Scientific Data

Characterizing Lossless GPU Data Compression Across AMD CDNA and RDNA Architectures

(2027)

Data movement remains a major bottleneck in high-performance computing workflows, making GPU-accelerated compression an increasingly important optimization. While NVIDIA platforms benefit from mature compression libraries such as nvCOMP, the performance of practical lossless compression on AMD GPUs remains undercharacterized. This paper addresses that gap by porting LZ4, Snappy, and Cascaded from CUDA to HIP and evaluating them across four AMD GPUs spanning three CDNA generations (MI50, MI210, MI300X) and RDNA 3 (RX 7900 XT), using both synthetic datasets and seismic simulation data. Our study makes three contributions. First, we provide a cross-generational characterization of lossless GPU compression on recent AMD accelerators. Second, we present a functional CUDA-to-HIP port of three widely used algorithms, establishing an open baseline for future AMD-specific optimization. Third, we analyze how architectural differences shape compression and decompression behavior across algorithms and datasets. The results show that the MI300X achieves up to 11×$$11\times $$ higher decompression throughput than the MI50, while RDNA 3 is competitive for compression workloads characterized by irregular memory accesses. We also show that transfers dominate end-to-end execution time on all evaluated platforms, indicating that the main benefits of GPU compression are most likely to emerge in GPU-resident workflows. For realistic seismic floating-point data, lossless compression ratios remain modest (1.03-1.04×$$1.03-1.04\times $$), suggesting that error-bounded lossy compression is a promising direction for future work.

Correction: Application of geometric deep learning for tracking of hyperons in a straw tube detector

(2026)

Cover page of Regional Earthquake Ground Motion Simulations for Southern California With EQSIM: Insights From the 2008 Chino Hills, 2024 Highland Park, and 2021 Carson Earthquakes

Regional Earthquake Ground Motion Simulations for Southern California With EQSIM: Insights From the 2008 Chino Hills, 2024 Highland Park, and 2021 Carson Earthquakes

(2026)

This study presents physics‐based, 3D simulations using the EQSIM framework for several earthquakes in the Los Angeles region. The primary objective was to assess the ability of deterministic physics‐based ground motion simulations to reproduce the observed motions from historical events. The selected events included the 5.4 2008 Chino Hills, the 4.4 2024 Highland Park, and the 4.3 2021 Carson events. The simulated motions were evaluated by comparing the recorded and simulated seismograms, as well as the Fourier amplitude spectra, across multiple seismic stations. The SCEC 3D velocity model, CVM‐S4.26.M01, was used to represent the regional geology, and ground motion simulations were carried out with a resolution of up to 5 Hz. The results indicate that the simulated motions captured the recorded motions up to approximately 4 Hz. While careful iterations regarding source parameters and corner frequencies were required, and, for the case of the Highland Park event, some of the near‐source stations had relatively low accuracy, the present study established a positive step toward the utilization of physics‐based simulations in practical applications. The computational efficiencies exhibited by EQSIM, especially on GPU clusters, further supported this assertion, as wall‐clock times of simulations involving more than 10 billion grid points were as low as minutes. This permits ensemble simulations for a considered scenario event so that modeling uncertainties (e.g., source and geology) can be bracketed.

Cover page of Data Cards for Standardized Metadata Across DOE-Aligned Data Initiatives: Toward Transparent, Interoperable, and Governed Dataset Documentation

Data Cards for Standardized Metadata Across DOE-Aligned Data Initiatives: Toward Transparent, Interoperable, and Governed Dataset Documentation

(2026)

Combination of Measurements of CP Properties of Higgs Boson Interactions with Vector Bosons Using Proton-Proton Collisions at s=13 TeV with the ATLAS Detector

(2026)

A combination of measurements of the properties of Higgs boson interactions with electroweak gauge bosons is presented, using of proton-proton collisions at recorded by the ATLAS detector. Results from vector boson fusion , inclusive , and channels are combined. No evidence of violation is observed, and constrains on the -violating operators in the Standard Model effective field theory framework (SMEFT) are set in the Warsaw basis. The results from the combination improve by over 40% on previous individual limits on and, for the first time, simultaneous constraints on three coefficients , , and are set. These limits are the most stringent constraints to date on the relevant Wilson coefficients in the SMEFT framework with minimum model dependence.

Search for long-lived particles using displaced vertices of oppositely charged leptons in 140 fb − 1 of pp collisions at s = 13 TeV with the ATLAS detector

(2026)

A search is presented for long-lived particles decaying into an oppositely charged lepton pair, μ + μ − , e + e − , or e  ±  μ ∓, that form a vertex within the inner tracking system of the ATLAS detector at the Large Hadron Collider, displaced from the primary proton–proton interaction region. The analysis uses the 140 fb − 1 of Run-2 data collected at s = 13 TeV by the ATLAS experiment in 2015–2018. The results of the analysis are interpreted in the context of three benchmark models covering masses from 0.1 to 2.2 TeV and a range of mean proper lifetimes times the speed of light from 1 to 10 000 mm. The first model is a generic Z′ boson pair-produced by a new heavy scalar, with the Z′ decaying into lepton pairs. The remaining two models are R-parity violating supersymmetric models in which the lightest neutralino χ ˜ 1 0 decays into ℓ + ℓ ′ − ν ( ℓ , ℓ ′ = e , μ). The models differ by the mode of production of the χ ˜ 1 0 , which can be produced via the decay of pairs of gluinos or of pairs of charginos and neutralinos ( χ ˜ 1 ± χ ˜ 1 0 , χ ˜ 1 ± χ ˜ 2 0 , or χ ˜ 2 0 χ ˜ 1 0 ). Although each benchmark sample includes pair-produced LLPs, only a single vertex is required to be reconstructed. No dilepton displaced vertex candidate is observed and the results are presented as upper limits on the production cross-sections. This analysis sets leading limits on the production cross-sections for multiple models, including parameter space that has never been directly probed.

Search for massive, long-lived particles in events with displaced vertices and displaced muons in pp collisions at s = 13.6 TeV with the ATLAS experiment

(2026)

A search is presented for massive long-lived particles in events featuring at least one displaced vertex and at least one displaced muon, using proton–proton collision data collected by the ATLAS detector at the Large Hadron Collider from 2022 to 2024 at a centre-of-mass energy of 13.6 TeV. The data sample corresponds to an integrated luminosity of 164 fb − 1 . The analysis targets scenarios in which long-lived particles decay inside the ATLAS inner detector, resulting in a topology of at least one massive, displaced vertex (DV) with multiple associated tracks, and at least one muon with a large transverse impact parameter relative to the primary interaction point. The muon is not required to be associated with the DV. Two signal regions are defined by the transverse distance of the reconstructed DV from the interaction point. Background contributions are estimated by using fully data-driven techniques. No significant excess above the expected background is observed. Upper limits at 95% confidence level are set on the visible cross-section and on the production cross-sections of several benchmark models of R-parity-violating supersymmetry.

Search for emerging jets in pp collisions at s=13TeV with the ATLAS experiment

(2026)

A search is presented for emerging jets using 140fb-1$$140~\textrm{fb}^{-1}$$ of proton–proton collision data at s=13TeV$$\sqrt{s} = 13~\textrm{TeV}$$, collected by the ATLAS experiment between 2015 and 2018. The search looks for the existence of a dark sector with symmetries similar to those in quantum chromodynamics. This dark sector is populated with dark quarks, which undergo showering similar to quarks in the Standard Model, leading to a high multiplicity of long-lived dark hadrons within a dark jet. These dark hadrons subsequently decay to Standard Model particles via a new heavy scalar mediating particle ϕ$$\phi $$. This results in jets which contain multiple displaced vertices, known as emerging jets. This analysis targets four-jet topologies, with two emerging jets and two Standard Model jets, resulting from the decay of pair-produced scalar mediators. No significant excess above the Standard Model background is observed. For dark pion proper decay lengths of 20mm$$20~\textrm{mm}$$, mediator masses are excluded between 1 and 2TeV$$2~\textrm{TeV}$$ assuming a dark pion mass of 20GeV$$20~\textrm{GeV}$$.

Search for electroweak tt¯Wj production in multileptonic final states at s=13 TeV with the ATLAS detector and bounds on effective field theory operators

(2026)

A search is presented for the electroweak production of a top-quark pair in association with a boson and at least one additional jet known as the process. This process has embedded within it a -scattering vertex, which is probed directly for the first time. The collision data were collected with the ATLAS detector during Run 2 of the LHC and correspond to an integrated luminosity of at . The search uses same-charge pairs of electrons and muons together with jets, of which at least one is -tagged. The properties of the most forward jet relative to the rest of the event are used to discriminate the electroweak production process from its strong production counterpart. A measured (expected) 95% CL upper limit on the cross section is set at (230 fb), to be compared with the expected Standard Model (SM) cross section of 47.7 fb. Limits are set on the SM effective field theory (EFT) operators and , which modify the electroweak couplings of the top quark through contributions to the -scattering vertex. The interpretation acts as a case study to emphasize the importance of energy-dependent sensitivity, multiprocess, and multioperator EFT contributions.

Quantifying Epistemic Uncertainty in Diffusion Models

(2026)

To ensure high quality outputs, it is important to quantify the epistemic uncertainty of diffusion models. Existing methods are often unreliable because they mix epistemic and aleatoric uncertainty. We introduce a method based on Fisher information that explicitly isolates epistemic variance, producing more reliable plausibility scores for generated data. To make this approach scalable, we propose FLARE (Fisher-Laplace Randomized Estimator), which approximates the Fisher information using a uniformly random subset of model parameters. Empirically, FLARE improves uncertainty estimation in synthetic time-series generation tasks, achieving more accurate and reliable filtering than other methods. Theoretically, we bound the convergence rate of our randomized approximation and provide analytic and empirical evidence that last-layer Laplace approximations are insufficient for this task.

Computing Sciences

Scientific Data