Engineering Papers⌕ Search

Engineering topics

Cai, Tejin

Publications and source records attributed to Cai, Tejin.

6 records
All sourcesNASA NTRSDOE OSTI
DOE OSTI2023

Accelerating Machine Learning Inference with GPUs in ProtoDUNE Data Processing

Abstract We study the performance of a cloud-based GPU-accelerated inference server to speed up event reconstruction in neutrino data batch jobs. Using detector data from the ProtoDUNE experiment and employing the standard DUNE grid job submission tools, we attempt to reprocess the data by running several thousand concurrent grid jobs, a rate we expect to be typical of current and future neutrino physics experiments. We process most of the dataset with the GPU version of our processing algorithm and the remainder with the CPU version for timing comparisons. We find that a 100-GPU cloud-based server is able to easily meet the processing demand, and that using the GPU version of the event processing algorithm is two times faster than processing these data with the CPU version when comparing to the newest CPUs in our sample. The amount of data transferred to the inference server during the GPU runs can overwhelm even the highest-bandwidth network switches, however, unless care is taken to observe network facility limits or otherwise distribute the jobs to multiple sites. We discuss the lessons learned from this processing campaign and several avenues for future improvements.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2021

Measurements of Nuclear Effects and the $\bar \nu_\mu +H \to \mu^+ + n$ Cross Section in MINERνA with Neutron Tagging

MINERνA, or Main INjector ExpeRiment for ν-A, at Fermilab, is an experimentdedicated to the study of neutrino-nucleus interactions in the GeV regime. Itsgoal is to illustrate the interplay between hadronic and nuclear physics and measure intranuclear dynamics crucial for the present and future neutrino oscillationmeasurements. We first measure a set of variables sensitive to how Monte Carlo(MC) simulations of neutrino-nucleus interactions implement binding energy andthen move on to measure the antineutrino CCQE cross section on the hydrogentargets in MINERνA’s CH detector. We have developed a method to preferentially select events on the hydrogen by comparing outgoing neutrons’ directions totheoretical neutron directions assuming two-body interactions. We measured thecross section, extracted the axial form factor, and performed a Z-expansion fit. Weobserve larger values in the axial form factor at high Q2than current best fits. Finally, we show a preliminary selection of events with both protons and neutrons toinvestigate nuclear processes responsible for producing these final states.

Cai, Tejin↗
DOE OSTI2020

Comparison of optical potential for nucleons and $\varDelta $ resonances

Precise modeling of neutrino interactions on nuclear targets is essential for neutrino oscillations experiments. The modeling of the energy of final state particles in quasielastic (QE) scattering and resonance production on bound nucleons requires knowledge of both the removal energy of the initial state bound nucleon as well as the average Coulomb and nuclear optical potentials for final state leptons and hadrons. We extract the average values of the real part of the nuclear optical potential for final state nucleons ( $U_{opt}^{QE}$ U opt QE ) as a function of the nucleon kinetic energy from inclusive electron scattering data on nuclear targets ( $_\mathbf{6 }^\mathbf{12 }{} \mathbf{C} $ 6 12 C + $_\mathbf{8 }^\mathbf{16 }{} \mathbf{O} $ 8 16 O , $_\mathbf{20 }^\mathbf{40 }{} \mathbf{Ca} $ 20 40 Ca + $_\mathbf{18 }^\mathbf{40 }{} \mathbf{Ar} $ 18 40 Ar , $_\mathbf{3 }^\mathbf{6 }{} \mathbf{Li} $ 3 6 Li , $_\mathbf{18 }^\mathbf{27 }{} \mathbf{Al} $ 18 27 Al , $_\mathbf{26 }^\mathbf{56 }{} \mathbf{Fe} $ 26 56 Fe , $_\mathbf{82 }^\mathbf{208 }{} \mathbf{Pb} $ 82 208 Pb ) in the QE region and compare to calculations. We also extract values of the average of the real part of the nuclear optical potential for a $\varDelta (1232)$ Δ ( 1232 ) resonance in the final state ( $U^\varDelta _{opt}$ U opt Δ ) within the impulse approximation. We find that $U^\varDelta _{opt}$ U opt Δ is more negative than $U_{opt}^{QE}$ U opt QE with $U^\varDelta _{opt}\approx $ U opt Δ ≈ 1.5 $U_{opt}^{QE}$ U opt QE for $_\mathbf{6 }^\mathbf{12 }{} \mathbf{C} $ 6 12 C .

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗