Showing papers by "Sandia National Laboratories published in 2022"

Kokkos 3: Programming Model Extensions for the Exascale Era

Abstract: As the push towards exascale hardware has increased the diversity of system architectures, performance portability has become a critical aspect for scientific software. We describe the Kokkos Performance Portable Programming Model that allows developers to write single source applications for diverse high-performance computing architectures. Kokkos provides key abstractions for both the compute and memory hierarchy of modern hardware. We describe the novel abstractions that have been added to Kokkos version 3 such as hierarchical parallelism, containers, task graphs, and arbitrary-sized atomic operations to prepare for exascale era architectures. We demonstrate the performance of these new features with reproducible benchmarks on CPUs and GPUs.

Benchmarking challenging small variants with linked and long reads

Abstract: Genome in a Bottle benchmarks are widely used to help validate clinical sequencing pipelines and develop variant calling and sequencing methods. Here we use accurate linked and long reads to expand benchmarks in 7 samples to include difficult-to-map regions and segmental duplications that are challenging for short reads. These benchmarks add more than 300,000 SNVs and 50,000 insertions or deletions (indels) and include 16% more exonic variants, many in challenging, clinically relevant genes not covered previously, such as PMS2. For HG002, we include 92% of the autosomal GRCh38 assembly while excluding regions problematic for benchmarking small variants, such as copy number variants, that should not have been in the previous version, which included 85% of GRCh38. It identifies eight times more false negatives in a short read variant call set relative to our previous benchmark. We demonstrate that this benchmark reliably identifies false positives and false negatives across technologies, enabling ongoing methods development.

A review for Ca(OH)2/CaO thermochemical energy storage systems

Abstract: Thermochemical energy storage is an essential component of thermal energy storage, which solves the intermittent and long-term energy storage problems of certain renewable energy sources. The appropriate decomposition temperature, high heat storage capacity of the CaO/Ca(OH)2 system makes it one of the successful thermochemical energy storage materials. To better predict reaction process of the thermochemical heat storage process, and lay a foundation for the application design and control of the thermochemical heat storage, we summarized and analyzed the research on the physical and chemical properties of the materials by scholars in recent years. These studies include the structure changes as the reaction occurs, the specific heat capacity and the energy density under different conditions and chemical reaction directions, reaction enthalpy and the reaction kinetics during dehydration/hydration. However, some inherent problem (e.g. low thermal conductivity, agglomeration and influence of carbon dioxide during the reaction) limit the performance of this material. By adding other materials to calcium hydroxide to prepare composite materials, researchers can solve these problems to a certain extent. Therefore, this paper reviews the research status of such composite materials from the perspective of the types of added materials. Finally, to better apply the CaO/Ca(OH)2 thermal storage system to actual production, the performance of this material in different types of reaction beds is summarized.

Evaluating Spatial Accelerator Architectures with Tiled Matrix-Matrix Multiplication

Abstract: There is a growing interest in custom spatial accelerators for machine learning applications. These accelerators employ a spatial array of processing elements (PEs) interacting via custom buffer hierarchies and networks-on-chip. The efficiency of these accelerators comes from employing optimized dataflow (i.e., spatial/temporal partitioning of data across the PEs and fine-grained scheduling) strategies to optimize data reuse. The focus of this work is to evaluate these accelerator architectures using a tiled general matrix-matrix multiplication (GEMM) kernel. To do so, we develop a framework that finds optimized mappings (dataflow and tile sizes) for a tiled GEMM for a given spatial accelerator and workload combination, leveraging an analytical cost model for runtime and energy. Our evaluations over five spatial accelerators demonstrate that the tiled GEMM mappings systematically generated by our framework achieve high performance on various GEMM workloads and accelerators.

Neuromorphic scaling advantages for energy-efficient random walk computations

Abstract: Computing stands to be radically improved by neuromorphic computing (NMC) approaches inspired by the brain's incredible efficiency and capabilities. Most NMC research, which aims to replicate the brain's computational structure and architecture in man-made hardware, has focused on artificial intelligence; however, less explored is whether this brain-inspired hardware can provide value beyond cognitive tasks. We demonstrate that high-degree parallelism and configurability of spiking neuromorphic architectures makes them well-suited to implement random walks via discrete time Markov chains. Such random walks are useful in Monte Carlo methods, which represent a fundamental computational tool for solving a wide range of numerical computing tasks. Additionally, we show how the mathematical basis for a probabilistic solution involving a class of stochastic differential equations can leverage those simulations to provide solutions for a range of broadly applicable computational tasks. Despite being in an early development stage, we find that NMC platforms, at a sufficient scale, can drastically reduce the energy demands of high-performance computing (HPC) platforms.

An All-Dielectric Polaritonic Metasurface with a Giant Nonlinear Optical Response

Abstract: Enhancing the efficiency of second-harmonic generation using all-dielectric metasurfaces to date has mostly focused on electromagnetic engineering of optical modes in the meta-atom. Further advances in nonlinear conversion efficiencies can be gained by engineering the material nonlinearities at the nanoscale, however this cannot be achieved using conventional materials. Semiconductor heterostructures that support resonant nonlinearities using quantum engineered intersubband transitions can provide this new degree of freedom. By simultaneously optimizing the heterostructures and meta-atoms, we experimentally realize an all-dielectric polaritonic metasurface with a maximum second-harmonic generation power conversion factor of 0.5 mW/W2 and power conversion efficiencies of 0.015% at nominal pump intensities of 11 kW/cm2. These conversion efficiencies are higher than the record values reported to date in all-dielectric nonlinear metasurfaces but with 3 orders of magnitude lower pump power. Our results therefore open a new direction for designing efficient nonlinear all-dielectric metasurfaces for new classical and quantum light sources.

Cascaded Optical Nonlinearities in Dielectric Metasurfaces

Abstract: Since the discovery of the laser, optical nonlinearities have been at the core of efficient light conversion sources. Typically, thick transparent crystals or quasi-phase matched waveguides are utilized in conjunction with phase-matching techniques to select a single parametric process. In recent years, due to the rapid developments in artificially structured materials, optical frequency mixing has been achieved at the nanoscale in subwavelength resonators arrayed as metasurfaces. Phase matching becomes relaxed for these wavelength-scale structures, and all allowed nonlinear processes can, in principle, occur on an equal footing. This could promote harmonic generation via a cascaded (consisting of several frequency mixing steps) process. However, so far, all reported work on dielectric metasurfaces have assumed frequency mixing from a direct (single step) nonlinear process. In this work, we prove the existence of cascaded second-order optical nonlinearities by analyzing the second- and third-wave mixing from a highly nonlinear metasurface in conjunction with polarization selection rules and crystal symmetries. We find that the third-wave mixing signal from a cascaded process can be of comparable strength to that from conventional third-harmonic generation and that surface nonlinearities are the dominant mechanism that contributes to cascaded second-order nonlinearities in our metasurface.

SPEARS: A Database-Invariant Spectral Modeling API

Abstract: The Spectral Physics Environment for Advanced Remote Sensing (SPEARS) application programming interface (API) is a Python-based, line-by-line, local thermal equilibrium (LTE) spectral modeling code which is optimized for simultaneously synthesizing optical spectra from any combination of fundamental spectroscopic databases. In this article, we contribute two novel spectral modeling techniques to the scientific literature. First we describe how SPEARS integrates a physics-based collisional model for calculating pressure broadening in the absence of available broadening coefficients. With this collisional model implementation, a generalized approach to fundamental spectroscopic databases can be achieved across multiple databases. We also detail our adaptive grid mesh algorithm developed to make the code scalable for simulating large spectral bandwidths at high spectral fidelity using intuitive grid parameters. We present comparisons to other modeling tools, experiments, and provide a discussion on the SPEARS user interface.

Loss of CD47 alters CD8+ T cell activation <i>in vitro</i> and immunodynamics in mice

Abstract: CD47 has established roles in the immune system for regulating macrophage phagocytosis and lymphocyte activation, with growing evidence of its cell-intrinsic regulatory roles in natural killer and CD8+ T cells. CD47 limits antigen-dependent cytotoxic activities of human and murine CD8+ T cells, but its role in T cell activation kinetics remains unclear. Using in vitro and in vivo models, we show here that CD47 differentially regulates CD8+ T cell responses to short- versus long-term activation. Although CD47 was not required for T cell development in mice and early activation in vitro, short-term stimuli elevated pathogen-reactive gene expression and enhanced proliferation and the effector phenotypes of Cd47-deficient relative to Cd47-sufficient CD8+ T cells. In contrast, persistent TCR stimulation limited the effector phenotypes of Cd47-/- CD8+ T cells and enhanced their apoptosis signature. CD8+ T cell expansion and activation in vivo induced by acute lymphocytic choriomeningitis virus (LCMV) infection did not differ in the absence of CD47. However, the frequency and effector phenotypes of Cd47-/- CD8+ T cells were constrained in chronic LCMV-infected as well as in mice bearing B16 melanoma tumors. Therefore, CD47 regulates CD8+ T cell activation, proliferation, and fitness in a context-dependent manner.