Search NASA⌕ Search

SEARCH · Search NASA

Results for “Applied physics”

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 559 records · Page 31

Primer on nuclear exchange models

Basic physics is applied to nuclear force exchange models between two nations. Ultimately, this scenario approach can be used to try and answer the age old question of 'how much is enough?' This work is based on Chapter 2 of Physics of Societal Issues: Calculations on National Security, Environment and Energy (Springer, 2007 and 2014)

98 NUCLEAR DISARMAMENT, SAFEGUARDS, AND PHYSICAL P↗

Field-induced spin level crossings within a quasi-$XY$ antiferromagnetic state in $Ba$ 2 $FeSi$ 2 $O$ 7

Here, we present a high-field study of the strongly anisotropic easy-plane square lattice $S$ = 2 quantum magnet $Ba$ 2 $FeSi$ 2 $O$ 7 . This compound is a rare high-spin antiferromagnetic system with very strong easy-plane anisotropy, such that the interplay between spin level crossings and antiferromagnetic order can be studied. We observe a magnetic field-induced spin level crossing occurring within an ordered state. This spin level crossing appears to preserve the magnetic symmetry while producing a nonmonotonic dependence of the order parameter magnitude. The resulting temperature–magnetic field phase diagram exhibits two dome-shaped regions of magnetic order overlapping around 30 T. The ground state of the lower-field dome is predominantly a linear combination of |$S^z$ = 0$\rangle$ and |$S^z$ = 1$\rangle$ states, while the ground state of the higher-field dome can be approximated by a linear combination of |$S^z$ = 1$\rangle$ and |$S^z$ = 2$\rangle$ states. At 30 T, where the spin levels cross, the magnetization exhibits a slanted plateau, the magnetocaloric effect shows a broad hump, and the electric polarization shows a weak slope change. We determined the detailed magnetic phase boundaries and the spin level crossings using measurements of magnetization, electric polarization, and the magnetocaloric effect in pulsed magnetic fields to 60 T. Furthermore, we calculate these properties using a mean-field theory based on direct products of SU(5) coherent states and find good agreement. Finally, we measure and calculate the magnetically induced electric polarization that reflects magnetic ordering and spin level crossings. This multiferroic behavior provides another avenue for detecting phase boundaries and symmetry changes.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Dynamics of Superfluid Helium in Low-Gravity

This report summarizes the work performed under a contract entitled 'Dynamics of Superfluid Helium in Low Gravity'. This project performed verification tests, over a wide range of accelerations of two Computational Fluid Dynamics (CFD) codes of which one incorporates the two-fluid model of superfluid helium (SFHe). Helium was first liquefied in 1908 and not until the 1930s were the properties of helium below 2.2 K observed sufficiently to realize that it did not obey the ordinary physical laws of physics as applied to ordinary liquids. The term superfluidity became associated with these unique observations. The low temperature of SFHe and it's temperature unifonrmity have made it a significant cryogenic coolant for use in space applications in astronomical observations with infrared sensors and in low temperature physics. Superfluid helium has been used in instruments such as the Shuttle Infrared Astronomy Telescope (IRT), the Infrared Astronomy Satellite (IRAS), the Cosmic Background Observatory (COBE), and the Infrared Satellite Observatory (ISO). It is also used in the Space Infrared Telescope (SIRTF), Relativity Mission Satellite formally called Gravity Probe-B (GP-B), and the Test of the Equivalence Principle (STEP) presently under development. For GP-B and STEP, the use of SFHE is used to cool Superconducting Quantum Interference Detectors (SQUIDS) among other parts of the instruments. The Superfluid Helium On-Orbit Transfer (SHOOT) experiment flown in the Shuttle studied the behavior of SFHE. This experiment attempted to get low-gravity slosh data, however, the main emphasis was to study the low-gravity transfer of SFHE from tank to tank. These instruments carried tanks of SFHE of a few hundred liters to 2500 liters. The capability of modeling the behavior of SFHE is important to spacecraft control engineers who must design systems that can overcome disturbances created by the movement of the fluid. In addition instruments such as GP-B and STEP are very sensitive to quasi-steady changes in the mass distribution of the liquid. The CFD codes were used to model the fluid's dynamic motion. Tests in one-g were performed with the main emphasis on being able to compute the actual damping of the fluid. A series of flights on the NASA Lewis reduced gravity DC-9 aircraft were performed with the Jet Propulsion Laboratory (JPL) Low Temperature Flight Facility and a superfluid Test Cell. The data at approximately 0.04g, lg and 2g were used to determine if correct fundamental frequencies can be predicted based on the acceleration field. Tests in zero gravity were performed to evaluate zero gravity motion.

Frank, David J.↗

Measurement of fifth- and sixth-order fluctuations of (net-)proton number in Au + Au collisions from phase II of the beam energy scan program at RHIC

We report high-statistics measurements of fifth- and sixth-order factorial cumulants and cumulant ratios of (net-)proton multiplicity distributions in Au+Au collisions at $\sqrt{s_{NN}}$ = 7.7–27GeV, using data from the STAR experiment collected during the Beam Energy Scan Phase II at RHIC. Protons and antiprotons are identified at midrapidity (|𝑦| < 0.5) with transverse momentum 0.4 < 𝑝 𝑇 < 2.0GeV/𝑐. The proton factorial cumulants 𝜅 4 , 𝜅 5 , and 𝜅 6 increase with order but exhibit no sign alternation within current uncertainties, offering no evidence for a two-component structure in the proton multiplicity distribution, as might be expected near a first-order phase transition. The cumulant ratios 𝐶 5 /𝐶 1 and 𝐶 6 /𝐶 2 fluctuate around zero in collisions at 0–40% centrality. The results are consistent with both the negative predictions from lattice QCD and the positive trends obtained from the ultrarelativistic quantum molecular dynamics (UrQMD) model. At $\sqrt{s_{NN}}$ ≳ 27GeV, the 𝐶 4 /𝐶 2 and 𝐶 5 /𝐶 1 results are compatible with predictions from lattice QCD, functional renormalization group (FRG), and hadron resonance gas (HRG) models, while UrQMD describes the data better at lower energies. Here, these measurements place constraints on baryon number fluctuations and offer valuable insights into the QCD phase structure.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Measurement of directed flow in Au + Au collisions at $\sqrt{s{NN}}$ = 19.6 and 27 GeV with the STAR event plane detector

In heavy-ion collision experiments, the global collectivity of final-state particles can be quantified by anisotropic flow coefficients (𝑣 𝑛 ) . The first-order flow coefficient, also referred to as the directed flow (𝑣 1 ), describes the collective sideward motion of produced particles and nuclear fragments in heavy-ion collisions. It carries information on the very early stage of the collision, especially at large pseudorapidity (𝜂), where it is believed to be generated during the nuclear passage time. Directed flow therefore probes the onset of bulk collective dynamics during thermalization, providing valuable experimental guidance to models of the pre-equilibrium stage. In 2018, the Event Plane Detector (EPD) was installed in STAR and used for the Beam Energy Scan phase-II (BES-II) data taking. The combination of EPD (2.1 < |𝜂| < 5.1) and high-statistics BES-II data enables us to extend the 𝑣 1 measurement to the forward and backward 𝜂 regions. In this paper, we present the measurement of 𝑣 1 over a wide 𝜂 range in Au + Au collisions at $\sqrt{s_{NN}}$ = 19.6 and 27 GeV using the STAR EPD. The results of the analysis at $\sqrt{s_{NN}}$ = 19.6 GeV exhibit excellent consistency with the previous PHOBOS measurement, while elevating the precision of the overall measurement. The increased precision of the measurement also revealed finer structures in heavy-ion collisions, including a potential observation of the first-order event-plane decorrelation. Multiple physics models were compared to the experimental results. Only a transport model and a three-fluid hybrid model can reproduce a sizable 𝑣 1 at large 𝜂 as was observed experimentally. The model comparison also indicates 𝑣 1 at large 𝜂 might be sensitive to the QGP phase transition.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Predicting departure from nucleate boiling with advanced data- and physics-driven approaches

The subcooled and low-quality saturated flow boiling critical heat flux (CHF) corresponding to the departure from nucleate boiling (DNB) is one of the major limiting factors in the design and operation of pressurized water reactors (PWRs). Due to lack of agreement on its physical mechanisms, development of an accurate DNB-type CHF model has been elusive. The various steady-state (SS) predictive tools that have been proposed can be grouped into two categories: (1) data-driven methods—namely best-fit empirical correlations and look-up tables that result in relatively good agreement with specific experimental datasets but often fail to extend beyond their ranges of validity, and (2) the physics-based mechanistic models that rely on reasonable yet limited understanding of the underlying physics and apply constitutive relations to close the conservation equations. On one hand, with recent advances in computational capabilities and optimization techniques, machine learning (ML)-based methods constitute an alternative and a more advanced, data-driven approach. With enough data, they can be particularly useful in engineering fields where the physical phenomena are complex and challenging to model. On the other hand, a state-of-the-art physics-driven approach would not only provide more accurate prediction of SS CHF, but it would also help better understand and model transient applications. The data-driven regression models use deep feed-forward neural networks (NNs) and random forests to cross-validate with 1,865 CHF test cases, covering a wide range of flow conditions and channel geometries. The best-estimate ML-based predictors compare favorably with the widely used look-up table for annulus and plate, and sensitivity analysis has confirmed their effectiveness. The key advantage of ML-based methods is their online extensibility of applicability domain. The proposed physics-driven approach combines key assumptions and parameters in the relatively well-accepted theories of liquid sublayer dryout and near-wall bubble crowding. A more realistic understanding of local mechanisms has been modeled. The new model has been optimized and validated against 1,439 tube data, and it shows considerably improved performance when compared to a recent mechanistic model, demonstrating unbiased close agreement with measurements over a wide range of operating conditions. Future work will extend the improved physics-driven model to non-tube geometry applications and will couple the model with ML via a hybrid approach. The mechanistic SS work will also serve to improve understanding and to model transient CHF scenarios.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Search for the chiral magnetic effect through beam energy dependence of charge separation using event shape selection

High-energy, heavy-ion collisions can create local domains of chirality-imbalanced quarks, reflecting the topological features of quantum chromodynamics. The chiral magnetic effect (CME) predicts an electric charge separation of quarks in such topological domains along the magnetic field ($\vec{B}$) generated by the passing of two high-Z nuclei. Here, we use a correlation observable Δ⁢𝛾$^{112}$ between charged meson pairs to detect the CME-induced charge separation and a novel event shape selection (ESS) method to mitigate the background effects related to elliptic flow (𝑣 2 ). The ESS method classifies events based on the emission pattern of final-state particles and determines Δ⁢𝛾$^{112}_{ESS}$ from the zero-flow limit. We reconstruct the $\vec{B}$ field direction from the spectator nucleons, which minimizes backgrounds unrelated to the collective motion of the system. In this work, we report the measurements of Δ⁢𝛾$^{112}$ and a background indicator Δ⁢𝛾$^{132}$ in Au+Au collisions from the Brookhaven National Laboratory Relativistic Heavy Ion Collider (RHIC) Beam Energy Scan phase II and at the top RHIC energy. After background suppression, Δ⁢𝛾$^{132}_{ESS}$ aligns with zero, and Δ⁢𝛾$^{112}_{ESS}$ is reduced to no more than 20% of Δ⁢𝛾$^{112}$. We observe a finite residual charge separation with 2.5⁢𝜎, 3⁢𝜎, and 3.2⁢𝜎 significance in the 20–50% centrality range of Au + Au collisions at 11.5, 14.6, and 19.6 GeV. The results at 17.3 and 27 GeV also show positive values but with a lower significance of 1.3⁢𝜎 and 1.1⁢𝜎, respectively. The corresponding Δ⁢𝛾$^{112}_{ESS}$ values at 7.7, 9.2, and 200 GeV are consistent with zero within uncertainties.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Correlator convolutional neural networks as an interpretable architecture for image-like quantum matter data

Image-like data from quantum systems promises to offer greater insight into the physics of correlated quantum matter. However, the traditional framework of condensed matter physics lacks principled approaches for analyzing such data. Machine learning models are a powerful theoretical tool for analyzing image-like data including many-body snapshots from quantum simulators. Recently, they have successfully distinguished between simulated snapshots that are indistinguishable from one and two point correlation functions. Thus far, the complexity of these models has inhibited new physical insights from such approaches. Here, we develop a set of nonlinearities for use in a neural network architecture that discovers features in the data which are directly interpretable in terms of physical observables. Applied to simulated snapshots produced by two candidate theories approximating the doped Fermi-Hubbard model, we uncover that the key distinguishing features are fourth-order spin-charge correlators. Our approach lends itself well to the construction of simple, versatile, end-to-end interpretable architectures, thus paving the way for new physical insights from machine learning studies of experimental and numerical data.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Measurement of medium-induced acoplanarity in central Au-Au and 𝑝⁢𝑝 collisions at $\sqrt{s_{NN}}$ = 200 GeV using direct-photon + jet and 𝜋 0 + jet correlations

The STAR Collaboration reports measurements of acoplanarity using semi-inclusive distributions of charged-particle jets recoiling from direct photon and 𝜋 0 triggers, in central Au–Au and 𝑝⁢𝑝 collisions at $\sqrt{s_{NN}}$ = 200 GeV. Significant medium-induced acoplanarity broadening is observed for large but not small recoil jet resolution parameter, corresponding to recoil jet yield enhancement up to a factor of ≈ 20 for trigger-recoil azimuthal separation far from 𝜋. This phenomenology is indicative of the response of the quark-gluon plasma to excitation, but not the scattering of jets off of its quasiparticles. As a result, the measurements are not well described by current theoretical models which incorporate jet quenching.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Energy dependence of polarized 𝛾⁢𝛾 → 𝑒 + ⁢ 𝑒− in peripheral Au+Au collisions at $\sqrt{s_{NN}}$ = 54.4 and 200 GeV with the STAR experiment at RHIC

We report the differential yields at mid-rapidity of the Breit-Wheeler process (𝛾⁢𝛾 → 𝑒 + ⁢𝑒 − ) in peripheral Au+Aucollisions at $\sqrt{s_{NN}}$ = 54.4 and 200 GeV with the STAR experiment at the Relativistic Heavy Ion Collider (RHIC), as a function of energy $\sqrt{s_{NN}}$, 𝑒 + ⁢𝑒 − transverse momentum 𝑝 T , 𝑝$^{2}_{T}$, invariant mass 𝑀 𝑒⁢𝑒 , and azimuthal angle. In the invariant mass range of 0.4 < 𝑀 𝑒⁢𝑒 < 2.6GeV/𝑐 2 at low transverse momentum (𝑝 T < 0.15GeV/𝑐), the yields increase while the pair √⟨𝑝$^{2}_{T}$⟩ decreases with increasing $\sqrt{s_{NN}}$, a feature that is correctly predicted by the QED calculation. Here, the energy dependencies of the measured quantities are sensitive to the nuclear form factor, infrared divergence and photon polarization. The data are compiled and used to extract the charge radius of the Au nucleus.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Advanced Power and Propulsion: Insuring Human Survival and Productivity in Deep Space Missions

Dr. Chang-Diaz gave an intriguing presentation of his research in advanced rocket propulsion and its relevance for planning and executing crewed deep space explorations. Though not necessarily exclusively Martian, his thrust looks critically at future Mars missions. Initially Dr. Chang-Diaz showed the time constraints of Mars missions due to orbital mechanics and our present chemically powered rocket technology. Since essentially all the energy required to place current generation spacecraft into a Martian trajectory must be expended in the early minutes of a flight, most of such a mission is spent in free-fall drift, captive to the gravitational forces among Earth, the Sun, and Mars. The simple physics of such chemically powered missions requires nearly a year in transit for each direction of a Mars mission. And the optimal orientations of Earth and Mars for rendezvous require further time on or around Mars to await return. These extensions of mission duration place any crew under a three-fold jeopardy: (1) physiological deconditioning (which in some aspects is still unknown and unpreventable), (2) psychological stress, and (3) ionizing radiation. This latter risk is due to exposure of crew members for extended time to the highly unpredictable and potentially lethal radiations of open space. Any gains in shortening mission duration would reap equivalent or greater benefits for these crew concerns. Dr. Chang-Diaz has applied his training and expertise (Ph.D. from Massachusetts Institute of Technology in applied plasma physics) toward development of continuous rocket propulsion which would offer great time advantages in travel, and also more launch options than are now available. He clearly explained the enormous gains from a relatively low thrust accelerative force applied essentially continuously versus the high, but short-lived propulsion of present chemical rockets. In fact, such spacecraft could be powered throughout the mission, accelerating to approximately the midpoint and decelerating during the latter half. This would not only provide some level of gravity (acceleration) throughout the mission but also allow very high velocities to be achieved, thus saving many months of travel time. In proposing the design of such a spacecraft propulsion system, Dr. Chang-Diaz was quick to acknowledge the need for a large power source, which undoubtedly must be nuclear fueled at the solar distances involved. He calls his system the Variable Specific Impulse Magnetoplasma Rocket (VASIMR). The other major ingredient is a mass (deuterium, which could also function as a radiation shield for crews) for energizing into the ultra hot, high velocity exhaust plasma. He foresees models now functional in the laboratory soon to be tested in space. In fact, some of these concepts have already been tried there. His optimism and determination would have operational rockets in the next decades.

Chang-Diaz, Franklin R.↗

Study of Vacuum Energy Physics for Breakthrough Propulsion

This report summarizes the accomplishments during a three year research project to investigate the use of surfaces, particularly in microelectromechanical systems (MEMS), to exploit quantum vacuum forces. During this project, we developed AFM instrumentation to repeatably measure Casimir forces in the nanoNewton range at 10 6 torr, designed an experiment to measure attractive and repulsive quantum vacuum forces, developed a QED based theory of Casimir forces that includes non-ideal material properties for rectangular cavities and for multilayer slabs, developed theoretical models for a variety of microdevices utilizing vacuum forces, applied vacuum physics to a gedanken spacecraft, and investigated a new material with a negative index of refraction.

Millis, Marc G.↗

Applying Bayesian inference and deterministic anisotropy to retrieve the molecular structure ∣Ψ(R)∣2 distribution from gas-phase diffraction experiments

Abstract Currently, our general approach to retrieving molecular structures from ultrafast gas-phase diffraction heavily relies on complex ab initio electronic or vibrational excited state simulations to make conclusive interpretations. Without such simulations, inverting this measurement for the structural probability distribution is typically intractable. This creates a so-called inverse problem. Here we address this inverse problem by developing a broadly applicable method that approximates the molecular frame structure ∣Ψ( R , t )∣ 2 distribution independent of these complex simulations. We retrieve the vibronic ground state ∣Ψ( R )∣ 2 for both simulated stretched NO 2 and measured N 2 O. From measured N 2 O, we observe 40 mÅ coordinate-space resolution from 3.75 Å −1 reciprocal space range and poor signal-to-noise, a 50X improvement over traditional Fourier transform methods. In simulated NO 2 diffraction experiments, typical to high signal-to-noise levels predict 100–1000X resolution improvements, down to 0.1 mÅ. By directly measuring the width of ∣Ψ( R )∣ 2 , we open ultrafast gas-phase diffraction capabilities to measurements beyond current analysis approaches. This method has the potential to effectively turn gas-phase ultrafast diffraction into a discovery-oriented technique to probe systems that are prohibitively difficult to simulate.

74 ATOMIC AND MOLECULAR PHYSICS↗

Diagrammatic approach to nonlinear optical response with application to Weyl semimetals

Nonlinear optical responses are a crucial probe of physical systems including periodic solids. In the absence of electron-electron interactions, they are calculable with standard perturbation theory starting from the band structure of Bloch electrons, but the resulting formulas are often large and unwieldy, involving many energy denominators from intermediate states. This work gives a Feynman diagram approach to calculating nonlinear responses. This diagrammatic method is a systematic way to perform perturbation theory, which often offers shorter derivations and also provides a natural interpretation of nonlinear responses in terms of physical processes. Applying this method to second-order responses concisely reproduces formulas for the second-order-harmonic shift current. We then apply this method to third-order responses and derive formulas for third-order-harmonic generation and self-focusing of light, which can be directly applied to tight-binding models. Third-order responses in the semiclasscial regime include a Berry curvature quadrupole term, whose importance is discussed including symmetry considerations and when the Berry curvature quadrupole becomes the leading contribution. In conclusion, the method is applied to compute third-order optical responses for a model Weyl semimetal, where we find a new topological contribution that diverges in a clean material, as well as resonances with a peculiar linear character.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

NASA Hybrid Reflectometer Project

Time-domain and frequency-domain reflectometry have been used for about forty years to locate opens and shorts in cables. Interpretation of reflectometry data is as much art as science. Is there information in the data that is being missed? Can the reflectometers be improved to allow us to detect and locate defects in cables that are not outright shorts or opens? The Hybrid Reflectometer Project was begun this year at NASA Ames Research Center, initially to model wire physics, simulating time-domain reflectometry (TDR) signals in those models and validating the models against actual TDR data taken on testbed cables. Theoretical models of reflectometry in wires will give us an understanding of the merits and limits of these techniques and will guide the application of a proposed hybrid reflectometer with the aim of enhancing reflectometer sensitivity to the point that wire defects can be detected. We will point out efforts by some other researchers to apply wire physics models to the problem of defect detection in wires and we will describe our own initial efforts to create wire physics models and report on testbed validation of the TDR simulations.

Lynch, Dana↗

Pseudotunnel Magnetoresistance in Twisted van der Waals Fe 3 GeTe 2 Homojunctions

Abstract Twistronics, a novel engineering approach involving the alignment of van der Waals (vdW) integrated two‐dimensional materials at specific angles, has recently attracted significant attention. Novel nontrivial phenomena have been demonstrated in twisted vdW junctions (the so‐called magic angle), such as unconventional superconductivity, topological phases, and magnetism. However, there have been only few reports on integrated vdW layers with large twist angles θ t , such as twisted interfacial Josephson junctions using high‐temperature superconductors. Herein, vdW homojunctions of the thin‐magnetic flakes, Fe 3 GeTe 2 (FGT), with large θ t ranging from 0° to 90°, without inserting any tunnel barriers are assembled. Nevertheless, these vdW homojunctions exhibit tunnel‐magnetoresistance (TMR) like behavior (pseudo‐TMR (PTMR) effect) with the ratios highly sensitive to the θ t values, revealing that the vdW gap at the junction interface between the twisted FGT layers behaves like a tunnel barrier and the θ t serves a control parameter for PTMR by drastically varying magnitudes of the lattice‐mismatch and the subsequent appearance of antiferromagnetic (AFM) spin alignment. First‐principles calculations considering vacuum gaps indicate strong dependence of TMR on the θ t driven by the sixfold screw rotational symmetry of bulk FGT. The present homojunctions hold promise as a platform for novel AFM spin‐dependent phenomena and spintronic applications.

Obata, Reiji [Faculty of Science and Engineering A↗

High Energy Electromagnetic Particle Transportation on the GPU

We present massively parallel high energy electromagnetic particle transportation through a finely segmented detector on a Graphics Processing Unit (GPU). Simulating events of energetic particle decay in a general-purpose high energy physics (HEP) detector requires intensive computing resources, due to the complexity of the geometry as well as physics processes applied to particles copiously produced by primary collisions and secondary interactions. The recent advent of hardware architectures of many-core or accelerated processors provides the variety of concurrent programming models applicable not only for the high performance parallel computing, but also for the conventional computing intensive application such as the HEP detector simulation. The components of our prototype are a transportation process under a non-uniform magnetic field, geometry navigation with a set of solid shapes and materials, electromagnetic physics processes for electrons and photons, and an interface to a framework that dispatches bundles of tracks in a highly vectorized manner optimizing for spatial locality and throughput. Core algorithms and methods are excerpted from the Geant4 toolkit, and are modified and optimized for the GPU application. Program kernels written in C/C++ are designed to be compatible with CUDA and OpenCL and with the aim to be generic enough for easy porting to future programming models and hardware architectures. To improve throughput by overlapping data transfers with kernel execution, multiple CUDA streams are used. Issues with floating point accuracy, random numbers generation, data structure, kernel divergences and register spills are also considered. Performance evaluation for the relative speedup compared to the corresponding sequential execution on CPU is presented as well.

Canal, P.↗