International Journal on Advanced Science, Engineering and Information Technology, Vol. 9 (2019) No. 2, List of accepted papers, DOI:10.18517/ijaseit.9.2.8308

Design and evaluation of a scalable engine for 3D-FFT computation in an FPGA cluster

Roberto Ammendola, Pierpaolo Loreti

Abstract

The 3D FFT is commonly used to solve the partial differential equations describing the system evolution in several physical
phenomena. Simulation of such problems requires the use of parallel High-Performance Computing architecture since the size of the
problem grows rapidly. Modern HPC systems are considering the FPGAs as components of the computing architecture because they
can combine effective hardware acceleration capability and dedicated communication facilities.
In this paper, we explore an HPC design that uses FPGA accelerators to compute the 3DFFT. We devise a scalable FFT engine based
on a radix-2 double-precision core. The FFT engine can be adapted to the different technology constraints adjusting the module
configuration parameters. We compare the various possible configurations with the technological limits of the available hardware.

Keywords:

3D-FFT; FPGA; High Performance Computing; Cluster

Viewed: 21 times (since Sept 4, 2017)

cite this paper