Search NASA⌕ Search

SEARCH · Search NASA

Results for “memory efficiency”

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.

At least 1,459 records · Page 81

MassiveGNN: Efficient Training via Prefetching for Massively Connected Distributed Graphs

Graph Neural Networks (GNN) are indispensable in learning from graph-structured data, yet their rising computational costs, especially on massively connected graphs, pose significant challenges in terms of execution performance. To tackle this, distributed-memory solutions such as partitioning the graph to concurrently train multiple replicas of GNNs are in practice. However, approaches requiring a partitioned graph usually suffer from communication overhead and load imbalance, even under optimal partitioning and communication strategies due to irregularities in the neighborhood minibatch sampling. This paper proposes practical trade-offs for improving the sampling and communication overheads for representation learn- ing on distributed graphs (using popular GraphSAGE architecture) by developing a parameterized prefetch and eviction scheme on top of the state-of-the-art Amazon DistDGL distributed GNN framework, demonstrating about 15–40% improvement in end-to-end training performance on the NERSC Perlmutter supercomputer for various OGB datasets.

Machine Leanring, high performance comptuing, grap↗

Some Advanced Concepts in Discrete Aerodynamic Sensitivity Analysis

An efficient incremental iterative approach for differentiating advanced flow codes is successfully demonstrated on a two-dimensional inviscid model problem. The method employs the reverse-mode capability of the automatic differentiation software tool ADIFOR 3.0 and is proven to yield accurate first-order aerodynamic sensitivity derivatives. A substantial reduction in CPU time and computer memory is demonstrated in comparison with results from a straightforward, black-box reverse-mode applicaiton of ADIFOR 3.0 to the same flow code. An ADIFOR-assisted procedure for accurate second-rder aerodynamic sensitivity derivatives is successfully verified on an inviscid transonic lifting airfoil example problem. The method requires that first-order derivatives are calculated first using both the forward (direct) and reverse (adjoinct) procedures; then, a very efficient noniterative calculation of all second-order derivatives can be accomplished. Accurate second derivatives (i.e., the complete Hesian matrices) of lift, wave drag, and pitching-moment coefficients are calculated with respect to geometric shape, angle of attack, and freestream Mach number.

Taylor, Arthur C., III↗

MOSIQS: Persistent Memory Object Storage With Metadata Indexing and Querying for Scientific Computing

Scientific applications often require high-bandwidth shared storage to perform joint simulations and collaborative data analytics. Shared memory pools provide a chance to satisfy such needs. Recently, a high-speed network such as Gen-Z utilizing persistent memory (PM) offers an opportunity to create a shared memory pool connected to compute nodes. However, there are several challenges to use scientific applications on the shared memory pool directly such as scalability, failure-atomicity, and lack of scientific metadata-based search and query. In this paper, we propose MOSIQS, a persistent memory object storage framework with metadata indexing and querying for scientific computing. We design MOSIQS based on the key idea that memory objects on PM pool can live beyond the application lifetime and can become the sharing currency for applications and scientists. MOSIQS provides an aggregate memory pool atop an array of persistent memory devices to store and access memory objects to accelerate scientific computing. MOSIQS uses a lightweight persistent memory key-value store to manage the metadata of memory objects, which enables memory object sharing. To facilitate metadata search and query over millions of memory objects resident on memory pool, we introduce Group Split and Merge (GSM), a novel persistent index data structure designed primarily for scientific datasets. GSM splits and merges dynamically to minimize the query search space and maintains low query processing time while overcoming the index storage overhead. MOSIQS is implemented on top of PMDK. We evaluate the proposed approach on many-core server with an array of real PM devices. Experimental results show that MOSIQS gains a 100% write performance improvement and executes multi-attribute queries efficiently with 2.7× less index storage overhead offering significant potential to speed up scientific computing applications.

97 MATHEMATICS AND COMPUTING↗

Electrically programmable-erasable In-Ga-Zn-O thin-film transistor memory with atomic-layer-deposited Al{sub 2}O{sub 3}/Pt nanocrystals/Al{sub 2}O{sub 3} gate stack

Amorphous indium-gallium-zinc oxide (a-IGZO) thin-film transistor (TFT) memory is very promising for transparent and flexible system-on-panel displays; however, electrical erasability has always been a severe challenge for this memory. In this article, we demonstrated successfully an electrically programmable-erasable memory with atomic-layer-deposited Al{sub 2}O{sub 3}/Pt nanocrystals/Al{sub 2}O{sub 3} gate stack under a maximal processing temperature of 300 {sup o}C. As the programming voltage was enhanced from 14 to 19 V for a constant pulse of 0.2 ms, the threshold voltage shift increased significantly from 0.89 to 4.67 V. When the programmed device was subjected to an appropriate pulse under negative gate bias, it could return to the original state with a superior erasing efficiency. The above phenomena could be attributed to Fowler-Nordheim tunnelling of electrons from the IGZO channel to the Pt nanocrystals during programming, and inverse tunnelling of the trapped electrons during erasing. In terms of 0.2-ms programming at 16 V and 350-ms erasing at −17 V, a large memory window of 3.03 V was achieved successfully. Furthermore, the memory exhibited stable repeated programming/erasing (P/E) characteristics and good data retention, i.e., for 2-ms programming at 14 V and 250-ms erasing at −14 V, a memory window of 2.08 V was still maintained after 10{sup 3} P/E cycles, and a memory window of 1.1 V was retained after 10{sup 5} s retention time.

36 MATERIALS SCIENCE↗

Characterization of PIPS detectors for measurement of radioxenon

The International Monitoring System arm of the Comprehensive Nuclear-Test-Ban Treaty Organization utilizes sampling of atmospheric radioxenon to scan the world for nuclear testing events. Current designs utilized in the field involve plastic scintillator cells run in coincidence with Sodium Iodide (NaI) or High Purity Germanium (HPGe) detectors. The use of plastic scintillators as electron detectors exhibit some issues in comparison to other materials. Silicon detectors are analyzed as an alternative to plastic scintillators for beta detection based on their higher energy resolution and lowered memory effect. A radioxenon detection setup was created with the intent of measuring radioxenon samples using coincidence counting between silicon and NaI detectors as well as silicon and HPGe. Finally, the absolute efficiency of both setups is calculated and compared.

38 RADIATION CHEMISTRY, RADIOCHEMISTRY, AND NUCLEA↗

A Diffusion‐Based Uncertainty Quantification Method to Advance E3SM Land Model Calibration

Abstract Calibrating land surface models and accurately quantifying their uncertainty are crucial for improving the reliability of simulations of complex environmental processes. This, in turn, advances our predictive understanding of ecosystems and supports climate‐resilient decision‐making. Traditional calibration methods, however, face challenges of high computational costs and difficulties in accurately quantifying parameter uncertainties. To address these issues, we develop a diffusion‐based uncertainty quantification (DBUQ) method. Unlike conventional generative diffusion methods, which are computationally expensive and memory‐intensive, DBUQ innovates by formulating a parameterized generative model and approximates this model through supervised learning, which enables quick generation of parameter posterior samples to quantify its uncertainty. DBUQ is effective, efficient, and general‐purpose, making it suitable for site‐specific ecosystem model calibration and broadly applicable for parameter uncertainty quantification across various earth system models. In this study, we applied DBUQ to calibrate the Energy Exascale Earth System Model land model at the Missouri Ozark AmeriFlux forest site. Results indicated that DBUQ produced accurate parameter posterior distributions similar to those from Markov Chain Monte Carlo sampling but with 30 times less computing time. This significant improvement in efficiency suggests that DBUQ can enable rapid, site‐level model calibration at a global scale, enhancing our predictive understanding of climate impacts on terrestrial ecosystems.

54 ENVIRONMENTAL SCIENCES↗

Efficient generation of universal two-dimensional cluster states with hybrid systems

We present a scheme to generate any two-dimensional cluster state efficiently. The number of the basic gate (entangler) operations is on the order of the entanglement bonds of a cluster state. This deterministic scheme, which demands few ancilla resources and no quantum memory, is suitable for large-scale quantum computation.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

A Co-Design Study Of Fusion Whole Device Modeling Using Code Coupling

Complex workflows consisting of multiple simulation and analysis codes running concurrently through in-memory coupling is becoming popular due to inherent advantages in online management of large-scale data, resilience, and the code development process. However, orchestrating such a multi-application workflow to efficiently utilize resources on a heterogeneous architecture is challenging. In this paper, we present our results with running the Fusion Whole Device Modeling benchmark workflow on Summit, a pre-exascale supercomputer at Oak Ridge National Laboratory. We explore various resource distribution and process placement mechanisms, including sharing compute nodes between processes from separate applications. We show that fine-grained process placement can have a significant impact towards efficient utilization of the compute power of a node on Summit, and conclude that sophisticated tools for performing co-design studies of multi-application workflows can play an important role towards efficient orchestration of such workflows.

Choi, Jong Youl↗

Sparse Binary Matrix-Vector Multiplication on Neuromorphic Computers

Neuromorphic computers offer the opportunity for low-power, efficient computation. Though they have been primarily applied to neural network tasks, there is also the opportunity to leverage the inherent characteristics of neuromorphic computers (low power, massive parallelism, collocated processing and memory) to perform non-neural network tasks. Here, we demonstrate how an approach for performing sparse binary matrix-vector multiplication on neuromorphic computers. We describe the approach, which relies on the connection between binary matrix-vector multiplication and breadth first search, and we introduce the algorithm for performing this calculation in a neuromorphic way. We validate the approach in simulation. Finally, we provide a discussion of the runtime of this algorithm and discuss where neuromorphic computers in the future may have a computational advantage when performing this computation.

Schuman, Catherine↗

The PetscSF Scalable Communication Layer

PetscSF, the communication component of the Portable, Extensible Toolkit for Scientific Computation (PETSc), is designed to provide PETSc's communication infrastructure suitable for exascale computers that utilize GPUs and other accelerators. PetscSF provides a simple application programming interface (API) for managing common communication patterns in scientific computations by using a star-forest graph representation. PetscSF supports several implementations based on MPI and NVSHMEM, whose selection is based on the characteristics of the application or the target architecture. An efficient and portable model for network and intra-node communication is essential for implementing large-scale applications. The Message Passing Interface, which has been the de facto standard for distributed memory systems, has developed into a large complex API that does not yet provide high performance on the emerging heterogeneous CPU-GPU-based exascale systems. Here, we discuss the design of PetscSF, how it can overcome some difficulties of working directly with MPI on GPUs, and we demonstrate its performance, scalability, and novel features.

97 MATHEMATICS AND COMPUTING↗

Quantifying NUMA and Contention Effects in Multi-GPU Systems

As system architects strive for increased density and power efficiency, the traditional compute node is being augmented with an increasing number of graphics processing units (GPUs). The integration of multiple GPUs per node introduces complex performance phenomena including non-uniform memory access (NUMA) and contention for shared system resources. Utilizing the Keeneland system, this paper quantifies these effects and presents some guidance on programming strategies to maximize performance in multi-GPU environments.

Spafford, Kyle L↗

EXAGRAPH: Graph and combinatorial methods for enabling exascale applications

Combinatorial algorithms in general and graph algorithms in particular play a critical enabling role in numerous scientific applications. However, the irregular memory access nature of these algorithms makes them one of the hardest algorithmic kernels to implement on parallel systems. With tens of billions of hardware threads and deep memory hierarchies, the exascale computing systems in particular pose extreme challenges in scaling graph algorithms. The codesign center on combinatorial algorithms, ExaGraph, was established to design and develop methods and techniques for efficient implementation of key combinatorial (graph) algorithms chosen from a diverse set of exascale applications. Algebraic and combinatorial methods have a complementary role in the advancement of computational science and engineering, including playing an enabling role on each other. In this paper, we survey the algorithmic and software development activities performed under the auspices of ExaGraph from both a combinatorial and an algebraic perspective. In particular, we detail our recent efforts in porting the algorithms to manycore accelerator (GPU) architectures. We also provide a brief survey of the applications that have benefited from the scalable implementations of different combinatorial algorithms to enable scientific discovery at scale. We believe that several applications will benefit from the algorithmic and software tools developed by the ExaGraph team.

97 MATHEMATICS AND COMPUTING↗

A variable rate speech compressor for mobile applications

One of the most promising speech coder at the bit rate of 9.6 to 4.8 kbits/s is CELP. Code Excited Linear Prediction (CELP) has been dominating 9.6 to 4.8 kbits/s region during the past 3 to 4 years. Its set back however, is its expensive implementation. As an alternative to CELP, the Base-Band CELP (CELP-BB) was developed which produced good quality speech comparable to CELP and a single chip implementable complexity as reported previously. Its robustness was also improved to tolerate errors up to 1.0 pct. and maintain intelligibility up to 5.0 pct. and more. Although, CELP-BB produces good quality speech at around 4.8 kbits/s, it has a fundamental problem when updating the pitch filter memory. A sub-optimal solution is proposed for this problem. Below 4.8 kbits/s, however, CELP-BB suffers from noticeable quantization noise as a result of the large vector dimensions used. Efficient representation of speech below 4.8 kbits/s is reported by introducing Sinusoidal Transform Coding (STC) to represent the LPC excitation which is called Sine Wave Excited LPC (SWELP). In this case, natural sounding good quality synthetic speech is obtained at around 2.4 kbits/s.

Yeldener, S.↗

Current capabilities for simulating the extreme distortion of thin structures subjected to severe impacts

The explicit transient dynamics technology in use today for simulating the impact and subsequent transient dynamic response of a structure has its origins in the 'hydrocodes' dating back to the late 1940's. The growth in capability in explicit transient dynamics technology parallels the growth in speed and size of digital computers. Computer software for simulating the explicit transient dynamic response of a structure is characterized by algorithms that use a large number of small steps. In explicit transient dynamics software there is a significant emphasis on speed and simplicity. The finite element technology used to generate the spatial discretization of a structure is based on a compromise between completeness of the representation for the physical processes modelled and speed in execution. That is, since it is expected in every calculation that the deformation will be finite and the material will be strained beyond the elastic range, the geometry and the associated gradient operators must be reconstructed, as well as complex stress-strain models evaluated at every time step. As a result, finite elements derived for explicit transient dynamics software use the simplest and barest constructions possible for computational efficiency while retaining an essential representation of the physical behavior. The best example of this technology is the four-node bending quadrilateral derived by Belytschko, Lin and Tsay. Today, the speed, memory capacity and availability of computer hardware allows a number of the previously used algorithms to be 'improved.' That is, it is possible with today's computing hardware to modify many of the standard algorithms to improve their representation of the physical process at the expense of added complexity and computational effort. The purpose is to review a number of these algorithms and identify the improvements possible. In many instances, both the older, faster version of the algorithm and the improved and somewhat slower version of the algorithm are found implemented together in software. Specifically, the following seven algorithmic items are examined: the invariant time derivatives of stress used in material models expressed in rate form; incremental objectivity and strain used in the numerical integration of the material models; the use of one-point element integration versus mean quadrature; shell elements used to represent the behavior of thin structural components; beam elements based on stress-resultant plasticity versus cross-section integration; the fidelity of elastic-plastic material models in their representation of ductile metals; and the use of Courant subcycling to reduce computational effort.

Key, Samuel W.↗

Efficient Calculation of a Jitter/Stability Metric

A tool for computing a jitter/stability metric used in NASA requirements statements is developed. An efficient algorithm is given for computing this metric. Two ways of implementing it on a computer are discussed. One is optimized for computational speed while the other sacrifices some speed to conserve memory. Timing studies are given to show that the improvement of computation times using the present algorithm over previously existing techniques can run to several orders of magnitude, and that previous techniques were so costly that the present algorithm represents enabling technology. Further comparisons show that the memory conservative implementation runs at about half the speed of the fast implementation, but can cut the major data storage requirement of the fast implementation by 95-99%, making the algorithm implementable on much smaller computers, such as PC's, than it would be otherwise. Software for both implementations is included in version 2 of the NASA time and frequency domain analysis program PLATSIM.

Giesy, Daniel P.↗

RACER: Effective Race Detection Using AspectJ

The limits of coding with joint constraints on detected and undetected error rates Programming errors occur frequently in large software systems, and even more so if these systems are concurrent. In the past, researchers have developed specialized programs to aid programmers detecting concurrent programming errors such as deadlocks, livelocks, starvation and data races. In this work we propose a language extension to the aspect-oriented programming language AspectJ, in the form of three new built-in pointcuts, lock(), unlock() and may be Shared(), which allow programmers to monitor program events where locks are granted or handed back, and where values are accessed that may be shared amongst multiple Java threads. We decide thread-locality using a static thread-local objects analysis developed by others. Using the three new primitive pointcuts, researchers can directly implement efficient monitoring algorithms to detect concurrent programming errors online. As an example, we expose a new algorithm which we call RACER, an adoption of the well-known ERASER algorithm to the memory model of Java. We implemented the new pointcuts as an extension to the Aspect Bench Compiler, implemented the RACER algorithm using this language extension and then applied the algorithm to the NASA K9 Rover Executive. Our experiments proved our implementation very effective. In the Rover Executive RACER finds 70 data races. Only one of these races was previously known.We further applied the algorithm to two other multi-threaded programs written by Computer Science researchers, in which we found races as well.

data analysis↗

Roofline Analysis in the Intel® Advisor to Deliver Optimized Performance for applications on Intel® Xeon Phi™ Processor

In this session we show, in two case studies, how the roofline feature of Intel Advisor has been utilized to optimize the performance of kernels of the XGC1 and PICSAR codes in preparation for Intel Knights Landing architecture. The impact of the implemented optimizations and the benefits of using the automatic roofline feature of Intel Advisor to study performance of large applications will be presented. This demonstrates an effective optimization strategy that has enabled these science applications to achieve up to 4.6 times speed-up and prepare for future exascale architectures. # Goal/Relevance of Session The roofline model [1,2] is a powerful tool for analyzing the performance of applications with respect to the theoretical peak achievable on a given computer architecture. It allows one to graphically represent the performance of an application in terms of operational intensity, i.e. the ratio of flops performed and bytes moved from memory in order to guide optimization efforts. Given the scale and complexity of modern science applications, it can often be a tedious task for the user to perform the analysis on the level of functions or loops to identify where performance gains can be made. With new Intel tools, it is now possible to automate this task, as well as base the estimates of peak performance on measurements rather than vendor specifications. The goal of this session is to demonstrate how the roofline feature of Intel Advisor can be used to balance memory vs. computation related optimization efforts and effectively identify performance bottlenecks. A series of typical optimization techniques: cache blocking, structure refactoring, data alignment, and vectorization illustrated by the kernel cases will be addressed. # Description of the codes ## XGC1 The XGC1 code [3] is a magnetic fusion Particle-In-Cell code that uses an unstructured mesh for its Poisson solver that allows it to accurately resolve the edge plasma of a magnetic fusion device. After recent optimizations to its collision kernel [4], most of the computing time is spent in the electron push (pushe) kernel, where these optimization efforts have been focused. The kernel code scaled well with MPI+OpenMP but had almost no automatic compiler vectorization, in part due to indirect memory addresses and in part due to low trip counts of low-level loops that would be candidates for vectorization. Particle blocking and sorting have been implemented to increase trip counts of low-level loops and improve memory locality, and OpenMP directives have been added to vectorize compute-intensive loops that were identified by Advisor. The optimizations have improved the performance of the pushe kernel 2x on Haswell processors and 1.7x on KNL. The KNL node-for-node performance has been brought to within 30% of a NERSC Cori phase I Haswell node and we expect to bridge this gap by reducing the memory footprint of compute intensive routines to improve cache reuse. ## PICSAR is a Fortran/Python high-performance Particle-In-Cell library targeting at MIC architectures first designed to be coupled with the PIC code WARP for the simulation of laser-matter interaction and particle accelerators. PICSAR also contains a FORTRAN stand-alone kernel for performance studies and benchmarks. A MPI domain decomposition is used between NUMA domains and a tile decomposition (cache-blocking) handled by OpenMP has been added for shared-memory parallelism and better cache management. The so-called current deposition and field gathering steps that compose the PIC time loop constitute major hotspots that have been rewritten to enable more efficient vectorization. Particle communications between tiles and MPI domain has been merged and parallelized. All considered, these improvements provide speedups of 3.1 for order 1 and 4.6 for order 3 interpolation shape factors on KNL configured in SNC4 quadrant flat mode. Performance is similar between a node of cori phase 1 and KNL at order 1 and better on KNL by a factor 1.6 at order 3 with the considered test case (homogeneous thermal plasma).

97 MATHEMATICS AND COMPUTING↗