Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Memory benchmark”

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 235 records · Page 13

Feedback-directed page placement for ccNUMA via hardware-generated memory traces

Non-uniform memory architectures with cache coherence (ccNUMA) are becoming increasingly common, not just for large-scale high performance platforms but also in the context of multi-core architectures. Under ccNUMA, data placement may influence overall application performance significantly as references resolved locally to a processor/core impose lower latencies than remote ones. This work develops a novel hardware-assisted page placement paradigm based on automated tracing of the memory references made by application threads. Two placement schemes, modeling both singlelevel and multi-level latencies, allocate pages near processors that most frequently access that memory page. These schemes leverage performance monitoring capabilities of contemporary microprocessors to efficiently extract an approximate trace of memory accesses. This information is used to decide page affinity, i.e., the node to which the page is bound. The method operates entirely in user space, is widely automated, and handles not only static but also dynamic memory allocation. Experiments show that this method, although based on lossy tracing, can efficiently and effectively improve page placement, leading to an average wall-clock execution time saving of over 20% for the tested benchmarks on the SGI Altix with a 2x remote access penalty and 12% on AMD Opterons with a 1.3–2.0x access penalty. This is accompanied by a one-time tracing overhead of 2.7% over the overall original program wallclock time.

Computer Science↗

Forecasting of in situ electron energy loss spectroscopy

Abstract Forecasting models are a central part of many control systems, where high-consequence decisions must be made on long latency control variables. These models are particularly relevant for emerging artificial intelligence (AI)-guided instrumentation, in which prescriptive knowledge is needed to guide autonomous decision-making. Here we describe the implementation of a long short-term memory model (LSTM) for forecasting in situ electron energy loss spectroscopy (EELS) data, one of the richest analytical probes of materials and chemical systems. We describe key considerations for data collection, preprocessing, training, validation, and benchmarking, showing how this approach can yield powerful predictive insight into order-disorder phase transitions. Finally, we comment on how such a model may integrate with emerging AI-guided instrumentation for powerful high-speed experimentation.

36 MATERIALS SCIENCE↗

GPU-Centric Communication on NVIDIA GPU Clusters with InfiniBand: A Case Study with OpenSHMEM

GPUs have become an essential component for building compute clusters with high compute density and high performance per watt. As such clusters scale to have 1000s of GPUs, efficiently moving data between the GPUs becomes imperative to get maximum performance. NVSHMEM is an implementation of the OpenSHMEM standard for NVIDIA GPU clusters which allows communication to be issued from inside GPU kernels. In earlier work, we have shown how NVSHMEM can be used to achieve better application performance on GPUs connected through PCIe or NVLink. As part of this effort, we implement IB verbs for Mellanox InfiniBand adapters in CUDA. We evaluate different design alternatives, taking into consideration the relaxed memory model, automatic memory access coalescing and thread hierarchy on the GPU. We also consider correctness issues that arise in these designs. We take advantage of these designs transparently or through API extensions in NVSHMEM. With micro-benchmarks, we show that a Nvidia Pascal P100 GPU is able saturate the network bandwidth using only one or two of its 56 available streaming multiprocessors (SM). On a single GPU using a single IB EDR adapter, we achieve a throughput of around 90 million messages per second. In addition, we implement a 2dstencil application kernel using NVSHMEM and compare its performance with a CUDA-aware MPI-based implementation that uses GPUDirect RDMA. Speedups in the range of 23% to 42% are seen for input sizes large enough to fill the occupancy of Nvidia Pascal P100 GPUs on 2 to 4 nodes indicating that there are gains to be had by eliminating the CPU from the communication path when all computation runs on the GPU.

Potluri, Sreeram↗

Scaling the memory wall using mixed-precision - HPG-MxP on an exascale-class machine

Mixed-precision algorithms have been proposed as a way for scientific computing to benefit from some of the gains seen for AI on recent high performance computing (HPC) platforms. A few applications dominated by dense matrix operations have seen substantial speedups by utilizing low precision formats such as FP16. However, a majority of scientific simulation applications are memory bandwidth limited. Beyond preliminary studies, the practical gain from using mixed-precision algorithms on a given high-performance computing (HPC) system is largely unclear. The High Performance GMRES Mixed Precision (HPG-MxP) benchmark has been proposed to measure the useful performance of a HPC system on sparse matrix-based mixed-precision applications. In this work, we present an implementation of the HPG-MxP benchmark for an exascale system and describe our algorithm enhancements. We show for the first time a speedup of 1.6x using a combination of double- and single-precision keeping the same residual level on modern GPU-based supercomputers.

Kashi, Aditya [ORNL] (ORCID:0000000325893792)↗

Monitoring of Liquid Metal Reactor Heater Zones with Recurrent Neural Network Learning of Temperature Time Series

Advanced high-temperature fluid reactors (ARs), such as sodium fast reactors (SFRs) and molten salt cooled reactors (MSCRs) utilize high-temperature fluids at ambient pressure. To melt the fluid during reactor startup and prevent fluid freezing during cooldown, the thermal–hydraulic systems of such ARs include heater zones consisting of specific heaters with controllers, temperature sensors, and thermal insulation. The failure of heater zones due to insulation material degradation or improper installation, resulting in parasitic heat losses, can lead to fluid freezing. The detection of faults using a heat-transfer model is difficult because of a lack of knowledge of the experimental details. Data-driven machine learning of heater zone temperature time series offers a viable alternative. In this study, we benchmarked the performance of recurrent neural networks (RNNs) in an analysis of heat-up transient temperature time series of heater zones installed on a liquid sodium vessel. The RNN models include long short-term memory (LSTM) and gated recurrent unit (GRU) networks, as well as their bi-directional variants, BiLSTM and BiGRU. Anomalous temperature points were designated using a percentile-based threshold applied to residual fluctuations in the detrended temperature time series. Additionally, the impact of the exponentially weighted moving average (EWMA) method on detection accuracy was examined. The RNN models’ performance was assessed using precision, recall, and F 1 score metrics. Results demonstrated that RNN models effectively detect anomalies in temperature time series with the best models for each heater zone achieving F 1 scores of over 93%. To explain the variations in RNN model performance across different heater zones, we used Kullback–Leibler (KL) divergence to quantify the relative entropy between training and testing data, and the Detrended Fluctuation Analysis (DFA) to assess long-range temporal correlations. For datasets with strong long-range correlations and minimal relative entropy between training and testing data, GRU is the best-performing model. When the data exhibits weaker long-term correlations and a significant relative entropy between training and testing distributions, BiGRU shows the best performance. For the data sets with intermediate values of both KL divergence and DFA, the best performance is obtained with LSTM and BiLSTM, respectively.

gated recurrent unit↗

Detection of Isotopes in Urban Source Search Low-Count Gamma Spectra Using Hopfield Neural Networks

Source search campaigns involve measurements of background gamma-ray spectra with a mobile detector-spectrometer traveling along arbitrarily chosen trajectories over a wide screening area. Radiation counts are typically measured with a tellurium-doped sodium iodide [NaI(Tl)] scintillator detector-spectrometer in short acquisition intervals, usually 1 s. The objective is to detect orphan isotopes with half-lives shorter than those of the isotopes in the natural background. In principle, radioisotopes can be identified by their unique gamma emission spectrum. However, detecting orphan isotopes in search data is challenging because low counts measured in short acquisition intervals result in incomplete spectral lines. In this study, we investigate the performance of a Hopfield neural network (HNN) that implements an auto-associative memory for the detection of isotopes of interest in an urban search campaign. The HNN is trained on one example of gamma spectra with well-resolved spectral lines of each isotope of interest. During testing, the auto-associative memory implementation of the HNN processes low-count gamma spectra with partially complete isotopic lines by matching incoming measurements to the closest one of its memory-stored patterns. The testing database consisted of almost 10 000 1-s gamma spectra, including measurements of orphan isotopes 137 Cs, 241 Am, and 131 I, obtained during two urban search surveys with a NaI(Tl) detector. The performance of the HNN detection algorithm was evaluated using precision, recall, and F1 scores, and benchmarked with a multiple linear regression (MLR) identification algorithm. In conclusion, the test results demonstrate that HNN outperforms MLR in the detection of all the isotopes of interest.

Auto associative memory↗

Low-Cost Parallel Algorithms for 2:1 Octree Balance

The logical structure of a forest of octrees can be used to create scalable algorithms for parallel adaptive mesh refinement (AMR), which has recently been demonstrated for several petascale applications. Among various frequently used octree-based mesh operations, including refinement, coarsening, partitioning, and enumerating nodes, ensuring a 2:1 size balance between neighboring elements has historically been the most expensive in terms of CPU time and communication volume. The 2:1 balance operation is thus a primary target to optimize. One important component of a parallel balance algorithm is the ability to determine whether any two given octants have a consistent distance/size relation. Based on new logical concepts we propose fast algorithms for making this decision for all types of 2:1 balance conditions in 2D and 3D. Since we are able to achieve this without constructing any parent nodes in the tree that would otherwise need to be sorted and communicated, we can significantly reduce the required memory and communication volume. In addition, we propose a lightweight collective algorithm for reversing the asymmetric communication pattern induced by non-local octant interactions. We have implemented our improvements as part of the opensource “p4est” software. Benchmarking this code with both synthetic and simulation-driven adapted meshes we are able to demonstrate much reduced runtime and excellent weak and strong scalability. On our largest benchmark problem with 5.13 × 10 11 octants the new 2:1 balance algorithm executes in less than 8 seconds on 112,128 CPU cores of the Jaguar Cray XT5 supercomputer.

Computer Science↗

PERKS: a Locality-Optimized Execution Model for Iterative Memory-bound GPU Applications

Iterative memory-bound solvers commonly occur in HPC codes. Typical GPU implementations have a loop on the host side that invokes the GPU kernel as much as time/algorithm steps there are. The termination of each kernel implicitly acts the barrier required after advancing the solution every time step. We propose an execution model for running memory-bound iterative GPU kernels: PERsistent KernelS (PERKS). In this model, the time loop is moved inside persistent kernel, and device-wide barriers are used for synchronization. We then reduce the traffic to device memory by caching subset of the output in each time step in the unused registers and shared memory. PERKS can be generalized to any iterative solver: they largely independent of the solver's implementation. We explain the design principle of PERKS and demonstrate effectiveness of PERKS for a wide range of iterative 2D/3D stencil benchmarks (geomean speedup of 2.12x for 2D stencils and 1.24x for 3D stencils over state-of-art libraries), and a Krylov subspace conjugate gradient solver (geomean speedup of 4.86x in smaller SpMV datasets from SuiteSparse and 1.43x in larger SpMV datasets over a state-of-art library). All PERKS-based implementations available at: https://github.com/neozhang307/PERKS.

Zhang, Lingqi↗

OpenSHMEM Implementation of IOR Benchmark

The benchmarking effort within the Computational Research & Development Programs at the Oak Ridge National Laboratory (ORNL) seeks to design and enable High Performance Computing (HPC) benchmarks and test suites. The work described in this report is a part of the effort focusing on the comparison and analysis of OpenSHMEM implementations using the Interleave Or Random (IOR) software for benchmarking parallel file system using POSIX, MPIIO, or HDF5 interfaces. We describe the effort to emulate the MPIIO parallel collective capabilities in the IOR benchmark using OpenSHMEM communication. One development effort was in emulating the MPI derived datatype used in the read/write operations and in setting the file view. Another effort was in implementing an internal cache in OpenSHMEM distributed shared memory to facilitate global collective I/O operations. Experiments comparing collective I/O in MPIIO implementations with the OpenSHMEM implementations were performed on the SGI Turing Cluster and the Cray XK7 Titan supercomputer at the Oak Ridge Leadership Computing Facility (OLCF). The preliminary results suggest that on the Cray XK7 Titan, the MPIIO implementations obtained higher write performance and the OpenSHMEM version obtained slightly higher read performance. On the SGI Turing Cluster, the MPIIO implementations obtained slightly higher performance over the OpenSHMEM implementations on large files.

97 MATHEMATICS AND COMPUTING↗

Study of Whisper Statistics for Criticality Safety Code Validation

Whisper is computational software designed to assist with validation studies with the Monte Carlo particle transport package MCNP. Historical approaches rely on expert based judgment to select benchmarks while Whisper uses sensitivity/uncertainty (S/U) methods to select relevant benchmarks to a particular application model. Whisper will also derive a calculational margin and margin of subcriticality (MOS) that may be used to generate an upper subcritical limit (USL) for a particular application model. Whisper finds benchmark experiments whose isotope sensitivity profiles align best with the MCNP model in question, and assign a correlation coefficient ('ck' in Whisper parlance) to each selected benchmark as a measure of how closely those profiles agree. When ck values are high, there is less uncertainty between model and benchmarks and a higher USL is possible, relative to cases when ck values are low. The default Whisper settings recommends a 100,000 particles per cycle run for 500 active cycles, resulting in 50 M active neutron histories per application run. This is computationally intensive and may not be necessary to produce accurate results. A Whisper statistical study was done varying the neutron histories per cycle from 10,000 to the default 100,000. A lower number of neutron histories will allow Whisper to be used on low memory systems such as laptops. The intent of this paper is to present the changes in the Whisper results as the number of neutrons histories run is varied.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

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.

97 MATHEMATICS AND COMPUTING↗

A Programming Model for Massive Data Parallelism with Data Dependencies

Accelerating processors can often be more cost and energy effective for a wide range of data-parallel computing problems than general-purpose processors. For graphics processor units (GPUs), this is particularly the case when program development is aided by environments such as NVIDIA s Compute Unified Device Architecture (CUDA), which dramatically reduces the gap between domain-specific architectures and general purpose programming. Nonetheless, general-purpose GPU (GPGPU) programming remains subject to several restrictions. Most significantly, the separation of host (CPU) and accelerator (GPU) address spaces requires explicit management of GPU memory resources, especially for massive data parallelism that well exceeds the memory capacity of GPUs. One solution to this problem is to transfer data between the GPU and host memories frequently. In this work, we investigate another approach. We run massively data-parallel applications on GPU clusters. We further propose a programming model for massive data parallelism with data dependencies for this scenario. Experience from micro benchmarks and real-world applications shows that our model provides not only ease of programming but also significant performance gains.

43 PARTICLE ACCELERATORS↗

SharP Unified Memory Allocator: An Intent-Based Memory Allocator for Extreme-Scale Systems

The pre-exascale systems will soon be deployed with a deep, complex memory hierarchy composed of many heterogeneous memories. This presents multiple challenges for users including: how to allocate data objects with locality between memories and devices for the various memories in these systems, which includes DRAM, High-bandwidth Memory (HBM), and non-volatile random access memory (NVRAM), and how to perform these allocations while providing portability for their application. Currently, the user can make use of multiple, disjoint libraries to allocate data objects on these memories. However, it is difficult to obtain locality between memories and devices when using libraries that are unaware of each other. This paper presents the Unified Memory Allocator (UMA) of the SHARed data-structure centric Programming abstraction (SharP) library, which provides a unified interface for memory allocations across DRAM, HBM, and NVRAM and is extensible to support future memory types. In addition, the SharP UMA allows for portability between systems by supporting both explicit and implicit, intent-based memory allocations. To demonstrate the ease of use of the SharP UMA, we have extended both Open MPIand OpenSHMEM-Xto support SharP. We validate this work by evaluating the performance implications and intent-based approach with synthetic benchmarks as well as adaptations of the Graph500 benchmark.

Aderholdt, William↗

Maximal clique enumeration with data-parallel primitives

The enumeration of all maximal cliques in an undirected graph is a fundamental problem arising in several research areas. We consider maximal clique enumeration on shared-memory, multi-core architectures and introduce an approach consisting entirely of data-parallel operations, in an effort to achieve efficient and portable performance across different architectures. We study the performance of the algorithm via experiments varying over benchmark graphs and architectures. Overall, we observe that our algorithm achieves up to a 33-time speedup and 9-time speedup over state-of-the-art distributed and serial algorithms, respectively, for graphs with higher ratios of maximal cliques to total cliques. Further, we attain additional speedups on a GPU architecture, demonstrating the portable performance of our data-parallel design.

97 MATHEMATICS AND COMPUTING↗

Enabling HPC Scientific Workflows for Serverless

The convergence of edge computing, big data analytics, and AI with traditional scientific calculations is increasingly being adopted in HPC workflows. Workflow management systems are crucial for managing and orchestrating these complex computational tasks. However, it is difficult to identify patterns within the growing population of HPC workflows. Serverless has emerged as a novel computing paradigm, offering dynamic resource allocation, quick response time, fine-grained resource management and auto-scaling. In this paper, we propose a framework to enable HPC scientific workflows on serverless. Our approach integrates a widely used traditional HPC workflow generator with an HPC serverless workflow management system to create benchmark suites of scientific workflows with diverse characteristics. These workflows can be executed on different serverless platforms. We comprehensively compare executing workflows on traditional local containers and serverless computing platforms. Our results show that serverless can reduce CPU and memory usage respectively by 78.11% and 73.92% without compromising performance.

Andrei da silva, Anderson↗

Quicksilver

Quicksilver is proxy application for the Monte Carlo Transport code, Mercury. The purpose of the Quicksilver mini-app is to create a freely distributable code base that represents the key elements of the Mercury workload by solving a simplified time-dependent neutron transport problem. As an open source code, Quicksilver will be hosted on the CZ git/stash repo alongside other mini-apps in the ACS so-design suite. It will be used by vendors and other partners for co-design activities such as identifying root causes of Mercury performance bottlenecks on future architectures (i.e., memory bandwidth, memory latency, flop intensity) and developing Quicksilver will be used within LLNL as a nimble platform to evaluate programming models, algorithms, etc. that might be useful in Mercury. Quicksilver may also become a replacement for MCB in future procurement benchmarks, as well as a way to engage vendors in the Trinity and Sierra Centers of Excellence as well as other co-design efforts.

Richards, David↗

Implementation of on-the-fly doppler broadening in MCNP

A new method to obtain Doppler broadened cross sections has been implemented into MCNP, removing the need to generate cross sections for isotopes at problem temperatures. When a neutron of energy E enters a material region that is at some temperature T, the cross sections for that material at temperature T are immediately obtained 'on-the-fly' (OTF) by interpolation using a high order functional expansion for the temperature dependence of the Doppler-broadened cross section for that isotope at the neutron energy E. The OTF cross sections agree with the NJOY-based cross sections for all neutron energies and all temperatures in the range specified by the user, e.g., 250 K - 3200 K. The OTF methodology has been successfully implemented into the MCNP Monte Carlo code and has been tested on several test problems by comparing MCNP with conventional ACE cross sections versus MCNP with OTF cross sections. The test problems include the Doppler defect reactivity benchmark suite and two full-core VHTR configurations, including one with multiphysics coupling using RELAP5-3D/ATHENA for the thermal-hydraulic analysis. The comparison has been excellent, verifying that the OTF libraries can be used in place of the conventional ACE libraries generated at problem temperatures. In addition, it has been found that the OTF methodology greatly reduces the complexity of the input for MCNP, resulting in an order of magnitude decrease in the number of input lines for full-core configurations. Finally, for full-core problems with multiphysics feedback, the memory required to store the cross section data is considerably reduced with OTF cross sections and the additional computational effort with OTF is modest, on the order of 10-15%. (authors)

97 MATHEMATICS AND COMPUTING↗

A performance study of the time-varying cache behavior: a study on APEX, Mantevo, NAS, and PARSEC

Cache has long been used to minimize the latency of main memory accesses by storing frequently used data near the processor. Processor performance depends on the underlying cache performance. Therefore, significant research has been done to identify the most crucial metrics of cache performance. Although the majority of research focuses on measuring cache hit rates and data movement as the primary cache performance metrics, cache utilization is significantly important. We investigate the application’s locality using cache utilization metrics. In addition, we present cache utilization and traditional cache performance metrics as the program progresses providing detailed insights into the dynamic application behavior on parallel applications from four benchmark suites running on multiple cores. We explore cache utilization for APEX, Mantevo, NAS, and PARSEC, mostly scientific benchmark suites. Our results indicate that 40% of the data bytes in a cache line are accessed at least once before line eviction. Also, on average a byte is accessed two times before the cache line is evicted for these applications. Moreover, we present runtime cache utilization, as well as, conventional performance metrics that illustrate a holistic understanding of cache behavior. To facilitate this research, we build a memory simulator incorporated into the Structural Simulation Toolkit (Rodrigues et al. in SIGMETRICS Perform Eval Rev 38(4):37–42, 2011). Finally, our results suggest that variable cache line size can result in better performance and can also conserve power.

97 MATHEMATICS AND COMPUTING↗