Hardware acceleration of a Quantum Monte Carlo application

We are currently exploring the use of reconfigurable computing using Field Programmable Gate Arrays (FPGAs) to accelerate kernels of scientific applications. Here, we present a hardware architecture targeted towards the acceleration of two scientific kernels in a Quantum Monte Carlo (QMC) application applied to N-body systems. Quantum Monte Carlo methods enable us to determine the ground-state properties of atomic or molecular clusters. Here, we focus on two key kernels of the QMC application: acceleration of potential energy and wavefunction calculations. Our current platform consisting of a dual processor Intel Xeon 2.4 GHz augmented with two reconfigurable FPGA development boards provides a 3x speedup over the equivalent software only implementation. Targeting our design onto a High Performance Computing (HPC) system like the Cray XD1 or XT4 platform with high gate-density FPGAs will allow us to operate multiple instances of our design thereby providing additional parallelism.