Precision Cascade: A novel algorithm for multi-precision extreme compression
Explore the source record for details and available documents.
73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
SEARCH · Search NASA
Results for “precision”
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.
ASSESSMENT OF SYSTEMATIC CHROMATIC ERRORS THAT IMPACT SUB-1% PHOTOMETRIC PRECISION IN LARGE-AREA SKY SURVEYS
Meeting the science goals for many current and future ground-based optical large-area sky surveys requires that the calibrated broadband photometry is both stable in time and uniform over the sky to 1% precision or better. Past and current surveys have achieved photometric precision of 1%–2% by calibrating the survey’s stellar photometry with repeated measurements of a large number of stars observed in multiple epochs. The calibration techniques employed by these surveys only consider the relative frame-by-frame photometric zeropoint offset and the focal plane position-dependent illumination corrections, which are independent of the source color. However, variations in the wavelength dependence of the atmospheric transmission and the instrumental throughput induce source color-dependent systematic errors. These systematic errors must also be considered to achieve the most precise photometric measurements. In this paper, we examine such systematic chromatic errors (SCEs) using photometry from the Dark Energy Survey (DES) as an example. We first define a natural magnitude system for DES and calculate the systematic errors on stellar magnitudes when the atmospheric transmission and instrumental throughput deviate from the natural system. We conclude that the SCEs caused by the change of airmass in each exposure, the change of the precipitable water vapor and aerosol in the atmosphere over time, and the non-uniformity of instrumental throughput over the focal plane can be up to 2% in some bandpasses. We then compare the calculated SCEs with the observed DES data. For the test sample data, we correct these errors using measurements of the atmospheric transmission and instrumental throughput from auxiliary calibration systems. The residual after correction is less than 0.3%. Moreover, we calculate such SCEs for Type Ia supernovae and elliptical galaxies and find that the chromatic errors for non-stellar objects are redshift-dependent and can be larger than those for stars at certain redshifts.
79 ASTRONOMY AND ASTROPHYSICS↗
Towards in-situ high precision local material velocity measurements in lattice materials under dynamic compression
The study of dynamic compression of materials requires high precision spatial and temporal resolution to infer the material state at any given time. In ultra-fast imaging experiments such as those performed at the Dynamic Compression Sector at the Advanced Photon Source, the current CCD frame rate is less than the X-ray pulse rate necessitating that 4 CCDs be used to image a single scintillator screen to obtain the same projection of transmitted or diffracted beam at time intervals synchronized with the X-ray beam. The use of multiple-camera setups is often employed for a variety of reasons including enhancing the temporal resolution through interleaving frames or increasing the number of viewpoints, for example. Physical corrections are made to have each CCD approximately viewing the same field of view, however we present corrections that can be made after the data collection to further refine and be able to compare movements and changes in the sample between independent CCDs. These software corrections enable higher precision and higher fidelity comparison of position changes due to sample deformation. Furthermore, we present a method to correct images (or frames) as a demonstration toward high fidelity displacement measurements in these types of ultra-fast multi-camera setup experiments.
36 MATERIALS SCIENCE↗
Atomically Precise Crystalline Materials Based on Kinetically Inert Metal Ions via Reticular Mechanopolymerization
Explore the source record for details and available documents.
99 GENERAL AND MISCELLANEOUS↗
High-precision γ -ray spectroscopy of the cardiac PET imaging isotope Rb 82 and its impact on dosimetry
82 Rb is a positron-emitting isotope used in cardiac positron emission tomography (PET) imaging which has been reported to deliver a significantly lower effective radiation dose than analogous imaging isotopes like 201 Tl and 99m Tc sestamibi. High-quality β-decay data are essential to accurately appraise the total dose received by the patients. A source of 82 Sr was produced at the Brookhaven Linac Isotope Producer (BLIP), transported to Argonne National Laboratory, and studied with the Gammasphere facility. Significant revisions have been made to the level scheme of 82 Kr including 12 new levels, 50 new γ-ray transitions, and the determination of many new spin assignments through angular correlations. Lastly, these new high-quality data allow a precise reappraisal of the β-decay strength function and thus the consequent dose received by patients.
73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
Fine-Grained Exploitation of Mixed Precision for Faster CNN Training
As deep convolutional neural networks (CNNs) have become increasingly popular and successful at an ever-widening number of machine learning tasks specialized hardware has become increasingly available for training and deploying them. NVIDIA's recent Volta architecture includes tensor cores which perform a fused operation reduced and mixed precision (16-bit multiply, 32-bit accumulate). Recent research indicates that, typically, very little is lost (in terms of training accuracy) when half precision is used in place of single precision, and performance gains can be made by doing arithmetic in reduced precision. In this work we demonstrate that making layer-by-layer choices as to the arithmetic/data precision can lead to further performance improvement. In our study of 25,200 CNNs we demonstrate an average speedup (over purely half precision) of 1.27x and speedups as high as 3.64x by appropriately combining single and half precision arithmetic and data types on a layer-by-layer basis.
Johnston, Travis↗
Mixed-Precision S/DGEMM Using the TF32 and TF64 Frameworks on Low-Precision AI Tensor Cores
Using NVIDIA graphics processing units (GPUs) equipped with Tensor Cores has enabled the significant acceleration of general matrix multiplication (GEMM) for applications in machine learning (ML) and artificial intelligence (AI) and in high-performance computing (HPC) generally. The use of such power-efficient, specialized accelerators can provide a performance increase between 8 × and 20 ×, albeit with a loss in precision. However, a high level of precision is required in many large scientific and HPC applications, and computing in single or double precision is still necessary for many of these applications to maintain accuracy. Fortunately, mixed-precision methods can be employed to maintain a higher level of numerical precision while also taking advantage of the performance increases from computing with lower-precision AI cores. With this in mind, we extend the state of the art by using NVIDIA’s new TF32 framework. This new framework not only burdens some constraints of the previous frameworks, such as costly 32 16-bit castings but also provides an equivalent precision and performance by using a much simpler approach. We also propose a new framework called TF64 that attempts double-precision arithmetic with low-precision Tensor Cores. Although this framework does not exist yet, we validated the correctness of this idea and achieved an equivalent of 64-bit precision on 32-bit hardware.
Valero Lara, Pedro↗
The Aemulus Project. I. Numerical Simulations for Precision Cosmology
The rapidly growing statistical precision of galaxy surveys has led to a need for ever more precise predictions of the observables used to constrain cosmological and galaxy formation models. The primary avenue through which such predictions will be obtained is suites of numerical simulations. These simulations must span the relevant model parameter spaces, be large enough to obtain the precision demanded by upcoming data, and be thoroughly validated in order to ensure accuracy. In this paper, we present one such suite of simulations, forming the basis for the AEMULUS Project, a collaboration devoted to precision emulation of galaxy survey observables. We have run a set of 75 (1.05 h(-1) Gpc)(3) simulations with mass resolution and force softening of 3.51 x 10(10) (Omega(m)/0.3) h(-1) M-circle dot and 20 h(-1) kpc, respectively, in 47 different wCDM cosmologies spanning the range of parameter space allowed by the combination of recent cosmic microwave background, baryon acoustic oscillation, and Type Ia supernova results. We present convergence tests of several observables including spherical overdensity halo mass functions, galaxy projected correlation functions, galaxy clustering in redshift space, and matter and halo correlation functions and power spectra. We show that these statistics are converged to 1% (2%) or to the sample variance of the statistic, whichever is larger, for halos with more than 500 (200) particles, respectively, and scales of r > 200 h(-1) kpc. in real space or k similar to 3 h Mpc(-1). in harmonic space for z <= 1. We find that the dominant source of uncertainty comes from varying the particle loading of the simulations. This leads to large systematic errors for statistics using halos with fewer than 200 particles and scales smaller than k similar to 4 h Mpc(-1). We provide the halo catalogs and snapshots detailed in this work to the community at. https://AemulusProject.github.io.
large-scale structure of universe↗
Precision determination of electron scattering angle by differential nuclear recoil energy method
Not Available
Liyanage, N.↗
Engineering on-chip nanoporous gold material libraries via precision photothermal treatment [Precision Photothermal Annealing of Nanoporous Gold Thin Films for the Microfabrication of a Single-ship Material Libraries]
Single-chip material libraries of thin films of nanostructured materials are a promising approach for high throughput studies of structure-property relationship in the fields of physics and biology. Nanoporous gold (np-Au), produced by an alloy corrosion process, is a nanostructured material of specific interest in both these fields. One attractive property of np-Au is its self-similar coarsening behavior by thermally induced surface diffusion. However, traditional heat application techniques for the modification of np-Au are bulk processes that cannot be used to generate a library of different pore sizes on a single chip. Laser micromachining offers an attractive solution to this problem by providing a means to apply energy with high spatial and temporal resolution. In our present study we use finite element multiphysics simulations to predict the effects of laser mode (continuous-wave vs. pulsed) and supporting substrate thermal conductivity on the local np-Au film temperatures during photothermal annealing and subsequently investigate the mechanisms by which the np-Au network is coarsening. Our simulations predict that continuous-wave mode laser irradiation on a silicon supporting substrate supports the widest range of morphologies that can be created through the photothermal annealing of thin film np-Au. Using this result we successfully fabricate a single-chip material library consisting of 81 np-Au samples of 9 different morphologies for use in increased throughput material interaction studies.
36 MATERIALS SCIENCE↗
Precision measurement of the top-quark mass in lepton plus jets final states
N/A
72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
High Precision Magnetic Field Measurement for the Muon g-2 Experiment
The Muon g-2 experiment at Fermilab will measure the anomalous magnetic moment of the muon, aµ with a precision of 140 ppb, a four-fold improvement over the former BNL E821 result. If the current discrepancy of about 3.5σ between the Standard Model value for aµ and the former experiments was real, then the new experiment would verify the deviation with >7σ significance. The experiment uses muons with the magic momentum of 3.09 GeV/c stored in a very homogeneous magnetic field of a 45 m long storage ring. The determination of aµ requires the precise knowledge of both the anomalous muon spin precession, ωa, and the magnetic field, ωp, expressed in terms of the free-proton Larmor frequency. The field measurement uses the high precision tool of pulsed Nuclear Magnetic Resonance (pNMR) and requires cross-calibrating a set of probes: a very accurate NMR probe calibrates the 17 probes on the in-vacuum field mapping system (the so-called trolley). This trolley in turn is used to map the field over the full azimuth of the storage ring. The trolley measurement also provides the calibration of the 378 fixed probes around the ring that constantly monitor the field drift. This article presents the summary of the entire field measurement system needed to determine ωp. This work was supported in part by the US DOE, Fermilab and US DOE OHEP under contract No. KA2201020.
72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
Precise ROCOF estimation algorithm for low inertia power grids
The precise estimation of Rate Of Change Of Frequencies (ROCOFs) in a generation trip event can be helpful on power system inertia estimation, fast system response, and accurate event capturing. However, the ROCOF estimations from the existing Synchronized Measurement Devices (SMDs) are usually simply calculated by the finite difference between two adjacent frequency measurement points. The ineluctable noises, disturbances, and spikes from real-world frequency measurements can bring large dynamics to the ROCOF estimation and thus can result in an inaccurate estimation of the initial ROCOF. This issue becomes more serious when the target power grid has low inertia, and a large amount of distributed energy sources are deployed. To address this issue, a precise ROCOF estimation algorithm is designed based on least square method with flexible window size. In addition, a median filter is also designed and applied on the frequency measurements before using the proposed algorithm. The window size and thresholds in the proposed algorithm are determined with historical event data analysis. The proposed algorithm is deployed in a low cost, flexible, and distribution level universal grid analyzer (UGA) platform. In conclusion, multiple experiments are conducted in both a laboratory and the Hawaiian Islands to verify the effectiveness of the proposed algorithm.
24 POWER TRANSMISSION AND DISTRIBUTION↗
Precision cosmology with time delay lenses: High resolution imaging requirements
Lens time delays are a powerful probe of cosmology, provided that the gravitational potential of the main deflector can be modeled with sufficient precision. Recent work has shown that this can be achieved by detailed modeling of the host galaxies of lensed quasars, which appear as ``Einstein Rings'' in high resolution images. The distortion of these arcs and counter-arcs, as measured over a large number of pixels, provides tight constraints on the difference between the gravitational potential between the quasar image positions, and thus on cosmology in combination with the measured time delay. We carry out a systematic exploration of the high resolution imaging required to exploit the thousands of lensed quasars that will be discovered by current and upcoming surveys with the next decade. Specifically, we simulate realistic lens systems as imaged by the Hubble Space Telescope (HST), James Webb Space Telescope (JWST), and ground based adaptive optics images taken with Keck or the Thirty Meter Telescope (TMT). We compare the performance of these pointed observations with that of images taken by the Euclid (VIS), Wide-Field Infrared Survey Telescope (WFIRST) and Large Synoptic Survey Telescope (LSST) surveys. We use as our metric the precision with which the slope γ' of the total mass density profile ρ tot ∝ r–γ' for the main deflector can be measured. Ideally, we require that the statistical error on γ' be less than 0.02, such that it is subdominant to other sources of random and systematic uncertainties. We find that survey data will likely have sufficient depth and resolution to meet the target only for the brighter gravitational lens systems, comparable to those discovered by the SDSS survey. For fainter systems, that will be discovered by current and future surveys, targeted follow-up will be required. Furthermore, the exposure time required with upcoming facilitites such as JWST, the Keck Next Generation Adaptive Optics System, and TMT, will only be of order a few minutes per system, thus making the follow-up of hundreds of systems a practical and efficient cosmological probe.
79 ASTRONOMY AND ASTROPHYSICS↗
Complementarity between nonstandard Higgs boson searches and precision Higgs boson measurements in the MSSM
Precision measurements of the Higgs boson properties at the LHC provide relevant constraints on possible weak-scale extensions of the Standard Model (SM). In the context of the minimal supersymmetric Standard Model (MSSM) these constraints seem to suggest that all the additional, non-SM-like Higgs bosons should be heavy, with masses larger than about 400 GeV. This article shows that such results do not hold when the theory approaches the conditions for “alignment independent of decoupling,” where the lightest CP-even Higgs boson has SM-like tree-level couplings to fermions and gauge bosons, independently of the nonstandard Higgs boson masses. In addition, the combination of current bounds from direct Higgs boson searches at the LHC, along with the alignment conditions, have a significant impact on the allowed MSSM parameter space yielding light additional Higgs bosons. In particular, after ensuring the correct mass for the lightest CP-even Higgs boson, we find that precision measurements and direct searches are complementary and may soon be able to probe the region of non-SM-like Higgs boson with masses below the top quark pair mass threshold of 350 GeV and low to moderate values of tanβ.
72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
THE MIRA–TITAN UNIVERSE: PRECISION PREDICTIONS FOR DARK ENERGY SURVEYS
Large-scale simulations of cosmic structure formation play an important role in interpreting cosmological observations at high precision. The simulations must cover a parameter range beyond the standard six cosmological parameters and need to be run at high mass and force resolution. A key simulation-based task is the generation of accurate theoretical predictions for observables using a finite number of simulation runs, via the method of emulation. Using a new sampling technique, we explore an eight-dimensional parameter space including massive neutrinos and a variable equation of state of dark energy. We construct trial emulators using two surrogate models (the linear power spectrum and an approximate halo mass function). The new sampling method allows us to build precision emulators from just 26 cosmological models and to systematically increase the emulator accuracy by adding new sets of simulations in a prescribed way. Emulator fidelity can now be continuously improved as new observational data sets become available and higher accuracy is required. Finally, using one ΛCDM cosmology as an example, we study the demands imposed on a simulation campaign to achieve the required statistics and accuracy when building emulators for investigations of dark energy.
79 ASTRONOMY AND ASTROPHYSICS↗
Towards Precision-Aware Fault Tolerance Approaches for Mixed-Precision Applications
Graphics Processing Units (GPUs), the dominantly adopted accelerators in HPC systems, are susceptible to transient hardware fault. New generation of GPUs feature mixed-precision architectures such as NVIDIA Tensor Cores to accelerate matrix multiplications. While being widely adapted, how would they behave under transient hardware faults remain unclear. In this study, we conduct a large-scale fault injection experiments on GEMM kernels implemented with different floating-point data types on the V100 and A100 Tensor Cores, and show distinct error resilience characteristics for the GEMMS with different formats. In the future, we plan to explore this space by building precision-aware floating-point fault tolerance techniques for applications such as DNNs that exercise low-precision computations.
Fang, Bo↗
Deterministic Design of High Precision Slicing Machine Final Report CRADA No. T-160-91-1-C (Advanced manufacturing initiative final report. Deterministic design of high precision slicing machine)
The project is designed to merge the expertise of LLNL and a US slicing machine builder in the development and production of a family of ultra-precise, slicing/grinding machines. The project will lead to the joint development of a modular family of slicers which are easily configured to a given accuracy and size. The slicers can then be profitably produced and marketed for similar precision machining applications. This project is directed towards maintaining a competitive edge for US manufacturers in the area of high precision slicing and grinding machines.
36 MATERIALS SCIENCE↗