Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “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.

At least 955 records · Page 53

UNITY: Unified Memory and File Space. In: ROSS '17 Proceedings of the 7th International Workshop on Runtime and Operating Systems for Supercomputers ROSS 2017, Article No. 6

This paper describes the vision for UNITY, a new high-performance computing focused data storage abstraction that places the entire memory hierarchy, including both traditionally separated memory-and file-based data storage, into one storage continuum. Through the use of a novel API and a set of services centered around a smart runtime system, UNITY is able to provide a number of valuable and interesting benefits. The unified storage space provides a scalable and resilient data environment that dynamically manages the mapping of data onto available resources based on multiple factors, including desired persistence and energy budget considerations. By eliminating the need for high-performance computing domain scientists to develop architecture-dependent optimizations for rapidly evolving data storage technologies, UNITY addresses both ease-of-use and performance.

Jones, Terry↗

BonVoision: Leveraging Spatial Data Smoothness For Recovery From Memory Soft Errors

The increasing soft error rates in memory systems raise an emerging concern for modern computing systems. As a result, detectable but uncorrectable errors (DUEs) become potentially more frequent and affect HPC applications. Today, upon encountering a DUE, HPC applications crash, incurring significant performance, storage, and energy overheads. In this paper, we propose a technique to continue application execution past a DUE through the repair of the corrupted memory data by leveraging spatial data smoothness. We present BonVoision, a run-time system that intercepts DUE events, analyzes the binary to identify data elements in the structural neighborhood of the event, and fixes the corrupted data elements by interpolating from the values in their neighborhood. Our evaluation demonstrates that BonVoision incurs negligible overhead and outperforms other recovery strategies by a factor of 2×, on average. We demonstrate that BonVoision also improves the efficiency of existing checkpointing/restart schemes by approximately increasing the optimal checkpoint interval by 23%.

Fang, Bo↗

CoREC: Scalable and Resilient In-memory Data Staging for In-situ Workflows

The dramatic increase in the scale of current and planned high-end HPC systems is leading new challenges, such as the growing costs of data movement and IO, and the reduced mean time between failures (MTBF) of system components. In-situ workflows, i.e., executing the entire application workflows on the HPC system, have emerged as an attractive approach to address data-related challenges by moving computations closer to the data, and staging-based frameworks have been effectively used to support in-situ workflows at scale. However, the resilience of these staging-based solutions has not been addressed, and they remain susceptible to expensive data failures. Furthermore, naive use of data resilience techniques such as n-way replication and erasure codes can impact latency and/or result in significant storage overheads. In this article, we present CoREC, a scalable and resilient in-memory data staging runtime for large-scale in-situ workflows. CoREC uses a novel hybrid approach that combines dynamic replication with erasure coding based on data access patterns. It also leverages multiple levels of replications and erasure coding to support diverse data resiliency requirements. Furthermore, the article presents optimizations for load balancing and conflict-avoiding encoding, and a low overhead, lazy data recovery scheme. We have implemented the CoREC runtime and have deployed with the DataSpaces staging service on leadership class computing machines and present an experimental evaluation in the article. Here, the experiments demonstrate that CoREC can tolerate in-memory data failures while maintaining low latency and sustaining high overall storage efficiency at large scales.

97 MATHEMATICS AND COMPUTING↗

Timely Reporting of Heavy Hitters Using External Memory

Given an input stream S of size N, a Φ-heavy hitter is an item that occurs at least ΦN times in S. The problem of finding heavy-hitters is extensively studied in the database literature. In this work, we study a real-time heavy-hitters variant in which an element must be reported shortly after we see its T = Φ N-th occurrence (and hence it becomes a heavy hitter). We call this the Timely Event Detection (TED) Problem. The TED problem models the needs of many real-world monitoring systems, which demand accurate (i.e., no false negatives) and timely reporting of all events from large, high-speed streams with a low reporting threshold (high sensitivity). Like the classic heavy-hitters problem, solving the TED problem without false-positives requires large space (Ω (N) words). Thus in-RAM heavy-hitters algorithms typically sacrifice accuracy (i.e., allow false positives), sensitivity, or timeliness (i.e., use multiple passes). We show how to adapt heavy-hitters algorithms to external memory to solve the TED problem on large high-speed streams while guaranteeing accuracy, sensitivity, and timeliness. Our data structures are limited only by I/O-bandwidth (not latency) and support a tunable tradeoff between reporting delay and I/O overhead. With a small bounded reporting delay, our algorithms incur only a logarithmic I/O overhead. We implement and validate our data structures empirically using the Firehose streaming benchmark. Multi-threaded versions of our structures can scale to process 11M observations per second before becoming CPU bound. In comparison, a naive adaptation of the standard heavy-hitters algorithm to external memory would be limited by the storage device’s random I/O throughput, i.e., ≈100K observations per second.

97 MATHEMATICS AND COMPUTING↗

Synchronization for CXL Based Memory

Compute Express Link (CXL) is an important emerging standard for disaggregated memory. While this standard provisions coherency across numerous hosts and devices, implementing hardware support for type three devices is challenging. In this work, we look at the overhead of software synchronization and using software-based coherency. Moreover, we discuss the limits of software-based coherency in fully expressing modern synchronization techniques for a CXL-based disaggregate memory system. We demonstrate our approach using a CXL hardware prototype and running a version of the famous Peterson Lock (enhanced to run with more than two threads). We analyze its performance and share how more advanced synchronization techniques might interact with software-based coherence CXL hardware and program execution models.

High Performance Computing (HPC)↗

Studying CPU and memory utilization of applications on Fujitsu A64FX and Nvidia Grace Superchip

ARM-based manycore CPU architectures are well-positioned to provide the rising memory throughput requirements of modern data intensive scientific applications in High Performance Computing (HPC). The Fujitsu A64FX CPU platform is based on the ARM v8.2A architecture, and is the processor of the flagship Japanese supercomputer - "Fugaku", which was previously ranked as the #1 supercomputer in the world according to the Top500 list. The Nvidia Grace superchip features 144 Neoverse V2 cores based on the ARMv9 architecture with 4x128b SVE2, providing exceptional computational power. The chip supports up to 480GB of memory, making it ideal for AI, machine learning, and scientific computing workloads. In this paper, we conduct a thorough performance exploration of a variety of parallel bandwidth-sensitive benchmarks and applications compiled with the native Fujitsu compiler on a Fugaku A64FX compute node and ARM (LLVM) Compiler on an NVIDIA Grace superchip compute node, engaging all the computational cores per cluster using OpenMP multithreading (assuming the cores can drive the available bandwidth). Our ultimate goals are to study the resource utilization of scientific applications and benchmarks on A64FX and Grace superchip, considering graph application scenarios ( GAP Benchmark suite) and eleven appli- cation proxies from the Rodinia heterogeneous benchmark suite (considering domains such as Data Mining, Bioinformatics, Fluid Dynamics, Pattern Recognition, etc.). Through exhaustive performance monitoring, we quantify the resource utilization of diverse OpenMP-based HPC applications on both the Fujitsu A64FX and the Nvidia Grace Superchip platforms.

benchmarking, Performance Analysis, High performan↗

Materials and Physics Challenges for Spin Transfer Torque Magnetic Random Access Memories

Magnetic random access memories utilizing the spin transfer torque effect for writing information are a strong contender for non-volatile memories scalable to the 20 nm node, and perhaps beyond. I will here examine how these devices behave as the device size is scaled down from 70 nm size to 20 nm. As device sizes go below ~50 nm, the size becomes comparable to intrinsic magnetic length scales and the device behavior does not simply scale with size. This has implications for the device design and puts additional constraints on the materials in the device.

Heinonen, O.↗

Data-Intensive Memory-Map simulator and runtime

DI-MMAP is a simulator for modeling the performance of next generation non-volatile random access memory technologies (NVRAM) and a high-perfromance memory-map runtime for the Linux operating system. It is implemented as a device driver for the Linux operating system. It will be used by algorithm designers to unserstand the impact of future NVRAM on their algorithms and will be used by application developers for high-performance access to NVRAM storage.

Essen, B.B.↗

Memory Analysis Tool

This tool traces all memory accesses to stack (static allocation) and heap (dynamic allocation) on a trace run in a particular hardware and then estimates execution time on given arbitrary hardware configurations for hardware design space exploration. It can also compute memory access statistics, such as reuse distance.

Sato, Kento↗

High-temperature superelasticity in NiTiPt and NiTiPd shape memory alloys

Alloying NiTi shape memory alloys with sufficient Pt and Pd at the expense of Ni has been shown to increase transformation temperatures well above those possible with the binary composition, enabling consideration of a new class of high-temperature mechanical sensors and actuators. To achieve shape memory behavior at temperatures greater than 200 °C, previous studies of NiTiPt and NiTiPd compositions have focused primarily on alloys with greater than 16 at.% Pt and 25 at.% Pd. While generally not the focus, none of these previous studies demonstrated superelastic behavior at high temperatures. In an effort to identify alloys that transform to austenite at approximately 200 °C and exhibit superelastic behavior with recoverable strains greater than 2%, a series of Ti-rich NiTiPt ternary alloys with 15.5–16.5 at.% Pt, and Ti-rich NiTiPd ternary alloys with 25–26 at.% Pd have been targeted. The polycrystalline alloys were processed via arc melting followed by homogenization heat treatment. Thermomechanical properties were measured using electro-discharge machined tension test coupons and a miniature load frame equipped with a clamshell furnace. Ultimately, superelasticity in uniaxial tension with clear plateau regions was demonstrated and characterized in these alloys at temperatures between 219 °C and 270 °C.

Massad, Jordan E.↗

Memory and CPU efficient coherent mode decomposition of partially coherent synchrotron radiation with subtraction of common quadratic phase terms

Application examples of a memory and CPU efficient coherent mode decomposition (CMD) method for wave-optics based simulation of the partially coherent undulator radiation propagation through a hard X-ray beamline in a 3rd generation synchrotron radiation source are presented. The high efficiency of the method is achieved thanks to the analytical treatment of the common quadratic phase terms that are developed in the phase of cross-spectral density (CSD) of partially coherent radiation at a distance from source. This treatment allows for a considerable, several orders of magnitude, reduction of the 4D CSD mesh density (and the memory occupied by the CSD) required for ensuring sufficient accuracies of wavefront propagation simulations with the modes produced by the CMD at a beamline entrance. This method, implemented in the “Synchrotron Radiation Workshop” open-source software, dramatically increases the feasibility of the CMD of 4D CSD for producing 2D coherent modes for a large variety of applications at storage rings and other types of radiation sources.

36 MATERIALS SCIENCE↗

A program for undergraduate research into the mechanisms of sensory coding and memory decay

This is the final technical report for this DOE project, entitltled "A program for undergraduate research into the mechanisms of sensory coding and memory decay". The report summarizes progress on the three research aims: 1) to identify phyisological and genetic correlates of long-term habituation, 2) to understand mechanisms of olfactory coding, and 3) to foster a world-class undergraduate neuroscience program. Progress on the first aim has enabled comparison of learning-regulated transcripts across closely related learning paradigms and species, and results suggest that only a small core of transcripts serve truly general roles in long-term memory. Progress on the second aim has enabled testing of several mutant phenotypes for olfactory behaviors, and results show that responses are not fully consistent with the combinitoral coding hypothesis. Finally, 14 undergraduate students participated in this research, the neuroscience program attracted extramural funding, and we completed a successful summer program to enhance transitions for community-college students into 4-year colleges to persue STEM fields.

59 BASIC BIOLOGICAL SCIENCES↗

What Scientific Applications can Benefit from Hardware Transactional Memory?

Achieving efficient and correct synchronization of multiple threads is a difficult and error-prone task at small scale and, as we march towards extreme scale computing, will be even more challenging when the resulting application is supposed to utilize millions of cores efficiently. Transactional Memory (TM) is a promising technique to ease the burden on the programmer, but only recently has become available on commercial hardware in the new Blue Gene/Q system and hence the real benefit for realistic applications has not been studied, yet. This paper presents the first performance results of TM embedded into OpenMP on a prototype system of BG/Q and characterizes code properties that will likely lead to benefits when augmented with TM primitives. We first, study the influence of thread count, environment variables and memory layout on TM performance and identify code properties that will yield performance gains with TM. Second, we evaluate the combination of OpenMP with multiple synchronization primitives on top of MPI to determine suitable task to thread ratios per node. Finally, we condense our findings into a set of best practices. These are applied to a Monte Carlo Benchmark and a Smoothed Particle Hydrodynamics method. In both cases an optimized TM version, executed with 64 threads on one node, outperforms a simple TM implementation. MCB with optimized TM yields a speedup of 27.45 over baseline.

97 MATHEMATICS AND COMPUTING↗

Parallel Breadth-First Search on Distributed Memory Systems

Data-intensive, graph-based computations are pervasive in several scientific applications, and are known to be quite challenging to implement on distributed memory systems. In this work, we explore the design space of parallel algorithms for Breadth-First Search (BFS), a key subroutine in several graph algorithms. We present two highly-tuned parallel approaches for BFS on large parallel systems: a level-synchronous strategy that relies on a simple vertex-based partitioning of the graph, and a two-dimensional sparse matrix-partitioning-based approach that mitigates parallel communication overhead. For both approaches, we also present hybrid versions with intra-node multithreading. Our novel hybrid two-dimensional algorithm reduces communication times by up to a factor of 3.5, relative to a common vertex based approach. Our experimental study identifies execution regimes in which these approaches will be competitive, and we demonstrate extremely high performance on leading distributed-memory parallel systems. For instance, for a 40,000-core parallel execution on Hopper, an AMD Magny-Cours based system, we achieve a BFS performance rate of 17.8 billion edge visits per second on an undirected graph of 4.3 billion vertices and 68.7 billion edges with skewed degree distribution.

97 MATHEMATICS AND COMPUTING↗

Ultra Low Density Shape Memory Polymer Foams With Tunable Physicochemical Properties for Treatment of intracranial Aneurysms

Shape memory polymers (SMPs) are a rapidly emerging class of smart materials that can be stored in a deformed temporary shape, and can actively return to their original shape upon application of an external stimulus such as heat, pH or light. This behavior is particularly advantageous for minimally invasive biomedical applications comprising embolic/regenerative scaffolds, as it enables a transcatheter delivery of the device to the target site. The focus of this work was to exploit this shape memory behavior of polyurethanes, and develop an efficient embolic SMP foam device for the treatment of intracranial aneurysms.In summary, this work reports a novel family of ultra low density polymer foams which can be delivered via a minimally invasive surgery to the aneurysm site, actuated in a controlled manner to efficiently embolize the aneurysm while promoting physiological fluid/blood flow through the reticulated/open porous structure, and eventually biodegrade leading to complete healing of the vasculature.

36 MATERIALS SCIENCE↗

Improving Physics Based Electron Neutrino Appearance Identication with a Long Short-Term Memory Network

The NOvA experiment is a long baseline neutrino oscillation experiment with the objective to measure the oscillation probability of muon type neutrinos ($\nu_{\mu}$) into electron type neutrinos ($\nu_e$). NOvA measures the interactions of neutrinos from the NuMI beam in two functionally identical liquid scintillator detectors. The far detector detects the appearance of electron neutrinos, leading to measurement of the oscillation parameters under study. Using an off -axis beam with an 810 km baseline length, NOvA is sensitive to measuring the neutrino mass hierarchy, the CP violating parameter, and the octant of the mixing angle, $\theta_{23}$. The data presented in this thesis has been collected from October 2013 until May 2018. The first NOvA $\nu_e$ charged current identifier utilized an artificial neural network with the physical features of the highest energy reconstructed shower as inputs. The $\nu_ e$ charged current identifier in this thesis utilizes a Long Short-Term Memory network with the physical features of every reconstructed shower in a particular interaction. In addition to the Long Short-Term Memory network, there are two Boosted Decision Trees to assist in event level selection. In the analysis of the data, 54 $\nu_e$ candidate events were detected with an expected background of 15 events. The results of this analysis prefer the normal mass hierarchy with maximal mixing and $\frac{\delta_{CP}}{\pi}=1.92_{-1.19}^{0.08}$. Results from this analysis differ from the published NOvA analysis, due to the differences of electron identification techniques.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Sparse Data Acquisition on Emerging Memory Architectures

Emerging memory devices, such as resistive crossbars, have the capacity to store large amounts of data in a single array. Acquiring the data stored in large-capacity crossbars in a sequential fashion can become a bottleneck. We present practical methods, based on sparse sampling, to quickly acquire sparse data stored on emerging memory devices that support the basic summation kernel, reducing the acquisition time from linear to sub-linear. Experimental results show that at least an order of magnitude improvement in acquisition time can be achieved when the data is sparse.

97 MATHEMATICS AND COMPUTING↗

Multi-Level Memory Algorithmics for Large, Sparse Problems

In this report, we abstract eleven papers published during the project and describe preliminary unpublished results that warrant follow-up work. The topic is multi-level memory algorithmics, or how to effectively use multiple layers of main memory. Modern compute nodes all have this feature in some form.

97 MATHEMATICS AND COMPUTING↗