High-order Finite Difference Modeling On Reconfigurable Computing Platform

High-order finite difference schemes are always considered appropriate for seismic modeling applications because of their excellent numerical properties in dispersion, dissipation, and anisotropy. However, these algorithms are too computationally intensive to be used routinely, especially for 3D acoustic and elastic modeling cases. In this paper, an inexpensive Field Programmable Gate Array (FPGA) based reconfigurable computing (RC) platform is proposed to accelerate the execution of ordinary secondorder and high-order FD schemes. Our early experiments demonstrated 4 X ~ 19 X performance improvements for different FD schemes comparing with their conventional pure software solutions running on a 3.0GHz referential Linux workstation. This RC platform is also consistent with the prevailing PC-Cluster systems and can achieve much better price-performance ratio along with much lower power consumption.

[1]  Leigh House,et al.  3-D elastic numerical modeling of a complex salt structure , 2000 .

[2]  J. Virieux.  P-SV wave propagation in heterogeneous media: Velocity‐stress finite‐difference method , 1986 .

[3]  Dennis W. Prather,et al.  FPGA-based acceleration of the 3D finite-difference time-domain method , 2004, 12th Annual IEEE Symposium on Field-Programmable Custom Computing Machines.

[4]  Keith D. Underwood,et al.  FPGAs vs. CPUs: trends in peak floating-point performance , 2004, FPGA '04.

[5]  Shawn Larsen,et al.  Elastic modeling initiative, part III: 3-D computational modeling , 1998 .

[6]  Mi Lu,et al.  Accelerating seismic migration using FPGA-based coprocessor platform , 2004, 12th Annual IEEE Symposium on Field-Programmable Custom Computing Machines.

[7]  J. Kristek,et al.  3D Displacement finite differences and a combined memory optimization , 1999 .

[8]  Karl S. Hemmert,et al.  Closing the gap: CPU and FPGA trends in sustainable floating-point BLAS performance , 2004, 12th Annual IEEE Symposium on Field-Programmable Custom Computing Machines.

[9]  Mi Lu,et al.  Time domain numerical simulation for transient waves on reconfigurable coprocessor platform , 2005, 13th Annual IEEE Symposium on Field-Programmable Custom Computing Machines (FCCM'05).