Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “MONTE”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 451 records · Page 25

Quantum Monte Carlo solution of the dynamical mean field equations in real time

In this paper we present real-time inchworm quantum Monte Carlo results for single-site dynamical mean field theory on an infinite coordination number Bethe lattice. Our numerically exact results are obtained on the L-shaped Keldysh contour and, being evaluated in real time, avoid the analytic continuation issues typically encountered in Monte Carlo calculations. Our results show that inchworm Monte Carlo methods have now reached a state where they can be used as dynamical mean field impurity solvers and the dynamical sign problem can be overcome. As nonequilibrium problems can be simulated at the same cost, we envisage the main use of these methods as dynamical mean field solvers for time-dependent problems far from equilibrium.

36 MATERIALS SCIENCE↗

Lattice Monte Carlo for quantum Hall states on a torus

The Monte Carlo method is very useful for studying various model states proposed for the fractional quantum Hall systems. In this paper, we introduce a lattice Monte Carlo method based on an exact lattice formalism for quantum Hall problems defined on a torus. This “lattice representation” is applied to study the many body Berry phase of the composite-Fermi liquid phase in a half-filled Landau level. Here, the Monte Carlo result agrees with the exact numerical result found on small systems sizes and is consistent with the prediction from the Dirac fermion low-energy effective theory. Many other aspects of other model states, including the structure factor, Coulomb energy and particle-hole symmetry, are studied and discussed.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Quantum Monte Carlo method applied to non-Markovian barrier transmission

In nuclear fusion and fission, fluctuation and dissipation arise because of the coupling of collective degrees of freedom with internal excitations. Close to the barrier, quantum, statistical, and non-Markovian effects are expected to be important. In this work, a new approach based on quantum Monte Carlo addressing this problem is presented. The exact dynamics of a system coupled to an environment is replaced by a set of stochastic evolutions of the system density. The quantum Monte Carlo method is applied to systems with quadratic potentials. In all ranges of temperature and coupling, the stochastic method matches the exact evolution, showing that non-Markovian effects can be simulated accurately. A comparison with other theories, such as Nakajima-Zwanzig or time-convolutionless, shows that only the latter can be competitive if the expansion in terms of coupling constant is made at least to fourth order. A systematic study of the inverted parabola case is made at different temperatures and coupling constants. The asymptotic passing probability is estimated by different approaches including the Markovian limit. Large differences with an exact result are seen in the latter case or when only second order in the coupling strength is considered, as is generally assumed in nuclear transport models. In contrast, if fourth order in the coupling or quantum Monte Carlo method is used, a perfect agreement is obtained.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Advanced Monte Carlo modeling of prompt fission neutrons for thermal and fast neutron-induced fission reactions on {sup 239}Pu

Prompt fission neutrons following the thermal and 0.5 MeV neutron-induced fission reaction of {sup 239}Pu are calculated using a Monte Carlo approach to the evaporation of the excited fission fragments. Exclusive data such as the multiplicity distribution P({nu}), the average multiplicity as a function of fragment mass {nu}-bar(A), and many others are inferred in addition to the most used average prompt fission neutron spectrum {chi}(E{sub in},E{sub out}), as well as average neutron multiplicity {nu}-bar. Experimental information on these more exclusive data help constrain the Monte Carlo model parameters. The calculated average total neutron multiplicity is {nu}-bar{sub c}=2.871 in very close agreement with the evaluated value {nu}-bar{sub e}=2.8725 present in the ENDF/B-VII.0 library. The neutron multiplicity distribution P({nu}) is in very good agreement with the evaluation by Holden and Zucker. The calculated average spectrum differs in shape from the ENDF/B-VII.0 spectrum, evaluated with the Madland-Nix model. In particular, we predict more neutrons in the low-energy tail of the spectrum (below about 300 keV) than the Madland-Nix calculations, casting some doubts on how much scission neutrons contribute to the shape of the low-energy tail of the spectrum. The spectrum high-energy tail is very sensitive to the total kinetic energy distribution of the fragments as well as to the total excitation energy sharing at scission. Present experimental uncertainties on measured spectra above 6 MeV are too large to distinguish between various theoretical hypotheses. Finally, comparisons of the Monte Carlo results with experimental data on {nu}-bar(A) indicate that more neutrons are emitted from the light fragments than the heavy ones, in agreement with previous works.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Efficient Monte Carlo event generation for neutrino-nucleus exclusive cross sections

Modern neutrino-nucleus cross section computations need to incorporate sophisticated nuclear models to achieve greater predictive precision. However, the computational complexity of these advanced models often limits their practicality for experimental analyses. To address this challenge, we introduce a new Monte Carlo method utilizing normalizing flows to generate surrogate cross sections that closely approximate those of the original model while significantly reducing computational overhead. As a case study, we built a Monte Carlo event generator for the neutrino-nucleus cross section model developed by the Ghent group. This model employs a Hartree-Fock procedure to establish a quantum mechanical framework in which both the bound and scattering nucleon states are solutions to the mean-field nuclear potential. The surrogate cross sections generated by our method demonstrate excellent accuracy with a relative effective sample size of more than 98.4%, providing a computationally efficient alternative to traditional Monte Carlo sampling methods for differential cross sections.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Improving light propagation Monte Carlo simulations with accurate 3D modeling of skin tissue

In this paper, we present a 3D light propagation model to simulate multispectral reflectance images of large skin surface areas. In particular, we aim to simulate more accurately the effects of various physiological properties of the skin in the case of subcutaneous vein imaging compared to existing models. Our method combines a Monte Carlo light propagation model, a realistic three-dimensional model of the skin using parametric surfaces and a vision system for data acquisition. We describe our model in detail, present results from the Monte Carlo modeling and compare our results with those obtained with a well established Monte Carlo model and with real skin reflectance images.

60 APPLIED LIFE SCIENCES↗

Rare Events via Cross-Entropy Population Monte Carlo

Rare events are events that happen with very low frequency. Estimating rare event probabilities using Monte Carlo techniques is computationally expensive, often to the point of intractability, and special methods are required. Importance sampling (IS) is a well known technique that uses a proposal distribution in place of a target distribution to lower the variance of the estimator. Key to the success of IS methods is the choice of a proposal distribution, or the parameters governing the distribution. Adaptive importance sampling improves the parameters of a family or population of proposal distributions iteratively through trials. We present a novel cross-entropy population Monte Carlo algorithm, which adapts the parameters of proposals through the cross-entropy method. The proposed method stands apart from previous work in that we are not optimizing a mixture distribution. Instead, we leverage deterministic mixture weights and optimize the distributions individually through a reinterpretation of the typical derivation of the cross-entropy method. Demonstrations on rare event examples show that the algorithm can outperform existing resampling based population Monte Carlo methods, especially for higher-dimensional problems. Finally, we also demonstrate efficacy on a conjunction analysis problem.

97 MATHEMATICS AND COMPUTING↗

Hardware-Accelerated Ray Tracing of CAD-Based Geometry for Monte Carlo Radiation Transport

Monte Carlo radiation transport (MCRT) methods have been used to simulate radiation environments for many decades by tracking individual particles through a model to accumulate statistical information. MCRT geometry is historically formed using the constructive solid geometry (CSG). Recently, significant work has been performed to support simulations using computer-aided design (CAD)-based tessellated surfaces to support highly complex geometries. Ray tracing acceleration data structures from the rendering and visualization community are applied to accelerate particle tracking in CAD-based models. Despite these efforts, CSG representations provide the superior performance in surface intersection operations during particle flight. Concurrently, pseudo Monte Carlo methods have become prevalent in rendering applications to support more realistic models for scattering media, motivating innovations that are advantageous for MCRT simulations. Finally, the authors’ work extends these innovations by employing Intel’s Embree ray tracing kernel within a geometry toolkit for Monte Carlo to improve the simulation performance using CAD-based models by factors of 1.5 to 2.

42 ENGINEERING↗

Quantitative PET Imaging Using A Comprehensive Monte Carlo System Model

We present the complete image generation methodology developed for the RatCAP PET scanner, which can be extended to other PET systems for which a Monte Carlo-based system model is feasible. The miniature RatCAP presents a unique set of advantages as well as challenges for image processing, and a combination of conventional methods and novel ideas developed specifically for this tomograph have been implemented. The crux of our approach is a low-noise Monte Carlo-generated probability matrix with integrated corrections for all physical effects that impact PET image quality. The generation and optimization of this matrix are discussed in detail, along with the estimation of correction factors and their incorporation into the reconstruction framework. Phantom studies and Monte Carlo simulations are used to evaluate the reconstruction as well as individual corrections for random coincidences, photon scatter, attenuation, and detector efficiency variations in terms of bias and noise. Finally, a realistic rat brain phantom study reconstructed using this methodology is shown to recover >; 90% of the contrast for hot as well as cold regions. The goal has been to realize the potential of quantitative neuroreceptor imaging with the RatCAP.

38 RADIATION CHEMISTRY, RADIOCHEMISTRY, AND NUCLEA↗

Correlated Production and Analog Transport of Fission Neutrons and Photons Using Fission Models FREYA, FIFRELIN, and the Monte Carlo Code TRIPOLI-4

Fission modeling in general-purpose Monte Carlo transport codes often relies on average nuclear data provided by the international evaluation libraries. As such, only average fission multiplicities are available and correlations between fission neutrons and photons are missing. Whereas uncorrelated fission physics is usually sufficient for standard reactor core and radiation shielding calculations, correlated fission secondaries are required for specialized nuclear instrumentation and detector modeling. For coincidence counting detector optimization for instance, the precise simulation of fission neutrons and photons that remain correlated in time from birth to detection is essential. New developments were recently integrated into the Monte Carlo transport code TRIPOLI-4 to model fission physics more precisely, the purpose being to access event-by-event fission events from two different fission models: Fission Reaction Event Yield Algorithm (FREYA) and Fission Fragment Evaporation Leading to an Investigation of Nuclear Data (FIFRELIN). TRIPOLI-4 simulations can now be performed, either by connecting via an application programming interface to the Lawrence Livermore National Laboratory fission library including FREYA, or by reading the external fission event data files produced by FIFRELIN beforehand. These new capabilities enable us to easily compare results from the Monte Carlo transport calculations using the two fission models in a nuclear instrumentation application. In the first part of this paper, the broad underlying principles of the two fission models are recalled. We then present the experimental measurements of neutron angular correlations for 252 Cf(sf) and 240 Pu(sf). The correlations were measured for several neutron kinetic energy thresholds. In the latter part of this paper, simulation results are compared with the experimental data. Spontaneous fissions in 252 Cf and 240 Pu are modeled by FREYA or FIFRELIN. Emitted neutrons and photons are subsequently transported to an array of scintillators by TRIPOLI-4 in the analog mode to preserve their correlations. Angular correlations between fission neutrons obtained independently from these TRIPOLI-4 simulations, using either FREYA or FIFRELIN, are compared with the experimental results. In conclusion, for 240 Pu(sf), the measured correlations were used to tune one of the FREYA model parameters. Similarly, different sets of parameters have been tested in FIFRELIN to try to improve the agreement with the experimental data.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Experimental results and numerical modeling of a high-performance large-scale cryopump. I. Test particle Monte Carlo simulation

For the torus of the nuclear fusion project ITER (originally the International Thermonuclear Experimental Reactor, but also Latin: the way), eight high-performance large-scale customized cryopumps must be designed and manufactured to accommodate the very high pumping speeds and throughputs of the fusion exhaust gas needed to maintain the plasma under stable vacuum conditions and comply with other criteria which cannot be met by standard commercial vacuum pumps. Under an earlier research and development program, a model pump of reduced scale based on active cryosorption on charcoal-coated panels at 4.5 K was manufactured and tested systematically. The present article focuses on the simulation of the true three-dimensional complex geometry of the model pump by the newly developed ProVac3D Monte Carlo code. It is shown for gas throughputs of up to 1000 sccm ({approx}1.69 Pa m{sup 3}/s at T = 0 deg. C) in the free molecular regime that the numerical simulation results are in good agreement with the pumping speeds measured. Meanwhile, the capture coefficient associated with the virtual region around the cryogenic panels and shields which holds for higher throughputs is calculated using this generic approach. This means that the test particle Monte Carlo simulations in free molecular flow can be used not only for the optimization of the pumping system but also for the supply of the input parameters necessary for the future direct simulation Monte Carlo in the full flow regime.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Benchmarking of Monte Carlo simulation of bremsstrahlung from thick targets at radiotherapy energies

Several Monte Carlo systems were benchmarked against published measurements of bremsstrahlung yield from thick targets for 10-30 MV beams. The quantity measured was photon fluence at 1 m per unit energy per incident electron (spectra), and total photon fluence, integrated over energy, per incident electron (photon yield). Results were reported at 10-30 MV on the beam axis for Al and Pb targets and at 15 MV at angles out to 90 degree sign for Be, Al, and Pb targets. Beam energy was revised with improved accuracy of 0.5% using an improved energy calibration of the accelerator. Recently released versions of the Monte Carlo systems EGSNRC, GEANT4, and PENELOPE were benchmarked against the published measurements using the revised beam energies. Monte Carlo simulation was capable of calculation of photon yield in the experimental geometry to 5% out to 30 degree sign , 10% at wider angles, and photon spectra to 10% at intermediate photon energies, 15% at lower energies. Accuracy of measured photon yield from 0 to 30 degree sign was 5%, 1 s.d., increasing to 7% for the larger angles. EGSNRC and PENELOPE results were within 2 s.d. of the measured photon yield at all beam energies and angles, GEANT4 within 3 s.d. Photon yield at nonzero angles for angles covering conventional field sizes used in radiotherapy (out to 10 degree sign ), measured with an accuracy of 3%, was calculated within 1 s.d. of measurement for EGSNRC, 2 s.d. for PENELOPE and GEANT4. Calculated spectra closely matched measurement at photon energies over 5 MeV. Photon spectra near 5 MeV were underestimated by as much as 10% by all three codes. The photon spectra below 2-3 MeV for the Be and Al targets and small angles were overestimated by up to 15% when using EGSNRC and PENELOPE, 20% with GEANT4. EGSNRC results with the NIST option for the bremsstrahlung cross section were preferred over the alternative cross section available in EGSNRC and over EGS4. GEANT4 results calculated with the ''low energy'' physics list were more accurate than those calculated with the ''standard'' physics list.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Microbeam radiation therapy: A Monte Carlo study of the influence of the source, multislit collimator, and beam divergence on microbeams

Microbeam radiation therapy (MRT) is a new oncology method currently under development for the treatment of inoperable pediatric brain tumors. Monte Carlo simulation, or the computational study of radiation transport in matter, is often used in radiotherapy to theoretically estimate the dose required for treatment. However, its potential use in MRT dose planning systems is currently hindered by the significant discrepancies that have been observed between measured and theoretical dose and the PVDR (peak to valley dose ratio). The need to resolve these discrepancies is driven by the desirability of making MRT available to humans in the next few years. This article aims to resolve some of the discrepancies by examining the simplifications adopted in previous MRT Monte Carlo studies, such as the common practice of commencing microbeam transport on the surface of the target which neglects the influence of the distributed synchrotron source, multislit collimator, and the beam divergence between them. This article uses PENELOPE Monte Carlo simulation to investigate the influence of these beamline components upstream of the target on the lateral dose profiles and PVDRs of an array of 25 microbeams. It also compares the dose profiles and PVDRs of a microbeam array produced from a single simulation (full array) to those produced from the superposition of a single microbeam profile (sup array). The effect of modeling the distributed source and the beam divergence was an increase in the absorbed dose in the penumbral and valley regions of the microbeam profiles. Inclusion of the multislit collimator resulted in differences of up to 5 {mu}m in the FWHM of microbeam profiles across the array, which led to minor variations in the corresponding PVDR yields.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Considerations and limitations of fast Monte Carlo electron transport in radiation therapy based on precalculated data

The purpose of this work is to revisit the impediments and characteristics of fast Monte Carlo techniques for applications in radiation therapy treatment planning using new methods of utilizing pregenerated electron tracks. The limitations of various techniques for the improvement of speed and accuracy of electron transport have been evaluated. A method is proposed that takes advantage of large available memory in current computer hardware for extensive generation of precalculated data. Primary tracks of electrons are generated in the middle of homogeneous materials (water, air, bone, lung) and with energies between 0.2 and 18 MeV using the EGSnrc code. Secondary electrons are not transported, but their position, energy, charge, and direction are saved and used as a primary particle. Based on medium type and incident electron energy, a track is selected from the precalculated set. The performance of the method is tested in various homogeneous and heterogeneous configurations and the results were generally within 2% compared to EGSnrc but with a 40-60 times speed improvement. In a second stage the authors studied the obstacles for further increased speed-ups in voxel geometries by including ray-tracing and particle fluence information in the pregenerated track information. The latter method leads to speed increases of about a factor of 500 over EGSnrc for voxel-based geometries. In both approaches, no physical calculation is carried out during the runtime phase after the pregenerated data has been stored even in the presence of heterogeneities. The precalculated data are generated for each particular material and this improves the performance of the precalculated Monte Carlo code both in terms of accuracy and speed. Precalculated Monte Carlo codes are accurate, fast, and physics independent and therefore applicable to different radiation types including heavy-charged particles.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Monte Carlo simulations and radiation dosimetry measurements of peripherally applied HDR {sup 192}Ir breast brachytherapy D-shaped applicators

Conformal dose coverage for accelerated partial breast irradiation or radiotherapy boost can be obtained with AccuBoost registered D-shaped brachytherapy applicators using a flattened surface positioned near the patient. Three D-shaped applicators (D45/D53/D60) were dosimetrically characterized using Monte Carlo methods (MCNP5), air ionization chambers (Farmer and Markus), and radiochromic film (GafChromic EBT) in polystyrene and ICRU 44 breast tissue. HDR {sup 192}Ir source dwell times were either constant or optimized to improve skin dose uniformity. Scatter dose decreased as depth decreased. 10 mm beyond the applicator aperture, dose reductions of 90% and 51% were observed at depths of 0 and 30 mm, respectively. Similarly, planar dose uniformity improved as depth decreased and was also due to scatter and applicator geometry. Dose uniformity inside the applicator aperture was approximately 11% and 15% for all three applicators at the skin and 30 mm deep, respectively. Depth dose measurements in polystyrene using ion chamber and radiochromic film agreed with Monte Carlo results within 2%. Discrepancies between film and Monte Carlo dose profiles at 30 mm depth were within 1%.

61 RADIATION PROTECTION AND DOSIMETRY↗

Solid Water as phantom material for dosimetry of electron backscatter using low-energy electron beams: A Monte Carlo evaluation

This study evaluated the dosimetry of electron backscatter when Solid Water is used to substitute water as phantom in electron radiotherapy. Monte Carlo simulation (EGSnrc-based code) was employed to predict electron energy spectra and depth doses for the 0.5 and 1 cm of Solid Water and water slabs above 3 mm of lead (Pb) layers using electron beams with energies of 4 and 6 MeV. For comparison, Monte Carlo simulations were repeated with Pb layers taken out from the phantoms using the same experimental configuration. Analyses on electron energy spectra for the 4 and 6 MeV electron beams showed that deviations of electron energy distributions between the Solid Water and water phantom were more significant in the high-energy range (i.e., close to the maximal electron energy) than the lower range corresponding to the electron backscatter. These deviations of electron energy spectra varied with depth and were mainly due to the electron fluence or beam attenuation. Dosimetry results from Monte Carlo simulations showed that the Solid Water phantom had lower depth dose compared to water with the same experimental setup. For the 4 MeV electron beams with 0.5 cm of Solid Water, depth doses were 1.8%-3.9% and 2.3%-4.4% lower than those in water, with and without the Pb layer underneath, respectively. Thicker Solid Water of 1 cm resulted in different decreases in depth doses of 1.8%-4.6% (with Pb) and 2.3%-4.4% (without Pb) compared to water. For higher nominal electron beam energy of 6 MeV with 0.5 cm of Solid Water, depth doses decreased 1.7%-2.9% (with Pb) and 1.6%-2.1% (without Pb) compared to water. These decreases in depth doses changed to 1.7%-3.7% (with Pb) and 1.7%-3% (without Pb) when the thickness of Solid Water was increased to 1 cm. The dosimetry data in this study are useful in determining the correction factor when using Solid Water to substitute water for the electron backscatter measurement in electron radiotherapy.

61 RADIATION PROTECTION AND DOSIMETRY↗

EGSnrc-based Monte Carlo dosimetry of CSA1 and CSA2 {sup 137}Cs brachytherapy source models

Purpose: AAPM TG-56 recommends the use of a specific dosimetric dataset for each brachytherapy source model. In this study, a full dosimetric dataset for indigenously developed {sup 137}Cs source models, namely, the CSA1 and CSA2, in accordance with the AAPM TG-43U1 formalism is presented. The study includes calculation of dose-to-kerma ratio D/K in water around these sources including stainless steel encapsulated {sup 137}Cs sources such as RTR, 3M, and selectron/LDR {sup 137}Cs. Methods: The Monte Carlo-based EGSnrcMP code system is employed for modeling the sources in vacuum and in water. Calculations of air-kerma strength, S{sub K} for the investigated sources and collision kerma in water along the transverse axis of the RTR source are based on the FLURZnrc code. Simulations of water-kerma and dose in water for the CSA1, CSA2, RTR, 3M, and selectron/LDR {sup 137}Cs sources are carried out using the DOSRZnrc code. In DOSRZnrc calculations, water-kerma and dose are scored in a cylindrical water phantom having dimensions of 80 cm diameterx80 cm height. Results: The calculated dose-rate constants for the CSA1 and CSA2 sources are 0.945(1) and 1.023(1) cGy/(h U), respectively. The calculated value of S{sub K} per unit source activity, S{sub K}/A for the CSA1 and CSA2 sources is 7.393(7)x10{sup -8} cGy cm{sup 2}/(h Bq). The EGSnrcMP-based collision kerma rates for the RTR source along the transverse axis (0.25-10 cm) agree with the corresponding GEANT4-based published values within 0.5%. Anisotropy profiles of the CSA1 and CSA2 sources are significantly different from those of other sources. For the selectron/LDR single pellet {sup 137}Cs spherical source (modeled as a cylindrical pellet with dimensions similar to the seed selectron), the values of D/K at 1 and 1.25 mm from the capsule are 1.023(1) and 1.029(1), respectively. The value of D/K at 1 mm from the CSA1, CSA2, RTR, and 3M {sup 137}Cs source capsules (all sources have an external radius of 1.5 mm) is 1.017(1) and this ratio is applicable to axial positions z=0 to z={+-}L/2. This is in contrast to a published GEANT4-based Monte Carlo dosimetric study on RTR and 3M {sup 137}Cs sources wherein the authors have assumed that collision kerma is approximately equal to absorbed dose at 1 mm from the source capsules. Collision kerma is approximately equal to absorbed dose for distances {>=}2 mm from source capsules as opposed to {>=}1 mm reported in published studies. A detailed electron transport is necessary up to 2 mm from source capsules. Conclusions: The Monte Carlo-calculated dose-rate data for the CSA1 and CSA2 sources can be used as input data for treatment planning or to verify the calculations by radiotherapy treatment planning system.

61 RADIATION PROTECTION AND DOSIMETRY↗

Accelerating Monte Carlo simulations of photon transport in a voxelized geometry using a massively parallel graphics processing unit

Purpose: It is a known fact that Monte Carlo simulations of radiation transport are computationally intensive and may require long computing times. The authors introduce a new paradigm for the acceleration of Monte Carlo simulations: The use of a graphics processing unit (GPU) as the main computing device instead of a central processing unit (CPU). Methods: A GPU-based Monte Carlo code that simulates photon transport in a voxelized geometry with the accurate physics models from PENELOPE has been developed using the CUDA programming model (NVIDIA Corporation, Santa Clara, CA). Results: An outline of the new code and a sample x-ray imaging simulation with an anthropomorphic phantom are presented. A remarkable 27-fold speed up factor was obtained using a GPU compared to a single core CPU. Conclusions: The reported results show that GPUs are currently a good alternative to CPUs for the simulation of radiation transport. Since the performance of GPUs is currently increasing at a faster pace than that of CPUs, the advantages of GPU-based software are likely to be more pronounced in the future.

62 RADIOLOGY AND NUCLEAR MEDICINE↗