power efficiency with comparison to other multi-core devices.