SEARCH · Engineering Papers
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.
UMT is a Sn deterministic thermal radiative transport Fortran proxy application. Its primary purpose is to research how data layout, programming paradigms and architectures effect the implementation and performance of Sn transport in Fortran on unstructured meshes. It is based on the LLNL Teton Sn deterministic thermal radiative transport code.
Abstract not provided.
Workflow and serverless frameworks have empowered new approaches to distributed application design by abstracting compute resources. However, their typically limited or one-size-fits-all support for advanced data flow patterns leaves optimization to the application programmer—optimization that becomes more difficult as data become larger. The transparent object proxy, which provides wide-area references that can resolve to data regardless of location, has been demonstrated as an effective low-level building block in such situations. Here we propose three high-level proxy-based programming patterns—distributed futures, streaming, and ownership—that make the power of the proxy pattern usable for more complex and dynamic distributed program structures. We motivate these patterns via careful review of application requirements and describe implementations of each pattern. As a result, we evaluate our implementations through a suite of benchmarks and by applying them in three meaningful scientific applications, in which we demonstrate substantial improvements in runtime, throughput, and memory usage.
Summary The Oak Ridge Leadership Computing Facility (OLCF) has a long history of supporting and promoting GPU‐accelerated computing starting with the deployment of the Titan supercomputer in 2021 and continuing with the Summit supercomputer which has a theoretical peak performance of approximately 200 petaflops. Because the majority of Summit's computational power comes from its 27,972 GPUs, users must port their applications to one of the supported programming models in order to make efficient use of the system. To prepare the transition to Frontier, the OLCF's exascale supercomputer, users will need to adapt to an entirely new ecosystem which will include new hardware and software technologies. First, users will need to familiarize themselves with the AMD Radeon GPU architecture. Furthermore, users who have been previously relying on CUDA will need to transition to the Heterogeneous‐Computing Interface for Portability (HIP) or one of the other supported programming models (e.g., OpenMP, OpenACC). In this work, we describe our initial experiences and lessons learned in porting three applications or proxy apps currently running on Summit to the HPE/Cray ecosystem to leverage the compute power from AMD GPUs: minisweep, GenASiS, and Sparkler. Each one is representative of current production workloads utilized at the OLCF, different programming languages, and different programming models.
This document details the development and current state of MGMC: a proxy Multi-Group Monte Carlo transport application. The goal of this work was to develop a light-weight Monte Carlo transport solver that could be easily reconfigured to test parallelization strategies targeting advanced architectures and heterogeneous computing environments. Additionally, MGMC is able to act as a proxy-app for testing the development of algorithms and the integration of other libraries and into Monte Carlo transport codes. MGMC is currently able to run in parallel using OpenMP for CPU threads or CUDA for NVIDIA GPUs. MGMC was also used to investigate the using SYCL for portable parallelization targeting either CPU threads, vender-agnostic GPUs, and ARM chips, detailed in Section II. In the process of developing MGMC, several other C++ libraries have been created with the intent for re-use in future research projects and potential inclusion in production-level codes, detailed in Section III. The physics capabilities in MGMC are detailed in Section IV, code verification is detailed in Section V, and performance results are detailed in Section VI.
Boron isotope systematics of planktonic foraminifera from core-top sediments and culture experiments have been studied to investigate the sensitivity ofδ11B of calcite tests to seawater pH. However, our knowledge of the relationship between δ11B and pH remains incomplete for many taxa. Thus, to expand the potential scope of application of this proxy, we report δ11B data for seven different species of planktonic foraminifera from sediment core tops. We utilize a method for the measurement of small samples of foraminifera and calculate the δ11B-calcite sensitivity to pH for Globigerinoides ruber, Trilobus sacculifer (sacc or without sacc), Orbulina universa, Pulleniatina obliquiloculata, Neogloboquadrina dutertrei, Globorotalia menardii, and Globorotalia tumida, including for unstudied core tops and species. These taxa have diverse ecological preferences and are from sites that span a range of oceanographic regimes, including some that are in regions of air–sea equilibrium and others that are out of equilibrium with the atmosphere. The sensitivity of δ11Bcarbonate to δ11Bborate (e.g., Δδ11Bcarbonate/Δδ11Bborate) in core tops is consistent with previous studies for T. sacculifer and G. ruber and close to unity for N. dutertrei, O. universa, and combined deep-dwelling species. Deep-dwelling species closely follow the core-top calibration fo rO. universa, which is attributed to respiration-driven microenvironments likely caused by light limitation and/or symbiont–host interactions. Our data support the premise that utilizing boron isotope measurements of multiple species within a sediment core can be utilized to constrain vertical profiles of pH andpCO2 at sites spanning different oceanic regimes, thereby constraining changes in vertical pH gradients and yielding insights into the past behavior of the oceanic carbon pumps.
Version 4.0 of the ECP Proxy App Suite is practically unchanged from the previous release. The current set of proxies has proven useful for many aspects of benchmarking and co-design and we see little reason to alter the suite. Although there have been few changes to the ECP suite, the team has been hard at work in other areas. In the area of Machine Learning (ML) we have now created a separate proxy suite dedicated to this scientific applications of ML. The suite includes: miniGAN (Generative Adversarial Networks), miniRL (Reinforcement Learning), CRADL (inline inference), Cosmoflow-Benchmark (Convolutional Neural Network), and MLPerf-DeepCam (Climate Segmentation Benchmark). Section 3 contains more information about these proxies as well as the principles that are guiding the development of the suite. We have surveyed available proxies for several application domains including Computational Fluid Dynamics, Quantum Chemistry, Quantum Computing Simulation, Molecular Dynamics, Monte Carlo Transport, and Density Functional Theory to identify gaps in proxy coverage. Several new proxy apps are either already available or will be released soon to fill these gaps. Section 4 provides full details. Finally, section 5 reports on our continued collaboration with the ECP Continuous Integration (CI) effort to use proxy apps to help identify problems and roadblocks to cross-lab CI. We have also assisted the El Capitan Center of Excellence (COE) to stand up a CI system that will be used to test releases of HPE and AMD software stacks. We hope that the COE effort can serve as a model for ECP by showing how the Proxy App Team can work with AD and ST teams identify critical features, kernels, patterns, etc. and incorporate them into a CI system that will help ensure that Frontier and Aurora will provide those needed capabilities.
The cadmium isotope system has found use as a tracer for biological productivity, redox and organic carbon burial in the oceans. There are, however, very few observational constraints on Cd isotopic behaviour in modern sulfidic marine conditions, limiting our understanding of the modern Cd cycle, and our ability to use Cd isotope measurements of ancient sedimentary deposits as a paleoceanographic proxy. Here we study the behaviour of dissolved Cd in the water column of the Black Sea and its incorporation into sediments on the basin floor. The isotopic composition of dissolved Cd in the upper ∼50 m of the Black Sea water column is controlled by a combination of biological uptake and regeneration along with mixing of river water and Mediterranean seawater. Cadmium declines to <2 % of its peak subsurface concentration within the zone of nitrate-reduction but above the sulfide chemocline. The isotopic composition of dissolved Cd below the sulfide chemocline evolves in a manner that is consistent with experimental Cd sulfide fractionation factors, providing field-based support for these earlier studies. In contrast, the stability of the dissolved Cd isotopic signature above the chemocline, despite a large reduction in dissolved Cd concentrations, is explained by diffusion towards the Cd-sulfide sink below the chemocline. Sediments accumulating in the deep Black Sea below the sulfide chemocline have isotopic compositions that are similar to dissolved Cd in its near-surface aqueous concentration maximum. Sediments accumulating at shallower depths have lighter isotopic compositions that are likely affected by non-sulfidic burial pathways mediated by macrofaunal mixing and diagenesis of shallow burial phases. Our study shows that the isotopic composition of bulk sediments accumulating under sulfidic conditions is similar to dissolved Cd in the upper water column, supporting its use as a paleo-chemical proxy for ancient seawater as long as basin-scale controls on open ocean seawater sources are considered. We show that Cd burial in the deep Black Sea, predominantly as CdS, accounts for ∼0.5–15 % of total annual Cd burial in the global ocean.
Deoxycholic acid (DOC) is a unique, biologically derived surfactant with facial amphiphilicity that has been exploited, albeit minimally, in supramolecular assembly of materials. Here, we present the synthesis and structural characterization of three hybrid metal (Zn 2+ and Cd 2+ )-DOC compounds. Analysis by single-crystal Xray diffraction reveals the many interactions that are possible between these facial surfactants and the influence of solvent molecules that drive the assembly of materials. These structures are the first metal- DOC complexes besides those obtained from alkali and alkaline earth metals. We isolated polymeric chains of both Cd and Zn (Zn poly -DOC and Cd poly -DOC) from water. Major interactions between DOC molecules in these phases are hydrophobic in nature. Cd poly -DOC exhibits unique P1 symmetry, with complete interdigitation of the amphiphiles between neighboring polymeric chains. Zn 4 -DOC, obtained from methanol dissolution of Zn poly -DOC, features the OZn 4 tetrahedron, widely known in basic zinc acetate and MOF-5 (metal organic framework). We document a solvent-driven, roomtemperature transition between Zn poly -DOC and Zn 4 -DOC (in both directions) by scanning and transmission electron microscopies in addition to small-angle X-ray scattering, powder X-ray diffraction, and infrared spectroscopy. These studies show the methanoldriven transition of Zn poly -DOC to Zn 4 -DOC occurs via an intermediate with no long-range order of the Zn 4 clusters, indicating the strongest interactions driving assembly are intramolecular. On the contrary, water-driven solid-to-solid transformation from Zn 4 -DOC to Zn poly -DOC exhibits crystal-to-crystal transformation. Zn poly -DOC is robust, easy to synthesize, and comprised of biologically benign components, so we demonstrate dye absorption as a proxy for water treatment applications. It favors absorption of positively charged dyes. These studies advance molecular level knowledge of the supramolecular assembly of facial surfactants that can be exploited in the design of organic-inorganic hybrid materials. This research also highlights the potential of solvent for tuning supramolecular assembly processes, leading to new hybrid materials featuring facial surfactants.
Live Fuel Moisture (LFM) is a critical parameter for wildfire risk assessment, traditionally measured by labor-intensive field sampling. However, sampled LFM data are influenced by site-specific factors, such as local vegetation types and plant traits, and are often collected retrospectively after wildfire events, making it difficult to obtain pre-fire data for predictive applications. Here, we evaluate the relationship between LFM and Vegetation Water Content (VWC) and Soil Moisture (SM) retrieved from SMAP L-band brightness temperature using the Maximum Entropy Production (MEP) approach. The MEP-retrieved VWC exhibited strong correlation with in situ measurements of LFM ( r > 0.6) in the Western U.S. The integration of high-resolution vegetation coverage data enhances the detection of sub-grid vegetation heterogeneity. This study demonstrates the operational potential of remote sensing derived VWC as a scalable proxy of LFM, supporting its application in regional assessment of wildfire risk.
Climate field reconstructions (CFR) attempt to estimate spatiotemporal fields of climate variables in the past using climate proxies such as tree rings, ice cores, and corals. Data Assimilation (DA) methods are a recent and promising new means of deriving CFRs that optimally fuse climate proxies with climate model output. Despite the growing application of DA-based CFRs, little is understood about how much the assimilated proxies change the statistical properties of the climate model data. To address this question, we propose a robust and computationally efficient method, based on functional data depth, to evaluate differences in the distributions of two spatiotemporal processes. In this study, we apply our test to study global and regional proxy influence in DA-based CFRs by comparing the background and analysis states, which are treated as two samples of spatiotemporal fields. We find that the analysis states are significantly altered from the climate-model-based background states due to the assimilation of proxies. Moreover, the difference between the analysis and background states increases with the number of proxies, even in regions far beyond proxy collection sites. Our approach allows us to characterize the added value of proxies, indicating where and when the analysis states are distinct from the background states.
Abstract Carbon dioxide (CO 2 ) has been linked to many deleterious health effects, and it has also been used as a proxy for building occupancy measurements. These applications have created a need for low-cost and low-power CO 2 sensors that can be seamlessly incorporated into existing buildings. We report a resonant mass sensor coated with a solution-processable polymer blend of poly(ethylene oxide) (PEO) and poly(ethyleneimine) (PEI) for the detection of CO 2 across multiple use conditions. Controlling the polymer blend composition and nanostructure enabled better transport of the analyte gas into the sensing layer, which allowed for significantly enhanced CO 2 sensing relative to the state of the art. Moreover, the hydrophilic nature of PEO resulted in water uptake, which provided for higher sensing sensitivity at elevated humidity conditions. Therefore, this key integration of materials and resonant sensor platform could be a potential solution in the future for CO 2 monitoring in smart infrastructure.
Version 4.0 of the ECP Proxy App Suite is practically unchanged from the previous release. The current set of proxies has proven useful for many aspects of benchmarking and co-design and we see little reason to alter the suite. Although there have been few changes to the ECP suite, the team has been hard at work in other areas. In the area of Machine Learning (ML) we have now created a separate proxy suite dedicated to this scientific applications of ML.
ARM-based manycore CPU architectures are well-positioned to provide the rising memory throughput requirements of modern data intensive scientific applications in High Performance Computing (HPC). The Fujitsu A64FX CPU platform is based on the ARM v8.2A architecture, and is the processor of the flagship Japanese supercomputer - "Fugaku", which was previously ranked as the #1 supercomputer in the world according to the Top500 list. The Nvidia Grace superchip features 144 Neoverse V2 cores based on the ARMv9 architecture with 4x128b SVE2, providing exceptional computational power. The chip supports up to 480GB of memory, making it ideal for AI, machine learning, and scientific computing workloads. In this paper, we conduct a thorough performance exploration of a variety of parallel bandwidth-sensitive benchmarks and applications compiled with the native Fujitsu compiler on a Fugaku A64FX compute node and ARM (LLVM) Compiler on an NVIDIA Grace superchip compute node, engaging all the computational cores per cluster using OpenMP multithreading (assuming the cores can drive the available bandwidth). Our ultimate goals are to study the resource utilization of scientific applications and benchmarks on A64FX and Grace superchip, considering graph application scenarios ( GAP Benchmark suite) and eleven appli- cation proxies from the Rodinia heterogeneous benchmark suite (considering domains such as Data Mining, Bioinformatics, Fluid Dynamics, Pattern Recognition, etc.). Through exhaustive performance monitoring, we quantify the resource utilization of diverse OpenMP-based HPC applications on both the Fujitsu A64FX and the Nvidia Grace Superchip platforms.
We measure star-spot filling fractions for 240 stars in the Pleiades and M67 open star clusters using APOGEE high-resolution H-band spectra. For this work, we developed a modified spectroscopic pipeline which solves for star-spot filling fraction and star-spot temperature contrast. We exclude binary stars, finding that the large majority of binaries in these clusters (80 per cent) can be identified from Gaia DR3 and APOGEE criteria – important for field star applications. Our data agree well with independent activity proxies, indicating that this technique recovers real star-spot signals. In the Pleiades, filling fractions saturate at a mean level of 0.248 ± 0.005 for active stars with a decline at slower rotation; we present fitting functions as a function of Rossby number. In M67, we recover low mean filling fractions of 0.030 ± 0.008 and 0.003 ± 0.002 for main sequence GK stars and evolved red giants, respectively, confirming that the technique does not produce spurious spot signals in inactive stars. Star-spots also modify the derived spectroscopic effective temperatures and convective overturn time-scales. Effective temperatures for active stars are offset from inactive ones by −109 ± 11 K, in agreement with the Pecaut & Mamajek empirical scale. Star-spot filling fractions at the level measured in active stars changes their inferred overturn time-scale, which biases the derived threshold for saturation. Finally, we identify a population of stars statistically discrepant from mean activity–Rossby relations and present evidence that these are genuine departures from a Rossby scaling. Our technique is applicable to the full APOGEE catalogue, with broad applications to stellar, galactic, and exoplanetary astrophysics.
Recent induced polarization studies suggest that the real part of surface conductivity (σ' surf ) scales linearly with the imaginary conductivity (σ'' = σ'' surf ) or normalized chargeability (Mn) for a range of soil types. The coefficients of this relationship l and l_Mn ( l = σ''/σ' surf or l_Mn = Mn/σ' surf ) allow the separation of the surface and electrolytic conductivities from the bulk conductivity. However, the dependence of these constants on varying soil physicochemical properties, including under unsaturated conditions, is yet to be assessed. Using estimates of σ' surf from 18 undisturbed soil samples from a restored wetland and σ'' measured over a frequency range of 0.01 Hz to 10 kHz, the σ' surf and σ'' were compared with the laboratory measurements of soil properties. Also, l and l_Mn were calculated for each soil sample and regressed them against the soil properties. We find an apparent dependence of l on soil texture, bulk density, organic matter, and moisture contents, with coefficients of determination (R 2 ) ranging from 0.5 to 0.65 at low frequencies (e.g., 1 Hz) but not at high frequencies (e.g., 936 Hz). This dependence of l on soil texture results from the insensitivity of σ'' at low frequency to σ' surf and, by implication, to the soil properties controlling σ' surf . In contrast, l_Mn indicates no correlation with the soil properties because Mn is linearly correlated with σ' surf and correlated with the soil properties controlling σ' surf . Finally, our results call for caution on the application of σ'' at a single frequency as a proxy of σ' surf because σ'' is not necessarily correlated with σ' surf across all soil types. Although using l_Mn derived from multifrequency measurements overcomes this limitation, field acquisition of spectral information (e.g., up to 1000 Hz) remains a challenge.
The application of zinc (Zn) and cadmium (Cd) isotopes as palaeo-proxies in carbonate sediment is rapidly expanding due to their potential for tracing changes in biological productivity in the modern and past oceans. However, there are limited investigations into the chemical cleaning methods required to produce the most consistent and accurate data for these novel isotope systems. This could impact their use as palaeo-proxies to reconstruct ocean-atmosphere-climate interactions throughout Earth's history. To address this concern and expand the utility of the Zn and Cd stable isotope systems as palaeo-productivity tracers, the performance of two standard chemical cleaning protocols for acquiring robust and reliable Zn and Cd isotope datasets were assessed. These include (i) the Cd-cleaning method that uses a reductive step to selectively leach contaminating secondary iron (Fe)-manganese (Mn) (oxyhydr)oxide coatings from the carbonate surface, and an oxidative step that is used to remove post-depositional organic matter and sulphide precipitates; and (ii) the magnesium/calcium cleaning protocol that includes an oxidative step only, leaving secondary Fesingle bondMn (oxyhydr)oxide coatings largely intact. Well-preserved Holocene-, and Mesozoic-aged carbonate sediments were used to test the reliability of these two chemical cleaning methods. The Holocene samples comprised not only aliquots of bulk sediment, but also individual species of planktic and benthic foraminifera. Our results show that the best practice chemical cleaning method for retrieving consistent and accurate Zn and Cd isotope, and Zn/Ca and Cd/Ca datasets for carbonate sediments, requires both reductive and oxidative cleaning following the Cd-cleaning method. This differs from most methodological approaches applied to date that remove the reductive step from the chemical cleaning protocol and apply an oxidative step only, or no chemical cleaning at all. Inclusion of the reductive step in the chemical cleaning method typically shifts δ 66 Zn by ~0.1‰ lower and δ 114 Cd by 0.3‰ higher in the solid phase, while Zn/Ca and Cd/Ca typically decrease 2-fold. The benthic foraminifera, C. wuellerstorfi, that live in ocean bottom waters where the seawater Zn and Cd isotope composition is homogeneous display evidence of Zn and Cd isotope fractionation between seawater and carbonate on the order of 0.08 ± 0.08‰ (2SE, n = 4) and -0.25 ± 0.13‰ (2SE, n = 4), respectively, in agreement with experimental constraints. Furthermore, evidence of Zn isotope fractionation effects are recorded in a naturally-dissolved carbonate sediment, together with laboratory-controlled carbonate dissolution experiments. Based on these results, we recommend the Cd-cleaning method and the application of Zn and Cd isotope fractionation factors to accurately reconstruct past seawater Zn and Cd isotope compositions from carbonate sediments.
Synthetically generated, large graph networks serve as useful proxies to real-world networks for many graph-based applications. The ability to generate such networks helps overcome several limitations of real-world networks regarding their number, availability, and access. Here, we present the design, implementation, and performance study of a novel network generator that can produce very large graph networks conforming to any desired degree distribution. The generator is designed and implemented for efficient execution on modern graphics processing units (GPUs). Given an array of desired vertex degrees and number of vertices for each desired degree, our algorithm generates the edges of a random graph that satisfies the input degree distribution. Multiple runtime variants are implemented and tested: 1) a uniform static work assignment using a fixed thread launch scheme, 2) a load-balanced static work assignment also with fixed thread launch but with cost-aware task-to-thread mapping, and 3) a dynamic scheme with multiple GPU kernels asynchronously launched from the CPU. The generation is tested on a range of popular networks such as Twitter and Facebook, representing different scales and skews in degree distributions. Results show that, using our algorithm on a single modern GPU (NVIDIA Volta V100), it is possible to generate large-scale graph networks at rates exceeding 50 billion edges per second for a 69 billion-edge network. GPU profiling confirms high utilization and low branching divergence of our implementation from small to large network sizes. For networks with scattered distributions, we provide a coarsening method that further increases the GPU-based generation speed by up to a factor of 4 on tested input networks with over 45 billion edges.