SEMINARS IN BIOMEDICAL ENGINEERING

General-purpose computing on graphics processing units for high speed Monte Carlo simulation of photon migration

Erik Alerstam

In this work we highlight the major advantage of using general-purpose computing on graphics processing units (GPGPU) for Monte Carlo (MC) simulations of photon migration. Using a low cost graphics processing unit (GPU; NVIDIA 8800GT), a standard simulation of time-resolved photon migration in a semi-infinite geometry is executed over 1000x faster on the GPU compared to a standard single CPU (Intel Pentium 4 HT 3.4 GHz) while providing equivalent results.
This tremendous speedup of GPGPU over the traditional approach for MC simulations can be further improved by using one or several higher-end cards, providing an estimated 10000x speedup using current generation hardware. This massive improvement in MC computation speed is a major leap towards making MC a viable option for photon migration modelling, including for example iterative data evaluation or modelling of complex geometries. Due to the simplicity, low cost and great performance the technique is expected to become a standard method in MC simulations of photon migration

[ Back ]

sponsored by

The Laser Microbeam and Medical Program (LAMMP)
a NIH biotechnology resource facility at the Beckman Laser Institute