skip to main content

Search for: All records

Creators/Authors contains: "Jiang, C."

Note: When clicking on a Digital Object Identifier (DOI) number, you will be taken to an external site maintained by the publisher. Some full text articles may not yet be available without a charge during the embargo (administrative interval).
What is a DOI Number?

Some links on this page may take you to non-federal websites. Their policies may differ from this site.

  1. Disciplinary Data Science Topic: Statistics - Descriptive Statistics, Histograms, Scatter Plot Data Science Learning Goals: Students will know how to calculate basic statistics such as mean, standard deviation and relate the use of these statistics learned in class with real-world data and use them to describe the data Students will be able to construct visualization tools such as a histogram to get the range and distribution of the data set. Students will also learn how to interpret the results. Students will be able to use popular data science tools such as Python to analyze the data.
    Free, publicly-accessible full text available May 17, 2023
  2. AI and deep learning are experiencing explosive growth in almost every domain involving analysis of big data. Deep learning using Deep Neural Networks (DNNs) has shown great promise for such scientific data analysis applications. However, traditional CPU-based sequential computing can no longer meet the requirements of mission-critical applications, which are compute-intensive and require low latency and high throughput. Heterogeneous computing (HGC), with CPUs integrated with accelerators such as GPUs and FPGAs, offers unique capabilities to accelerate DNNs. Collaborating researchers at SHREC\inst{1} at the University of Florida, NERSC\inst{2} at Lawrence Berkeley National Lab, CERN Openlab, Dell EMC, and Intel are studying the application of heterogeneous computing (HGC) to scientific problems using DNN models. This paper focuses on the use of FPGAs to accelerate the inferencing stage of the HGC workflow. We present case studies and results in inferencing state-of-the-art DNN models for scientific data analysis, using Intel distribution of OpenVINO, running on an Intel Programmable Acceleration Card (PAC) equipped with an Arria 10 GX FPGA. Using the Intel Deep Learning Acceleration (DLA) development suite to optimize existing FPGA primitives and develop new ones, we were able accelerate the scientific DNN models under study with a speedup from 3x to 6x formore »a single Arria 10 FPGA against a single core (single thread) of a server-class Skylake CPU.« less
  3. Context. With the advent of space-based asteroseismology, determining accurate properties of red-giant stars using their observed oscillations has become the focus of many investigations due to their implications in a variety of fields in astrophysics. Stellar models are fundamental in predicting quantities such as stellar age, and their reliability critically depends on the numerical implementation of the physics at play in this evolutionary phase. Aims. We introduce the Aarhus red giants challenge, a series of detailed comparisons between widely used stellar evolution and oscillation codes that aim to establish the minimum level of uncertainties in properties of red giants arising solely from numerical implementations. We present the first set of results focusing on stellar evolution tracks and structures in the red-giant-branch (RGB) phase. Methods. Using nine state-of-the-art stellar evolution codes, we defined a set of input physics and physical constants for our calculations and calibrated the convective efficiency to a specific point on the main sequence. We produced evolutionary tracks and stellar structure models at a fixed radius along the red-giant branch for masses of 1.0  M ⊙ , 1.5  M ⊙ , 2.0  M ⊙ , and 2.5  M ⊙ , and compared the predicted stellar properties. Results. Oncemore »models have been calibrated on the main sequence, we find a residual spread in the predicted effective temperatures across all codes of ∼20 K at solar radius and ∼30–40 K in the RGB regardless of the considered stellar mass. The predicted ages show variations of 2–5% (increasing with stellar mass), which we attribute to differences in the numerical implementation of energy generation. The luminosity of the RGB-bump shows a spread of about 10% for the considered codes, which translates into magnitude differences of ∼0.1 mag in the optical V -band. We also compare the predicted [C/N] abundance ratio and find a spread of 0.1 dex or more for all considered masses. Conclusions. Our comparisons show that differences at the level of a few percent still remain in evolutionary calculations of red giants branch stars despite the use of the same input physics. These are mostly due to differences in the energy generation routines and interpolation across opacities, and they call for further investigation on these matters in the context of using properties of red giants as benchmarks for astrophysical studies.« less
  4. Contact. The large quantity of high-quality asteroseismic data that have been obtained from space-based photometric missions and the accuracy of the resulting frequencies motivate a careful consideration of the accuracy of computed oscillation frequencies of stellar models, when applied as diagnostics of the model properties. Aims. Based on models of red-giant stars that have been independently calculated using different stellar evolution codes, we investigate the extent to which the differences in the model calculation affect the model oscillation frequencies and other asteroseismic diagnostics. Methods. For each of the models, which cover four different masses and different evolution stages on the red-giant branch, we computed full sets of low-degree oscillation frequencies using a single pulsation code and, from these frequencies, typical asteroseismic diagnostics. In addition, we carried out preliminary analyses to relate differences in the oscillation properties to the corresponding model differences. Results. In general, the differences in asteroseismic properties between the different models greatly exceed the observational precision of these properties. This is particularly true for the nonradial modes whose mixed acoustic and gravity-wave character makes them sensitive to the structure of the deep stellar interior and, hence, to details of their evolution. In some cases, identifying these differences ledmore »to improvements in the final models presented here and in Paper I; here we illustrate particular examples of this. Conclusions. Further improvements in stellar modelling are required in order fully to utilise the observational accuracy to probe intrinsic limitations in the modelling and improve our understanding of stellar internal physics. However, our analysis of the frequency differences and their relation to stellar internal properties provides a striking illustration of the potential, in particular, of the mixed modes of red-giant stars for the diagnostics of stellar interiors.« less