Xeon e5 cpu

To validate the beam focusing performance, the flux output (optical fluence rate) generated by MCOCT was compared to the one from MCX [18 (link)]. The accumulated energy was stored within the 3D voxel matrix during the simulation. The 3D matrix with the deposited weight was normalized relative to the input power, resulting in the local absorption rate per watt (W) of the incident light in units of W/cm³/W_incident. Dividing this rate by the absorption coefficients of the corresponding voxels resulted in the local fluence rate $\mathrm{\varphi <!--\; \varphi \; -->}$ in W/cm²/W_incident. The tissue was represented in both simulators by a $200\times <!--\; \times \; -->200\times <!--\; \times \; -->300$ voxel geometry with isotropic voxels of $0.001$ cm³. The distance between the imaging lens and tissue surface $z_{f,I}$ was set at 0.85 cm with a beam width at imaging lens $D=0.1$ cm. In MCOCT, the illumination beam had a Gaussian profile at the imaging lens. In MCX, the trajectory was simulated as a hyperboloid of one sheet and the photon packets propagated in straight lines along the double-ruled surface, thereby mimicking a Gaussian-distributed photon launch [18 (link)]. For both simulations, optical properties were ${\mathrm{\mu <!--\; \mu \; -->}}_a=0.001$ cm^-1, ${\mathrm{\mu <!--\; \mu \; -->}}_s=0.01$ cm^-1, $n=1$ , and $g=0.5$ . MCOCT and MCX simulations were executed using 10⁶ photon packets and were run on a desktop computer equipped with an Intel Xeon E5 CPU at 3.50GHz, 16GB RAM and an NVIDIA TITAN Xp card.