Hardware implementation of the exponent based computational core for an exchange-correlation potential matrix generation

Authors:
Maciej Wielgosz;Ernest Jamro;Kazimierz Wiatr
Affiliations:
AGH University of Science and Technology, Krakow and ACK Cyfronet, Krakow;AGH University of Science and Technology, Krakow and ACK Cyfronet, Krakow;AGH University of Science and Technology, Krakow and ACK Cyfronet, Krakow
Venue:
PPAM'09 Proceedings of the 8th international conference on Parallel processing and applied mathematics: Part I
Year:
2009

Citing 7
Cited 1

Computer arithmetic systems: algorithms, architecture and implementation

Computer arithmetic systems: algorithms, architecture and implementation
Optimization and evaluation of Hartree-Fock application's I/O with PASSION

SC '97 Proceedings of the 1997 ACM/IEEE conference on Supercomputing
Architectures and APIs: assessing requirements for delivering FPGA performance to applications

Proceedings of the 2006 ACM/IEEE conference on Supercomputing
Reconfigurable accelerator for quantum Monte Carlo simulations in N-body systems

Proceedings of the 2006 ACM/IEEE conference on Supercomputing
Converting massive TLP to DLP: a special-purpose processor for molecular orbital computations

Proceedings of the 4th international conference on Computing frontiers
FPGA acceleration of a quantum Monte Carlo application

Parallel Computing
Highly Efficient Structure of 64-Bit Exponential Function Implemented in FPGAs

ARC '08 Proceedings of the 4th international workshop on Reconfigurable Computing: Architectures, Tools and Applications

Comparison of GPU and FPGA implementation of SVM algorithm for fast image segmentation

ARCS'13 Proceedings of the 26th international conference on Architecture of Computing Systems

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper presents an FPGA implementation of a calculation module for a finite sum of the exponential products (orbital function). The module is composed of several specially designed floating-point modules which are, fully pipelined and optimized for high speed performance. The hardware implementation revealed significant speed-up for the finite sum of the exponential products calculation ranging from 2.5× to 20× in comparison to the CPU. The orbital function is a computationally critical part of the Hartree-Fock algorithm. The presented approach aims to increase the performance of the part of the quantum chemistry computational system by employing FPGA-based accelerator. Several issues are addressed such as an identification of proper code fragments, porting a part of the Hartree-Fock algorithm to FPGA, data precision adjustment and data transfer overheads. The authors' intention was also to make hardware application of the orbital function universal and easily attachable to different systems.