SEARCH · Engineering Papers
Results for “compute in memory”
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.
Tiling Framework for Heterogeneous Computing of Matrix based Tiled Algorithms
Tiling matrix operations can improve the load balancing and performance of applications on heterogeneous computing resources. Writing a tile-based algorithm for each operation with a traditional, hand-tuned tiling approach that uses for loops in C/C++ is cumbersome and error prone. Moreover, it must enable and support the heterogeneous memory management of data objects and also explore architecture-supported, native, tiled-data transfer APIs instead of copying the tiled data to continuous memory before the data transfer. The tiling framework provides a tiled data structure for heterogeneous memory mapping and parameterization to a heterogeneous task specification API. We have integrated our tiled framework into MatRIS (Math kernels library using IRIS). IRIS is a heterogeneous run-time framework with a heterogeneous programming model, memory model, and task execution model. Experiments reveal that the tiled framework for BLAS operations has improved the programmability of tiled BLAS and improved performance by ~20% when compared against the traditional method that copies the data to continuous memory locations for heterogeneous computing.
Development of a military-specific mesh-type computational phantom library and its application to internal dosimetry and in-field radiological triage screening
Estimates of organ-absorbed and committed doses to individuals exposed to radioactive materials via acute inhalation often rely on internal dose coefficients and detector responses from reference human computational models. To achieve more accurate dose assessments to United States Armed Forces service members exposed in-field, computational models with varying morphometric parameters representative of this population are necessary. The International Commission on Radiological Protection (ICRP) Publication 145 provides detailed mesh reference computational phantoms (MRCPs) for adult males and females, with morphometric parameters matched to the 50th percentile. Previously, these phantoms were 2D and 3D scaled to match desired height, mass, and secondary anthropomorphic parameters in the creation of the University of Florida / Memorial Sloan Kettering (UF/MSK) computational phantom library. To achieve body fat percentage targets required for accession into the US Armed Forces, muscle and fat volumes were adjusted accordingly, thus, creating the UF/Department of Defence computational phantom library presented in this study. A comprehensive library of mesh-type computational human phantoms was created, including 57 adult males and 49 adult females with morphometric parameters aligned with United States Armed Forces service members. Phantoms were restricted to a body mass index between 19 and 27.5, with body fat percentages below 26% for males and 36% for females. Specific absorbed fractions were computed for selected source and target combinations, demonstrating how variations in height and body mass influence energy absorption in target regions relative to the ICRP MRCPs. Radiation detector responses were also computed, revealing that higher body masses resulted in decreased registered counts in the detection volume. These findings highlight the importance of incorporating morphometric variability in computational phantoms to achieve more accurate dose assessments and radiation detection responses for United States Armed Forces service members who inhale radioactive materials in-field.
Design and Performance of Kokkos Staging Space toward Scalable Resilient Application Couplings
With the growing number of applications designed for heterogeneous HPC devices, application programmers and users are finding it challenging to compose scalable workflows as ensembles of these applications, that are portable, performant and resilient. The Kokkos C++ library has been designed to simplify this cumbersome procedure by providing an intra-application uniform programming model and portable performance. However, assembling multiple Kokkos-enabled applications into a complex workflow is still a challenge. Although Kokkos enables a uniform programming model, the inter-application data exchange still remains a challenge from both performance and software development cost perspectives. In order to address this issue, we propose Kokkos data staging memory space, an extension of Kokkos' data abstraction (memory space) for heterogeneous computing systems. This new abstraction allows to express data on a virtual shared-space for multiple Kokkos applications, thus extending Kokkos to support inter-application data exchange to build an efficient application workflow. Additionally, we study the effectiveness of asynchronous data layout conversions for applications requiring different memory access patterns for the shared data. Our preliminary evaluation with a synthetic benchmark indicate the effectiveness of this conversion adapted to three different scenarios representing access frequency and use patterns of the shared data.
Fast Parallel Tensor Times Same Vector for Hypergraphs
Hypergraphs are a popular paradigm to rep- resent complex real-world networks exhibiting multi-way relationships of varying sizes. Mining centrality in hyper- graphs via symmetric adjacency tensors has only recently become computationally feasible for large and complex datasets. To enable scalable computation of these and related hypergraph analytics, here we focus on the Sparse Symmetric Tensor Times Same Vector (S3TTVC) oper- ation. We introduce the Compound Compressed Sparse Symmetric (CCSS) format, an extension of the compact CSS format for hypergraphs of varying hyperedge sizes and present a shared-memory parallel algorithm to compute S3TTVC. We experimentally show S3TTVC computation using the CCSS format achieves better performance than the naive baseline, and is subsequently more performant for hypergraph H-eigenvector centrality.
Distributed-Memory Sparse Deep Neural Network Inference Using Global Arrays
Partitioned Global Address Space (PGAS) models exhibit tremendous promise in developing efficient and productive distributed-memory parallel applications. They have been used extensively in scientific computations due to conveniently offering a ``shared-memory''-like model and convenient interfaces that separate communication with synchronization. Traditionally, PGAS communication models have been applied to dense/contiguously distributed data, but most modern applications depict varied levels of sparsity. Existing PGAS models require certain adaptations to support distributed sparse computations, since associated computations often require matrix arithmetic, in addition to data movement. The Global Arrays toolkit from Pacific Northwest National Laboratory (PNNL) is one of the earliest PGAS models to combine one-sided data communication and distributed matrix operations and is still used in the popular NWChem quantum chemistry suite. Recently, we have expanded the Global Arrays toolkit to support common sparse operations, like sparse matrix-dense matrix multiplies (SpMM), sparse matrix-sparse matrix multiplication (SpGEMM) and Sampled Dense-Dense Matrix Multiplication (SDDMM). As it turns out, these operations are the bedrock of sparse Deep Learning (DL); sparse deep neural networks and Graph Neural Networks (GNNs) have gained increasing attention recently in achieving speedups on training and inference with reduced memory footprints. Unlike scientific applications in High Performance Computing (HPC), modern (distributed-memory capable) DL toolkits often rely on non-standardized and closed-source vendor software optimizations, creating challenges in software-hardware co-design at scale. Our goal is to support a variety of distributed-memory sparse matrix operations and helper functions in the newly created Sparse Global Arrays (SGA), such that it is possible to build portable and productive Machine Learning scenarios for algorithm/software and hardware codesign purposes. Contemporary data-parallel schemes for training/inference are undergoing a major overhaul since model replication limits scalability and causes resource inefficiencies. As such, we have adopted tensor parallelism in decomposing the model and inputs, to mitigate memory issues. Current implementation is built on top of MPI and uses CPUs to maximize the portability across the platforms.
Tula: Optimizing Time, Cost, and Generalization in Distributed Large-Batch Training
Distributed training increases the number of batches processed per iteration either by scaling-out (adding more nodes) or scaling-up (increasing the batch-size). However, the largest configuration does not necessarily yield the best performance. Horizontal scaling introduces additional communication overhead, while vertical scaling is constrained by computation cost and device memory limits. Thus, simply increasing the batch-size leads to diminishing returns: training time and cost decrease initially but eventually plateaus, creating a knee-point in the time/cost vs. batch-size pareto curve. The optimal batch-size therefore depends on the underlying model, data and available compute resources. Large batches also suffer from worse model quality due to the well-known “generalization gap”. In this paper, we present Tula, an online service that automatically optimizes time, cost, and convergence quality for large-batch training of convolutional models. It combines parallel-systems modeling with statistical performance prediction to identify the optimal batchsize. Tula predicts training time and cost within 7.5−14% error across multiple models, and achieves up to 20× overall speedup and improves test accuracy by ≈9% on average over standard large-batch training on various vision tasks, thus successfully mitigating the generalization gap and accelerating training at the same time.
Generating An Advanced Cross-section Library For HTGR Pebble Bed Depletion Calculations Using Reduced-Order Model Generation Techniques
For code development, Advanced Reactor Technologies - Gas Cooled Reactors Program (ART-GCR) rely on a collaboration with the Nuclear Energy Advanced Modeling and Simulation (NEAMS) program, but the cross sections generation and the methodology definition is part of this program area goals. Based on previous studies in FY23, the size of microscopic cross section libraries increases rapidly with the number of tabulations, requiring significant amount of memory and drastically slowing down the Griffin calculations when evaluating cross sections via the multivariate linear interpolation approach. Rising to these challenges, this work investigates constructing Reduced-order Models (ROMs) for the multi-group microscopic cross sections to accelerate the cross section evaluation in Griffin. A database of multigroup cross sections is first collected considering all possible parameters that a designer could change for optimization. Down-selection of the ROM techniques afterward shows Deep Neural Network (DNN) as the best candidate when jointly consider memory efficiency, predictive accuracy, computational cost, scalability, flexibility and ease of implementation of the algorithms in comparison to the multidimensional interpolation. This work develops a specific interface that enables the cross section predictions using pre-trained DNN models into Griffin leveraging the existing ROM capabilities. DNNs have been trained for all isotopes for use in Griffin. Preliminary Griffin testing shows that DNNs exhibit exceptional predictive accuracy and the use of DNNs provides orders of magnitude improvement in memory efficiency compared to conventional interpolation techniques. With such ROM techniques, it holds great promise to further increase the fidelity of the Pebble Bed Reactor (PBR) simulation by increasing the number of tabulations/state variables during cross section evaluation, while maintaining the computational cost affordable in Griffin.
Suppressing Structural Relaxation in Nanoscale Antimony to Enable Ultralow‐Drift Phase‐Change Memory Applications
Abstract Phase‐change random‐access memory (PCRAM) devices suffer from pronounced resistance drift originating from considerable structural relaxation of phase‐change materials (PCMs), which hinders current developments of high‐capacity memory and high‐parallelism computing that both need reliable multibit programming. This work realizes that compositional simplification and geometrical miniaturization of traditional GeSbTe‐like PCMs are feasible routes to suppress relaxation. While to date, the aging mechanisms of the simplest PCM, Sb, at nanoscale, have not yet been unveiled. Here, this work demonstrates that in an optimal thickness of only 4 nm, the thin Sb film can enable a precise multilevel programming with ultralow resistance drift coefficients, in a regime of ≈10 −4 –10 −3 . This advancement is mainly owed to the slightly changed Peierls distortion in Sb and the less‐distorted octahedral‐like atomic configurations across the Sb/SiO 2 interfaces. This work highlights a new indispensable approach, interfacial regulation of nanoscale PCMs, for pursuing ultimately reliable resistance control in aggressively‐miniaturized PCRAM devices, to boost the storage and computing efficiencies substantially.
UNITY: Unified Memory and Storage Space
UNITY is a 36-month project focused on providing design and evaluate a new distributed storage paradigm that unifies the traditionally distinct application views of memory- and file-based data storage into a single scalable and resilient environment. The project is a collaboration among Oak Ridge National Laboratory (ORNL, Lead institution), Los Alamos National Labs (LANL), and Georgia Tech (GT). The main contributions of the GT team have been around development of low level systems software for best leveraging the capabilities of new types of persistent memory technologies, and for development of methods for intelligent data management across memory/storage substrates with heterogeneous components. GT contributed the Phoenix library for optimized checkpoint/restart for HPC I/O for systems with non-volatile memory (NVM), the NVStream library for NVM-specialized streaming I/O for HPC workflows, the CoMerge, Mnemo and Kleio solutions for intelligent data management on NVM-based systems. These contributions result in significant improvements in both application performance and system efficiency.
Complexity-calibrated benchmarks for machine learning reveal when prediction algorithms succeed and mislead
Abstract Recurrent neural networks are used to forecast time series in finance, climate, language, and from many other domains. Reservoir computers are a particularly easily trainable form of recurrent neural network. Recently, a “next-generation” reservoir computer was introduced in which the memory trace involves only a finite number of previous symbols. We explore the inherent limitations of finite-past memory traces in this intriguing proposal. A lower bound from Fano’s inequality shows that, on highly non-Markovian processes generated by large probabilistic state machines, next-generation reservoir computers with reasonably long memory traces have an error probability that is at least $$\sim 60\%$$ ∼ 60 % higher than the minimal attainable error probability in predicting the next observation. More generally, it appears that popular recurrent neural networks fall far short of optimally predicting such complex processes. These results highlight the need for a new generation of optimized recurrent neural network architectures. Alongside this finding, we present concentration-of-measure results for randomly-generated but complex processes. One conclusion is that large probabilistic state machines—specifically, large $$\epsilon$$ ϵ -machines—are key to generating challenging and structurally-unbiased stimuli for ground-truthing recurrent neural network architectures.
Performance Characteristics of the BlueField-2 SmartNIC
High-performance computing (HPC) researchers have long envisioned scenarios where application workflows could be improved through the use of programmable processing elements embedded in the network fabric. Recently, vendors have introduced programmable Smart Network Interface Cards (SmartNICs) that enable computations to be offloaded to the edge of the network. There is great interest in both the HPC and high-performance data analytics (HPDA) communities in understanding the roles these devices may play in the data paths of upcoming systems. This paper focuses on characterizing both the networking and computing aspects of NVIDIA’s new BlueField-2 SmartNIC when used in a 100Gb/s Ethernet environment. For the networking evaluation we conducted multiple transfer experiments between processors located at the host, the SmartNIC, and a remote host. These tests illuminate how much effort is required to saturate the network and help estimate the processing headroom available on the SmartNIC during transfers. For the computing evaluation we used the stress-ng benchmark to compare the BlueField-2 to other servers and place realistic bounds on the types of offload operations that are appropriate for the hardware. Our findings from this work indicate that while the BlueField-2 provides a flexible means of processing data at the network’s edge, great care must be taken to not overwhelm the hardware. While the host can easily saturate the network link, the SmartNIC’s embedded processors may not have enough computing resources to sustain more than half the expected bandwidth when using kernel-space packet processing. From a computational perspective, encryption operations, memory operations under contention, and on-card IPC operations on the SmartNIC perform significantly better than the general-purpose servers used for comparisons in our experiments. Therefore, applications that mainly focus on these operations may be good candidates for offloading to the SmartNIC.
Temperature effects in spin-dependent Hall currents in an ideal skyrmion gas
The quest of novel reliable and fast performing logic and memory elements in classical and quantum computing requires discoveries of new effects in novel quantum materials. Skyrmions are such magnetic textures where electron scattering can essentially change the shape and location of the skyrmion, which can be used as a memory element being more efficient than domain walls as memory device in classical computing. Because the skyrmion motion is sensitive even to very small currents, we study electron scattering by skyrmions in an ideal skyrmion gas in ferromagnetic environment. In such systems, the direct and Hall currents become spin-dependent. For applications it is important to consider a Hall effect in the whole range of temperatures under the assumption of the skyrmion existence. In this study we find the nonmonotonic temperature dependence of the direct spin-up conductivity, i. e., the conductivity for the current directed along the applied electric field due to the electrons with the spin parallel to the ferromagnetic moment. Such a behavior contradicts the traditional understanding where the temperature only increases the value of the conductivity. The spin-down Hall conductivity is found to be even more dramatic exhibiting the conductivity sign change (i.e., the change in the current direction) with temperature for small skyrmion sizes. The found effects strongly depend on Fermi energy. The most pronounced dependencies take place if the Fermi energy is slightly below and above the bottom of the upper (spin-down) energy band of an ideal 2D electron gas. In addition, we also find that the direct and Hall resistivities, ρ xx and ρ xy , are independent of the exchange integral J for 2J > ε F .
Single-Cell Universal Logic-in-Memory Using 2T-nC FeRAM: An Area and Energy-Efficient Approach for Bulk Bitwise Computation
This work presents a novel approach to configure 2T-nC ferroelectric RAM (FeRAM) for performing single cell logic-in-memory operations, highlighting its advantages in energy-efficient computation over conventional DRAM-based approaches. Unlike conventional 1T-1C dynamic RAM (DRAM), which incurs refresh overhead, 2T-nC FeRAM offers a promising alternative as a non-volatile memory solution with low energy consumption. Our key findings include the potential of quasi-nondestructive readout (QNRO) sensing in 2T-nC FeRAM for logic-in-memory (LiM) applications, demonstrating its inherent capability to perform inverting logic without requiring external modifications, a feature absent in traditional 1T-1C DRAM. We successfully implement the MINORITY function within a single cell of 2T-nC FeRAM, enabling universal NAND and NOR logic, validated through SPICE simulations and experimental data. Additionally, the research investigates the feasibility of 3D integration with 2T-nC FeRAM, showing substantial improvements in storage and computational density, facilitating bulk-bitwise computation. Our evaluation of eight real-world, data-intensive applications reveals that 2T-nC FeRAM achieves 2× higher performance and 2.5× lower energy consumption compared to DRAM. Furthermore, the thermal stability of stacked 2T-nC FeRAM is validated, confirming its reliable operation when integrated on a compute die. These findings emphasize the advantages of 2T-nC FeRAM for LiM, offering superior performance and energy efficiency over conventional DRAM.
System and method for controlling an energy beam of an additive manufacturing system
An additive manufacturing system includes an energy beam generator that generates an energy beam to melt and fuse raw materials for a part and a computing device that controls operation of the energy beam generator. The computing device includes a memory element that stores or access a three-dimensional model of the part, and a processing element that receives at least a portion of the three-dimensional model and controls a parameter of the energy beam generator according to a gradient function so as to apply a variable amount of heat to the raw material that forms an overhang on the part.
A scalable superconducting nanowire memory array with row–column addressing
Scalable superconducting memory is required for the development of low-energy superconducting computers and fault-tolerant quantum computers. Conventional superconducting logic-based memory cells possess a large footprint that limits scaling; nanowire-based superconducting memory cells, although more compact, have high error rates, which hinders integration into large arrays. Here we report a 4 × 4 superconducting nanowire memory array that is designed for scalable row–column operations and has a functional density of 2.6 Mbit cm −2 . Each memory cell is based on a nanowire loop consisting of two temperature-dependent superconducting switches and a variable kinetic inductor. The arrays operate at 1.3 K, where we implement and characterize multiflux quanta state storage and destructive read-out. By optimizing the write- and read-pulse sequences, we minimize bit errors and maximize operating margins. We achieve a minimum bit error rate of 10 −5 . Here, we also use circuit-level simulations to understand the memory cell’s dynamics, performance limits and stability under varying pulse amplitudes.
Pass-efficient methods for compression of high-dimensional turbulent flow data
The future of high-performance computing, specifically on future Exascale computers, will presumably see memory capacity and bandwidth fail to keep pace with data generated, for instance, from massively parallel partial differential equation (PDE) systems. Current strategies proposed to address this bottleneck entail the omission of large fractions of data, as well as the incorporation of in situ compression algorithms to avoid overuse of memory. To ensure that post-processing operations are successful, this must be done in a way that a sufficiently accurate representation of the solution is stored. Moreover, in situations where the input/output system becomes a bottleneck in analysis, visualization, etc., or the execution of the PDE solver is expensive, the number of passes made over the data must be minimized. In the interest of addressing this problem, this work focuses on the utility of pass-efficient, parallelizable, low-rank, matrix decomposition methods in compressing high-dimensional simulation data from turbulent flows. Additionally, a particular emphasis is placed on using coarse representation of the data – compatible with the PDE discretization grid – to accelerate the construction of the low-rank factorization. This includes the presentation of a novel single-pass matrix decomposition algorithm for computing the so-called interpolative decomposition. The methods are described extensively and numerical experiments on two turbulent channel flow data are performed. In the first (unladen) channel flow case, compression factors exceeding 400 are achieved while maintaining accuracy with respect to first- and second-order flow statistics. In the particle-laden case, compression factors of 100 are achieved and the compressed data is used to recover particle velocities. These results show that these compression methods can enable efficient computation of various quantities of interest in both the carrier and disperse phases.
A high-order computational framework for particle-resolved simulations of disperse multiphase flows
This work presents a high-order numerical approach for particle-resolved simulations of disperse multiphase flows, where the Navier-Stokes equations for fluid flow are solved using a high-order spectral element method in the Eulerian framework, and the particle phase is directly simulated with a discrete element method. The coupling between particles and fluids is explicitly handled using an adapted direct-forcing immersed boundary method. Unlike the conventional schemes, a high-order barycentric Lagrange interpolation method and a Gaussian projection kernel are used to ensure accurate momentum exchange between local boundary points and surrounding fluid nodes in the framework of high-order fluid solver. Benchmark tests of increasing complexity are conducted to demonstrate the accuracy and efficiency of our method. Here, it is found that our approach exhibits an excellent convergence performance, as the fluid element/grid is refined and the number of boundary points increases. Compared to conventional low-order methods, the proposed high-order framework enables the use of substantially larger fluid elements while maintaining high accuracy in modeling fluid-particle interactions, owing to the enhanced resolution of high-order basis functions. Moreover, since the primary unknowns are stored at element or grid nodes, the high-order approach offers improved efficiency in both CPU memory usage and total computational cost.