GPU ACCELERATION OF DOPPLER BROADENING FOR NEUTRON TRANSPORT CALCULATIONS1

In radiation transport calculations, the effects of material temperature on neutron/nucleus interactions must be taken into account through Doppler broadening adjustments to microscopic cross section data. Historically, Monte Carlo simulations have accounted for this dependence by interpolating among precalculated broadened sections at a variety temperatures. More recently, there has been much interest in on-the-fly methods, where reference data is on-demand during particle any temperature. Unfortunately, operations are expensive traditional central processing unit (CPU) architectures, making unaffordable without approximations or complex preprocessing. This work considers use graphics (GPU)s, which excel parallel processing, continuous-energy simulations. Two methods considered – GPU implementation standard SIGMA1 algorithm and novel vectorized that leverages convolution properties operation an attempt expose additional parallelism. Numerical results demonstrate similar lookup throughput obtained as with precomputed CPU, implying offloading may enable treatment noticeable reduction performance codes.