SEARCH · Search NASA

Results for “GPU”

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.

SourceResearch connectionFrom yearThrough year

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

At least 145 records · Page 8

All sources NASA NTRS DOE OSTI

DOE OSTI2025

FPGA-accelerated SpeckleNN with SNL for real-time X-ray single-particle imaging

We present the implementation of a specialized version of our previously published unified embedding model, SpeckleNN, for real-time speckle pattern classification in X-ray Single-Particle Imaging (SPI), using the SLAC Neural Network Library (SNL) on an FPGA platform. This hardware realization transitions SpeckleNN from a prototypic model into a practical edge solution, optimized for running inference near the detector in high-throughput X-ray free-electron laser (XFEL) facilities, such as those found at the Linac Coherent Light Source (LCLS). To address the resource constraints inherent in FPGAs, we developed a more specialized version of SpeckleNN. The original model, which was designed for broader classification across multiple biological samples, comprised ~5.6 million parameters. The new implementation, while reducing the parameter count to 64.6K (a 98.8% reduction), focuses on maintaining the model's essential functionality for real-time operation, achieving an accuracy of 90%. Furthermore, we compressed the latent space from 128 to 50 dimensions. This implementation was demonstrated on the KCU1500 FPGA board, utilizing 71% of available DSPs, 75% of LUTs, and 48% of FFs, with an average power consumption of 9.4W according to the Vivado post-implementation report. The FPGA performed inference on a single image with a latency of 45.015 microseconds at a 200 MHz clock rate. In comparison, running the same inference on an NVIDIA A100 GPU resulted in an average power consumption of ~73W and an image processing latency of around 400 microseconds. Our FPGA-accelerated version of SpeckleNN demonstrated significant improvements, achieving an 8.9 × speedup and a 7.8 × reduction in power consumption compared to the GPU implementation. Key advancements include model specialization and dynamic weight loading through SNL, which eliminates the need for time-consuming FPGA design re-synthesis, allowing fast and continuous deployment of models (re)trained online. These innovations enable real-time adaptive classification and efficient vetoing of speckle patterns, making SpeckleNN more suited for deployment in XFEL facilities. This implementation has the potential to significantly accelerate SPI experiments and enhance adaptability to evolving experimental conditions.

47 OTHER INSTRUMENTATION↗

DOE OSTI2026

Benchmarking of massively parallel phase-field codes for directional solidification

We present a detailed benchmark comparing two state-of-the-art phase-field implementations for simulating alloy solidification under experimentally relevant conditions. The study investigates the directional solidification of Al-3wt%Cu under high-velocity solidification conditions and SCN-0.46wt% camphor under microgravity conditions from National Aeronautics and Space Administration (NASA) DECLIC-DSI-R experiments. Both codes, one employing finite-difference discretization with uniform mesh and GPU-acceleration (GPU-PF) and the other one employing finite-element discretization with adaptive-mesh and CPU-parallelization (PRISMS-PF), solve the same quantitative phase-field formulation that incorporates an anti-trapping current for the solidification of dilute alloys. We evaluate the predictions of each code for dendritic morphology, primary spacing, and tip dynamics in both 2D and 3D, as well as their numerical convergence and computational performance. While existing benchmark problems have primarily focused on simplified or small-scale simulations, they do not reflect the computational and modeling challenges posed by employing experimentally relevant time and length scales. Our results provide a practical framework for assessing phase-field code performance as well as validating and facilitating their application in integrated computational materials engineering (ICME) workflows that require integration with realistic experimental data.

36 MATERIALS SCIENCE↗

DOE OSTI2026

Comparison of automated chemical-guided segmentation and human annotation of soil organic matter in X-ray microcomputed tomography imaging in contrasted soil types

Soil organic matter (OM) formation and persistence is strongly influenced by the spatial distribution of organic substrates and microscale soil heterogeneity by dictating OM accessibility to microorganisms. However, traditional size and/or density fractionation techniques disrupt aggregate architecture, eliminating spatial information needed to fully understand intra-aggregate OM distribution. To quantify three-dimensional OM spatial distribution and automate segmentation in X-ray microcomputed tomography (µCT) imaging without human annotation bias, we developed an iodine gas vapor (I2) based staining workflow that eliminates labor-intensive manual annotation while maintaining segmentation accuracy, using aggregates from four taxonomically diverse soils (Xerofluvent, Haploxeroll Sphagnofibrist, Palehumult) with an 8-fold range of soil organic carbon. Human annotation of 10 µCT slices by the experienced and inexperienced annotators resulted in variations up to 3% in the Dice similarity coefficient (DSC), reflecting a degree of inherent subjectivity of manual labeling. Such inconsistencies are expected to compound as the number of manually annotated slices increases. Dual-energy µCT imaging at 33.1 keV (below the iodine (I) K-edge) and 33.2 keV (above the I K-edge) was used to resolve aggregate microstructure following I2 staining. The automated image subtraction pipeline identified OM regions by the I Kedge induced brightness increases, achieving DSC values of 0.58–0.83 relative to an experienced annotator. Sensitivity analyses revealed that the reconstruction alpha value—optimized via the open-source tool TomocuPy—and the 3D registration slice count were the primary determinants of accuracy, providing a novel benchmark for dual-energy soil imaging. The pipeline without GPU acceleration achieved 9.6 to 43.2 times faster than manual annotation. Using GPU-accelerated image post-processing and affine transformation matrices, the pipeline successfully segmented OM elements for large-scale datasets (3232×3232 pixel, 2048 slices) within ~5200 s from raw file acquisition to segmented output. The high-throughput approach enables the quantification of OM spatial distribution across diverse and heterogeneous soil.

Soil microbial biomass↗

DOE OSTI2025

Multi-material ALE remap with interface sharpening using high-order matrix-free finite element methods

The arbitrary Lagrangian-Eulerian (ALE) technique involves remapping field quantities from a Lagrangian mesh to an optimized mesh in a conservative, accurate and bounds-preserving manner. For methods based on arbitrary order finite elements, as described in a reference, material volume fractions are advected in pseudo-time using flux-corrected transport (FCT) without any form of interface reconstruction. In practice, this can lead to excessive propagation of small volume fractions throughout the domain. In addition, this method requires assembly of a global advection matrix to compute the bounds-preserving low-order FCT solution. In this work, we introduce a new approach for ALE remap using a high-order matrix-free technique which incorporates a flux modification to sharpen material interfaces in a conservative manner. Our approach begins with computing a bounds-preserving low-order solution to the ALE remap equations at the element level. We then compute a sharp interface solution (not guaranteed to be bounds-preserving) which comes from solving an augmented version of the ALE remap equations with a conservative flux modification which acts to sharpen material volume fractions based on their gradients and transport directions. Using the sharp interface solution, we make global corrections to the bounds-preserving solution while maintaining preservation of bounds. By blending with the sharpened solution at the global level we are able to globally conserve mass without hindering the remap pseudo-time step. This new interface-aware ALE remap method is based entirely on partial assembly techniques where globally assembled matrix operators are no longer needed, resulting in a globally matrix-free FCT method for multi-material, multi-field ALE remap with high performance on GPU architectures. We present results of our new remap method on 1D, 2D and 3D benchmarks and describe the algorithmic tailoring for GPU architectures that was developed.

Vargas, Arturo [Lawrence Livermore National Labora↗

DOE OSTI2026

Gas permeation properties of amorphous zeolitic imidazolate framework membranes made by atomic/molecular layer deposition

Amorphous metal-organic framework (MOF) membranes are desirable because they may retain some of the molecular sieving properties of their crystalline counterparts while being free of grain boundary defects, which often hinder the consistent achievement of high membrane performance. However, current methods, like melting and compression, for fabricating amorphous MOF membranes involve multi-step processes that require the formation of a crystalline membrane first, that is then amorphized, and therefore, could be challenging to scale. Here, we utilize atomic/molecular layer deposition (ALD/MLD) of diethylzinc (DEZ) and 2-methylimidazole (2mIm) to directly synthesize ultrathin amorphous zeolitic imidazolate framework (aZIF) deposits on γ-alumina-coated α-alumina supports. As the number of ALD/MLD cycles increased from 10 to 300, gas permeances decreased while ideal selectivities increased. Mixture separation factors for C3H6/C3H8, CO2/N2, and H2/C3H8 as high as 4, 37, and 194, respectively, were obtained. At 200 °C and 2.5 bar equimolar feed of H2 and C3H8, a H2/C3H8 mixture separation factor of 185 is obtained with an H2 permeance of ca. 4.15x10-8 mol/m2-s-Pa (124 GPU). Additionally, the membrane achieves a CO2/N2 mixture separation factor of 37 at 25 °C and 1 bar with a CO2 permeance of ca. 3.96x10-8 mol/m2-s-Pa (118 GPU). Considering the vast array of compositionally distinct aZIFs that can be potentially deposited by this approach, an enormously large parameter space for membrane design is emerging to be explored.

36 MATERIALS SCIENCE↗

DOE OSTI2025

Small-scale properties from exascale computations of turbulence on a $\mathbf{32\,768^3}$ periodic cube

To study the physics of small-scale properties of homogeneous isotropic turbulence at increasingly high Reynolds numbers, direct numerical simulation results have been obtained for forced isotropic turbulence at Taylor-scale Reynolds number R λ = 2500 on a 32 768 3 three-dimensional periodic domain using a GPU pseudo-spectral code on a 1.1 exaflop GPU supercomputer (Frontier). These simulations employ the multi-resolution independent simulation (MRIS) technique (Yeung & Ravikumar 2020, Phys. Rev. Fluids, vol. 5, 110517) where ensemble averaging is performed over multiple short segments initiated from velocity fields at modest resolution, and subsequently taken to higher resolution in both space and time. Reynolds numbers are increased by reducing the viscosity with the large-scale forcing parameters unchanged. Although MRIS segments at the highest resolution for each Reynolds number last for only a few Kolmogorov time scales, small-scale physics in the dissipation range is well captured – for instance, in the probability density functions and higher moments of the dissipation rate and enstrophy density, which appear to show monotonic trends persisting well beyond the Reynolds number range in prior works in the literature. Attainment of range of length and time scales consistent with classical scaling also reinforces the potential utility of the present high-resolution data for studies of short-time-scale turbulence physics at high Reynolds numbers where full-length simulations spanning many large-eddy time scales are still not accessible. A single snapshot of the 32 768 3 data is publicly available for further analyses via the Johns Hopkins Turbulence Database.

intermittency↗

DOE OSTI2025

Packaging HEP Heterogeneous Mini-apps for Portable Benchmarking and Facility Evaluation on Modern HPCs

High Energy Physics (HEP) experiments are making increasing use of GPUs and GPU dominated High Performance Computer facilities. Both the software and hardware of these systems are rapidly evolving, creating challenges for experiments to make informed decisions as to where they wish to devote resources. In its first phase, the High Energy Physics Center for Computational Excellence (HEP-CCE) produced portable versions of a number of heterogeneous HEP mini-apps, such as p2r, FastCaloSim, Patatrack and the WireCell Toolkit, that exercise a broad range of GPU characteristics, enabling cross platform and facility benchmarking and evaluation. However, these miniapps still require a significant amount of manual intervention to deploy on a new facility. We present our work in developing turn-key deployments of these mini-apps, where by means of containerization and automated configuration and build techniques such as Spack, we are able to quickly test new hardware, software, environments and entire facilities with minimal user intervention, and then track performance metrics over time.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

DOE OSTI2026

Application of Portable Parallelization Strategies for GPUs on track reconstruction kernels

Utilizing the computational power of GPUs is one of the key ingredients to meet the computing challenges presented to the next generation of High-Energy Physics (HEP) experiments. Unlike CPUs, developing software for GPUs often involves using architecturespecific programming languages promoted by the GPU vendors and hence limits the platform that the code can run on. Various portability solutions have been developed to achieve portable, performant software across different GPU vendors. Given the rapid evolution of these portability solutions, an early adoption of them in simple HEP testbed applications will help us understand the strengths and weaknesses of respective approaches.We apply several portability solutions, including Alpaka, Kokkos, SYCL and std::execution::par, on kernels for track propagation extracted from the mkFit project. We report on the development experience of the same application with different portability solutions, as well as their performance on GPUs, measured as the throughput of the kernels, from different manufacturers such as NVIDIA, AMD and Intel.

Kwok, Martin [Fermilab] (ORCID:0000000286936146)↗

DOE OSTI2026

Developments in Performance and Portability of BlockGen

For more than a decade Monte Carlo event generators with the current matrix element algorithms have been used for generating hard scattering events on CPU platforms, with excellent flexibility and good efficiency. While the HL-LHC is approaching and precision requirements are becoming more demanding, many studies have been made to solve the bottleneck in the current Monte Carlo event generator tool chains. The novel BlockGen family of fast matrix element algorithms shown in this report, is one of the new developments that are more suitable for GPU acceleration. We report the development experience of porting BlockGen using Kokkos. Moreover, we discuss the performance of the Kokkos version in comparison with the dedicated GPU version in CUDA.

Bothmann, E. [Gottingen U.]↗

DOE OSTI2026

pop-cosmos : redshifts and physical properties of KiDS-1000 galaxies

ABSTRACT Principled Bayesian inference of galaxy properties has not previously been performed for wide-area weak-lensing surveys with millions of sources. We address this gap by applying the pop-cosmos generative model to perform spectral energy distribution (SED) fitting for 4 million KiDS (Kilo-Degree Survey)-1000 galaxies. Calibrated on deep COSMOS2020 photometric data, pop-cosmos specifies a physically motivated prior over the galaxy population up to $z \simeq 6$ in stellar population synthesis (SPS) parameter space. Using the Speculator SPS emulator with GPU (graphics processing unit)-accelerated Markov Chain Monte Carlo sampling, we perform full posterior inference at 8.2 GPU seconds per galaxy, obtaining joint constraints on galaxy redshifts and physical properties. We validate photometric redshifts against $\sim \!185\,\!000$ KiDS galaxies cross-matched to Dark Energy Spectroscopic Instrument Data Release 1 spectroscopic samples, achieving low bias ($2\times 10^{-3}$), scatter ($\sigma _{\mathrm{MAD}}=0.03$), and outlier fraction (3.2 per cent) for the Bright Galaxy Survey, with comparable performance (bias $3\times 10^{-2}$, $\sigma _{\mathrm{MAD}}=0.05$, 1.0 per cent outliers) for luminous red galaxies (LRGs). Within the LRG sample, we identify massive, dusty, star-forming contaminants at $z \simeq 0.4$ satisfying standard colour selections for quenched populations. We infer trends in stellar mass, star formation, metallicity, and dust across five tomographic redshift bins consistent with established scaling relations. Using specific star formation rate constraints, we identify $\sim$7 per cent of KiDS-1000 galaxies as quenched, versus 37 per cent implied by conservative colour cuts. This enables the construction of weak-lensing samples defined by physical properties while mitigating intrinsic alignment systematics and preserving statistical power. Our analysis validates pop-cosmos out of sample, establishing it as a scalable approach for galaxy evolution and cosmological analyses with photometric surveys.

Halder, Anik [Institute of Astronomy and Kavli Ins↗

DOE OSTI2025

A Parallel Alternative for Energy-Efficient Neural Network Training and Inferencing

Energy efficiency of training and inferencing with large neural network models is a critical challenge facing the future of sustainable large-scale machine learning workloads. This paper introduces an alternative strategy, called phantom parallelism, to minimize the net energy consumption of traditional tensor (model) parallelism, the most energy-inefficient component of large neural network training. The approach is presented in the context of feed-forward network architectures as a preliminary, but comprehensive, proof-of-principle study of the proposed methodology. We derive new forward and backward propagation operators for phantom parallelism, implement them as custom autograd operations within an end-to-end phantom parallel training pipeline and compare its parallel performance and energy-efficiency against those of conventional tensor parallel training pipelines. Formal analyses that predict lower bandwidth and FLOP counts are presented with supporting empirical results on up to 256 GPUs that corroborate these gains. Experiments are shown to deliver ∼50% reduction in the energy consumed to train FFNs using the proposed phantom parallel approach when compared with conventional tensor parallel methods. Additionally, the proposed approach is shown to train smaller phantom models to the same model loss on smaller GPU counts as larger tensor parallel models on larger GPU counts offering the possibility for even greater energy savings.

Seal, Sudip [ORNL] (ORCID:0000000332330656)↗

DOE OSTI2024

Picasso: Memory-Efficient Graph Coloring Using Palettes With Applications in Quantum Computing

A coloring of a graph is an assignment of colors to vertices such that no two neighboring vertices have the same color. The need for memory-efficient coloring algorithms is motivated by their application in computing clique partitions of graphs arising in quantum computations where the objective is to map a large set of Pauli strings into a compact set of unitaries. We present Picasso, a randomized memory-efficient iterative parallel graph coloring algorithm with theoretical sublinear space guarantees under practical assumptions. The parameters of our algorithm provide a trade-off between coloring quality and resource consumption. To assist the user, we also propose a machine learning model to predict the coloring algorithm’s parameters considering these trade-offs. We provide a sequential and a parallel implementation of the proposed algorithm. We perform an experimental evaluation on a 64-core AMD CPU equipped with 512 GB of memory and an Nvidia A100 GPU with 40GB of memory. For a small dataset where existing coloring algorithms can be executed within the 512 GB memory budget, we show up to 68× memory savings. On massive datasets we demonstrate that GPU-accelerated Picasso can process inputs with 49.5× more Pauli strings (vertex set in our graph) and 2,478× more edges than state-of-the-art parallel approaches.

artificial intelligence, quantum computing↗

DOE OSTI2025

Design-to-Deployment Continuum Platform for Microscopes and Computing Ecosystems

Science ecosystems with networked computing systems and physical instruments are increasingly being deployed with a goal to achieve the productivity promised by AI-supported remote automation. In support of these efforts, the virtual infrastructure twins (VITs) have been successfully utilized to develop the orchestration codes for these ecosystems without requiring physical access to expensive instruments, such as electron microscopes. Currently, the utility of such a VIT is severely limited by the computing capacity and capability of the computing system used as its host. Furthermore, codes developed on the VIT typically need to be transferred and refactored for production use, particularly, on high-performance systems with accelerators. In response, we develop a design-to-deployment continuum platform wherein a VIT runs natively on the ecosystem's own computing system, and thereby facilitates the continual in-situ testing and transition of codes for production use. Here, we describe the development and testing of software for remote microscope steering and GPU-based image reconstruction using this platform on a multi-GPU computing system networked to Nion microscopes. We demonstrate a continual transition of steering and reconstruction codes developed under VIT platform to production ecosystem deployment.

Al-Najjar, Anees [Oak Ridge National Laboratory (O↗

DOE OSTI2025

FedEFsz: Fair Cross-Silo Federated Learning System With Error-Bounded Lossy Compression

Cross-Silo federated learning systems have been identified as an efficient approach to scaling DNN training across geographically-distributed data silos to preserve the privacy of the training data. Communication efficiency and fairness are two major issues that need to be both satisfied when federated learning systems are deployed in practice. Simultaneously guaranteeing both of them, however, is exceptionally difficult because simply combining communication reduction and fairness optimization approaches often causes non-converged training or drastic accuracy degradation. Here, to bridge this gap, we propose FedEFsz. On the one hand, it integrates the state-of-the-art error-bounded lossy compressor SZ3 into cross-silo federated learning systems to significantly reduce communication traffic during the training. On the other hand, it achieves a high fairness (i.e., rather consistent model accuracy and performance across different clients) through a carefully designed heuristic algorithm that can tune the error-bound of SZ3 for different clients during the training. Extensive experimental results based on a GPU cluster with 65 GPU cards show that FedEFsz improves the fairness across different benchmarks by up to 60.88% and meanwhile reduces the communication traffic by up to 315×.

Cross-Silo Federated Learning Systems↗

DOE OSTI2026

Memory-Aware External Facelist Calculation: A Data-Parallel Atomic Hash Counting Approach

Unstructured volumetric meshes serve as fundamental data representations in various scientific simulations and analyses. They play a crucial role in representing complex computational domains and are essential for important numerical techniques, such as finite element analysis. Whenever such a mesh is read from a file, streamed in-situ, or generated by algorithms, scientific visualization libraries rely on calculating the external surface of a geometry, named “external facelist”, to produce a polygonal mesh for rendering. Consequently, external facelist calculation has become one of the most widely used algorithms in the scientific visualization domain, necessitating optimal performance. In this paper, we explore relevant work on external facelist calculation algorithms in two common visualization libraries, VTK and Viskores, assess their performance and memory constraints, and introduce a novel memory-aware external facelist calculation algorithm employing an atomic hash counting approach. This algorithm fully leverages Viskores' data-parallel primitive operations, facilitating its execution across diverse many-core architectures. Our algorithm features the lowest memory footprint on the GPU and the second-lowest on the CPU among all evaluated methods, and it also delivers the fastest performance on both CPU and GPU. It has been made available under an open-source license in the VTK and Viskores visualization systems.

Tsalikis, Spiros [Kitware] (ORCID:0000000151137195↗

DOE OSTI2025

Insights from Optimizing HPL Performance on Exascale Systems: A Comparative Analysis of Panel Factorization

High performance LINPACK (HPL) remains the primary benchmark for evaluating supercomputing performance. It includes many parts with substantial internal complexity, and its performance is affected by a large number of parameters that interact in ways that are difficult to predict on large-scale heterogeneous supercomputer systems. We present a comprehensive performance analysis of HPL on Frontier, the world’s first exascale supercomputer, which achieved HPL performance of 1.35 exaflops. Through empirical parameter tuning, detailed modeling, and comparative evaluation, we uncover critical performance insights, share lessons learned, and outline best practices for effective parameter tuning on exascale systems. We introduce and evaluate two novel PDFACT strategies: a dedicated-thread (DT) variant and a GPU-based variant (GPUPDFACT) implementation using HIP cooperative groups, demonstrating that GPU-based factorization outperforms conventional CPU-based PDFACT on Frontier’s architecture. Our findings establish key performance factors for HPL on exascale systems and offer valuable guidance for future high-performance computing and benchmarking efforts.

Lu, Hao [ORNL] (ORCID:000000018941870X)↗

DOE OSTI2025

Deploying and Tracking Software with NCCS Software Provisioning

The National Center for Computational Sciences (NCCS) at Oak Ridge National Laboratory has a long history of deploying ground-breaking leadership-class supercomputers for the U.S. Department of Energy. The latest in this line of supercomputers is Frontier, the first supercomputer to break the exascale barrier (1018 floating-point operations per second) on the TOP500 list. Frontier serves a wide array of scientific domains, from traditional simulation-based workloads to newer AI and Machine Learning workloads. To best serve the NCCS user community, NCCS uses Spack to deploy a comprehensive software stack of scientific software packages, providing straightforward access to these packages through Lmod Environment Modules. Maintaining a large software stack while also including multiple new compiler releases each year is a very time-consuming task. Additionally, it is not straightforward to provide a software stack alongside existing vendor-provided software such as the HPE/Cray Programming Environment (CPE), and existing CPE, Spack, and Lmod integration does not allow for multiple versions of GPU libraries such as AMD’s ROCm to be used. To address these challenges and shortcomings, NCCS has developed the NCCS Software Provisioning tool (NSP)1, a tool for deploying and monitoring software stacks on HPC systems. NSP allows NCCS to quickly and effectively provision software stacks from the ground up using template-driven recipes and configuration files. NSP is successfully deployed on Frontier and several other NCCS clusters, enabling the NCCS software team to quickly deploy software stacks for newly-released compilers, expand current software offerings, better support GPU-based software, and monitor Lmod module usage to identify unused software packages that can be removed from the software stack. In this work, we discuss the shortcomings of the previous CPE, Spack, and Lmod usage at NCCS, provide further details on the implementation and structure of NSP, then discuss the benefits that NSP provides.

Rentschler, Asa [ORNL] (ORCID:0009000597694743)↗

DOE OSTI2026

CORE-BFS: Communication-Optimized REctangular-partitioned BFS Achieving 160.845 TeraTEPS on Frontier Supercomputer

Distributed Breadth-First Search (BFS) is fundamental to many large-scale graph applications, but its performance on parallel systems is often limited by high communication overhead. This paper presents CORE-BFS, an extremely scalable GPU-based BFS implementation that introduces a unique rectangular 2D partitioning-based design for Frontier supercomputer. To further improve performance, we propose four key optimizations: (1) Rectangular 2D-partition specific data formats that use two compressed row and one compressed column status array bitmaps combined with a Double Compressed Sparse Row (DCSR) format per partition, reducing memory footprint and inter-rank traffic; (2) Adaptive frontier & communication strategy that unifies top-down and bottom-up traversal on the rectangular layout, uses lazy synchronization in top-down levels, and switches variants based on frontier size to minimize communication overhead; (3) Frontier-split degree-aware update that maps frontier vertices to thread-centric, wavefront-centric, and block-centric kernels based on their degree to improve GPU utilization and memory coalescing; (4) Row-reduction pipeline that overlaps bottom-up adjacency list processing with row-wise bitmap reduction to hide inter-rank latency. Together, these techniques increase parallelism while reducing memory and communication overhead. On the Graph500 benchmark, CORE - BFS scales up to 9,248 Frontier nodes with scale-42 graphs and reaches 160.845 TTEPS, delivering a 5.42 × speedup over our previous Frontier implementation.

Yang, Haoshen [Rutgers University]↗