Search NASA⌕ Search

SEARCH · Search NASA

Results for “hardware testing”

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 1,675 records · Page 93

Success in commissioning a key U.S. radioactive waste burial site through the technical support of Dufrane Nuclear Shielding Incorporated (DNSI) - 15476

With the -phase out- of the South Carolina based Barnwell Radioactive Waste Burial Site, and the stockpile and storage of radioactive waste and irradiated components throughout the U.S. Nuclear Industry for decades, the opening of the Texas Low-Level Radioactive Waste Disposal Compact Facility by Waste Control Specialists (WCS), is a game changer in solving this industry-wide issue. For the first time in decades, generators now have the opportunity to send their waste to a site specifically designed to permanently sequester the waste in a licensed facility that will protect public health and the environment, better than any built to date. As history has shown through the many decades of the U.S. Commercial and Federal Nuclear Industry, site operators and management have relied heavily on contractors and vendors with exceptional expertise and decades of experience to achieve success and meet their goals. The commissioning of WCS was no exception as DNSI was contracted to design, develop, build and deploy a major component required for the opening and successful operation of the WCS site. This component, being the Modular Concrete Canister (MCC), is the -work horse- of the site. With its enormous size, weight and strength, it provides many functions that meet and comply with the site performance assessment (PA) under 10 CFR Part 61, the site WAC, and Texas Low-Level Radioactive Waste Disposal Authority. Offsite contracting the production and delivery of this critical component was a major step toward success in meeting schedule and cost for the much needed opening of the Texas Compact and Federal site. In selecting a shielding vendor, WCS partnered with a U.S. Nuclear Service company with over 30 years of experience which spans the entire U.S. Commercial, Federal, and Nuclear Industry. The task given DNSI was not a simple one, it involved every aspect from start to finish of the design, production and delivery of rectangular, cylindrical and -irradiated hardware- (IH) MCCs, that could accept waste up to 100 Sv /Hr contact dose rates. The enormity of this project was quite clear and required extensive quality control, construction of batch plants, purchase of huge quantities of raw materials, setting up a production site in a neighboring city, relocation of personnel from company headquarters, and safely transporting 8,000 Tons of MCC components 75 miles one-way on West Texas highways daily. Most important of all, this was accomplished safely with no incidents, no component failures or rejects, and on schedule. The start of receipt, of -in-compact waste- and -out-of-compact waste- was conducted virtually in a 3 month period time span. To help guarantee success and quality of product, all forms had surfaces machined, heavy duty elastomer coatings were applied to concrete surfaces, only galvanized anchors were used, and the batch plant was run to strict ASME standards thus testing every 50 cubic yards of concrete poured. With significant teamwork and success from Vendors such as DNSI, WCS and the Texas Compact has evolved as one of the leading radioactive waste sites in the U.S. and the world, providing safe, secure and compliant permanent waste disposal services by using technical expertise and innovation of the highest standards. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Machine learning for continuous quantum error correction on superconducting qubits

Abstract Continuous quantum error correction has been found to have certain advantages over discrete quantum error correction, such as a reduction in hardware resources and the elimination of error mechanisms introduced by having entangling gates and ancilla qubits. We propose a machine learning algorithm for continuous quantum error correction that is based on the use of a recurrent neural network to identify bit-flip errors from continuous noisy syndrome measurements. The algorithm is designed to operate on measurement signals deviating from the ideal behavior in which the mean value corresponds to a code syndrome value and the measurement has white noise. We analyze continuous measurements taken from a superconducting architecture using three transmon qubits to identify three significant practical examples of non-ideal behavior, namely auto-correlation at temporal short lags, transient syndrome dynamics after each bit-flip, and drift in the steady-state syndrome values over the course of many experiments. Based on these real-world imperfections, we generate synthetic measurement signals from which to train the recurrent neural network, and then test its proficiency when implementing active error correction, comparing this with a traditional double threshold scheme and a discrete Bayesian classifier. The results show that our machine learning protocol is able to outperform the double threshold protocol across all tests, achieving a final state fidelity comparable to the discrete Bayesian classifier.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Classical combinatorial optimization scaling for random Ising models on 2D heavy-hex graphs

Motivated by near term quantum computing hardware limitations, combinatorial optimization problems that can be addressed by current quantum algorithms and noisy hardware with little or no overhead are used to probe capabilities of quantum algorithms such as the quantum approximate optimization algorithm. In this study, a specific class of near term quantum computing hardware defined combinatorial optimization problems, Ising models on heavy-hex graphs both with and without geometrically local cubic terms, are examined for their classical computational hardness via empirical computation time scaling quantification. Specifically the time-to-solution (TTS) metric using the classical heuristic simulated annealing is measured for finding optimal variable assignments (ground states), as well as the time required for the optimization software Gurobi to find an optimal variable assignment. Because of the sparsity of these Ising models, the classical algorithms are able to find optimal solutions efficiently even for large instances (i.e. 100 000 spin variables). The Ising models both with and without geometrically local cubic terms exhibit average-case linear-time or weakly quadratic scaling when solved exactly using Gurobi, and the Ising models with no cubic terms show evidence of exponential-time TTS scaling when sampled using simulated annealing. These findings point to the necessity of developing and testing more complex, namely more densely connected, optimization problems in order for quantum computing to ever have a practical advantage over classical computing. Our results are another illustration that different classical algorithms can indeed have exponentially different running times, thus making the identification of the best practical classical technique important in any quantum computing vs. classical computing comparison.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

FY2012 Report on Fast Reactor Toolset Work (PROTEUS-Fast)

As part of the NEAMS activity in DOE, a fast reactor simulation program was launched in April 2007, termed SHARP. The initial goal of SHARP was to develop a suite of modern simulation tools specifically for the analysis and design of sodium cooled fast reactors, and this goal has since been extended to consider all reactor types of interest. As part of that work, a high‐fidelity deterministic neutron transport code named UNIC was started, providing a common framework for fast reactor neutronics tools. Focusing on UNIC, spatial de‐homogenization increased the typical neutronics problem with a few hundred million degrees of freedom (DOF) into ones with hundreds of trillions of DOF. Such problems are beyond the capabilities of existing supercomputers. With the advent of future computer hardware, these high‐fidelity problems can be solved but only after a multiyear development program to adapt codes such as UNIC to that technology. UNIC can presently be used on smaller problems (single assembly, multiple assemblies) with today’s hardware, but it is also necessary to provide tools to solve the larger problems of interest, albeit without the aforementioned level of spatial de‐homogenization. To address the desire to demonstrate the coupling capability of NEAMS, part of the FY12 funding was focused on preparing the legacy codes, termed PROTEUS‐Fast, which are fully capable of solving full‐scale fast reactor problems. A bulk of the work this year was focused on preparing and distributing the PROTEUS‐Fast codes for use by all NEAMS users. The demonstration of a full coupling capability will require more follow on work, but significant progress was made this year preparing the interface coupling routines and preparing PROTEUS‐Fast for use. The primary targets of interest in PROTEUS‐Fast are DIF3D, DIF3D‐K (kinetics), and VARI3D (reactivity coefficients), with REBUS (fuel cycle) being an auxiliary component updated for convenience. All codes in PROTEUS‐Fast were verified and added to the BuildBot regression test. Unfortunately, all of the older codes suffered significant problems from the deprecated coding practices used within them and resulting conflicts with modern compilers. After some work, all of these issues were resolved and we feel confident in releasing them for use by others NEAMS members. With regard to specific upgrades, DIF3D‐VARIANT was upgraded with a few months of effort to allow the large scale problems of interest to be executed. After significant study, we choose not to modify VARI3D to include a transport option, but to develop a new code, PERSENT (partially funded by an external project) which uses the DIF3D‐VARIANT transport solution option. The outstanding features missing from MC 2 ‐3 were added, and MC 2 ‐3 has been released to RSICC. All of the remaining PROTEUS‐Fast codes are either finished with the copyright and licensing procedure or are actively going through it. While that process will delay the targeted August 2012 release to RSICC for a few more months, the codes are ready for immediate use and can be connected to the framework either by us or by others as the interface readers and documentation will be released with the PROTEUS‐Fast codes.

Wolters, E. R.↗

Extending Hypersonic Diagnostics to the Third Dimension

The design, construction, and initial testing of a high-magnification, long working-distance plenoptic camera is reported. A plenoptic camera uses a microlens array to enable resolution of the spatial and angular information of the incoming light field. With this, instantaneous images can be numerically refocused and perspective shifted in post-processing to enable instantaneous three-dimensional (3D) resolution of a scene. Prior to this work, most applications of plenoptic imaging were limited to relatively low magnifications (1× or less) or small working distances. Here, a unique system is developed with enables 5× magnification at a working distance of over a quarter meter. Experimental results demonstrate ~25 m spatial resolution with 3D imaging capabilities. This technology is demonstrated on two practical applications. First, burning aluminum particles on the order of 100 m in diameter are imaged near the reacting surface of a combusting solid rocket propellant. The long working distance is particularly advantageous for protection of the experimental hardware in this extremely hazardous environment. Next, background oriented schlieren is used to resolve the 3D structure of an underexpanded free jet. This demonstrates the ability to resolve index-of-refraction gradients at the working distances and spatial scales necessary to meet our ultimate goal of resolving 3D turbulent transition in the boundary layer of Sandia’s Hypersonic Wind Tunnel (HWT).

47 OTHER INSTRUMENTATION↗

TRIM: AI Guided Random Number Generation for Resource-Constrained IoT Systems

Random numbers often serve as the backbone for many security solutions in diverse domains such as cryptography, side channel leakage prevention, and moving target defense. However, generating true random numbers requires a physical source of entropy (e.g. hardware, quantum, environmental phenomenon) making it difficult to realize at a large scale and at a low cost. On the flip side, pseudorandom number generators (easy to implement) following a specific distribution (e.g. Gaussian) can be easily compromised given a sufficient amount of traces. In this work, we have developed a machine learning-guided generative approach that can be used to create portable, resource-efficient, and cost-effective random number generators with high throughput and true randomness characteristics. We implement the proposed approach as a highly parameterized framework and perform extensive evaluation for different settings. The framework was able to learn from true random sources such as irrational numbers and environmental audio noise and imitate those sources towards generating new good quality random numbers on demand. We have generated more than 1 billion bits and observed robust performance in terms of true randomness metrics obtained from NIST SP 800-22 and FIPS 140-1 randomness test suites achieving a throughput of up to 142.85 Mbps. Compared to the state-of-the-art (SOTA) technique, the iso-cost setup of our framework can achieve more than 500 Mbps in a distributed setting. We have evaluated the efficacy of running the true randomness imitation AI models on target edge devices such as Raspberry Pi 4 (Model B), Nvidia Jetson Nano, Nvidia Jetson Orin Nano and Nvidia Jetson Xavier. We have also looked at the security of the TRIM framework itself against different adversarial threat models.

Cybersecurity↗

Classically verifiable quantum advantage from a computational Bell test

Existing experimental demonstrations of quantum computational advantage have had the limitation that verifying the correctness of the quantum device requires exponentially costly classical computations. Here we propose and analyse an interactive protocol for demonstrating quantum computational advantage, which is efficiently classically verifiable. Our protocol relies on a class of cryptographic tools called trapdoor claw-free functions. Although this type of function has been applied to quantum advantage protocols before, our protocol employs a surprising connection to Bell’s inequality to avoid the need for a demanding cryptographic property called the adaptive hardcore bit, while maintaining essentially no increase in the quantum circuit complexity and no extra assumptions. Leveraging the relaxed cryptographic requirements of the protocol, we present two trapdoor claw-free function constructions, based on Rabin’s function and the Diffie–Hellman problem, which have not been used in this context before. We also present two independent innovations that improve the efficiency of our implementation and can be applied to other quantum cryptographic protocols. First, we give a scheme to discard so-called garbage bits, removing the need for reversibility in the quantum circuits. Second, we show a natural way of performing postselection that reduces the fidelity needed to demonstrate quantum advantage. Combining these results, we describe a blueprint for implementing our protocol on Rydberg atom-based quantum devices, using hardware-native operations that have already been demonstrated experimentally.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Asynchronous Runtimes in Action: An Introspective Framework for a Next Gen Runtime

One of the most critical challenges that new high performance systems face is the lack of system software support for these large scale systems. Investment on system stack components is essential in the development, debugging and optimization of the new emerging programming models. These emerging models have the promise to better utilize the vast hardware resources available in current and future systems. To aid in the development of applications and new system stacks, runtimes, as instances of their respective execution models, need to produce facilities to introspect their inner workings and allow an indepth attribution of performance bottlenecks and computational patterns. In other words, the runtime systems need to reduce their opacity to observers so that users of a novel program execution model can adapt their designs to fit the intended model usage, regardless of the layer that they are working on. This design/development loop (akin to co-design) enables synergistic opportunities across the entire computational stack. This paper presents the design and implementation of a simple “gray” box performance attribution harness running inside a fine grain runtime system: the Open Community Runtime (OCR). We showcase what such a framework can indicate regarding the runtime behavior while running at scale. To this end, we have designed a set of synthetic scenarios aimed to test the runtime at their best and worst cases. We present an analysis of the most important runtime features, properties and idiosyncrasies that will affect the development of new runtime features, algorithmic selection, and application development.

high performance distributed runtimes↗

wa-hls4ml: A Benchmark and Surrogate Models for hls4ml Resource and Latency Estimation

As machine learning (ML) is increasingly implemented in hardware to address real-time challenges in scientific applications, the development of advanced toolchains has significantly reduced the time required to iterate on various designs. These advancements have solved major obstacles, but also exposed new challenges. For example, processes that were not previously considered bottlenecks, such as hardware synthesis, are becoming limiting factors in the rapid iteration of designs. To mitigate these emerging constraints, multiple efforts have been undertaken to develop an ML-based surrogate model that estimates resource usage of ML accelerator architectures. We introduce wa-hls4ml, a benchmark for ML accelerator resource and latency estimation, and its corresponding initial dataset of over 680,000 fully connected and convolutional neural networks, all synthesized using hls4ml and targeting Xilinx FPGAs. The benchmark evaluates the performance of resource and latency predictors against several common ML model architectures, primarily originating from scientific domains, as exemplar models, and the average performance across a subset of the dataset. Additionally, we introduce GNN- and transformer-based surrogate models that predict latency and resources for ML accelerators. We present the architecture and performance of the models and find that the models generally predict latency and resources for the 75% percentile within several percent of the synthesized resources on the synthetic test dataset.

Hawks, Benjamin [Fermilab] (ORCID:0000000157000288↗

wa-hls4ml: A Benchmark and Surrogate Models for hls4ml Resource and Latency Estimation

As machine learning (ML) is increasingly implemented in hardware to address real-time challenges in scientific applications, the development of advanced toolchains has significantly reduced the time required to iterate on various designs. These advancements have solved major obstacles, but also exposed new challenges. For example, processes that were not previously considered bottlenecks, such as hardware synthesis, are becoming limiting factors in the rapid iteration of designs. To mitigate these emerging constraints, multipleefforts have been undertaken to develop an ML-based surrogate model that estimates resource usage of synthesized ML accelerator architectures. We introduce wa-hls4ml, a benchmark for ML accelerator resource and latency estimation, and its corresponding initial dataset of over 680 000 fully connected and convolutional neural networks, all synthesized using hls4ml and targeting Xilinx FPGAs. The benchmark evaluates the performance of resource and latency predictors against several common ML model architectures, primarily originating from scientific domains, as exemplar models, and the average performance across a subset of the dataset. Additionally, we introduce GNN- and transformer-based surrogate models that predict latency and resources for ML accelerators. We present the architecture and performance of the models and find that the models generally predict latency and resources for the 75% percentile within several percent of the synthesized resources on the synthetic test dataset.

Hawks, Benjamin G. [Fermilab]↗

First Argon Gas Puff Experiments With 500 ns Implosion Time On Sphinx Driver

Experiments have been performed at the SPHINX driver to study potential of an Argon Gas Puff load designed by AASC. We present here the gas Puff hardware and results of the last shot series.The Argon Gas Puff load used is injected thanks to a 20 cm diameter nozzle. The nozzle has two annuli and a central jet. The pressure and gas type in each of the nozzle plena can be independently adjusted to tailor the initial gaz density distribution. This latter is selected as to obtain an increasing radial density from outer shell towards the pinch axis in order to mitigate the RT instabilities and to increase radiating mass on axis. A flashboard unit produces a high intensity UV source to pre-ionize the Argon gas. Typical dimensions of the load are 200 mm in diameter and 40 mm height. Pressures are adjusted to obtain an implosion time around 550 ns with a peak current of 3.5 MA.With the goal of improving k-shell yield a mass scan of the central jet was performed and implosion time, mainly given by outer and middle plena settings, was kept constant. Tests were also done to reduce the implosion time for two configurations of the central jet. Strong zippering of the radiation production was observed mainly due to the divergence of the central jet over the 40 mm of the load height. Due to that feature k-shell radiation is mainly obtained near cathode. Therefore tests were done to mitigate this effect first by adjusting local pressure of middle and central jet and second by shortening the pinch length.At the end of this series, best shot gave 5 kJ of Ar k-shell yield. PCD detectors showed that k-shell x-ray power was 670 GW with a FWHM of less than 10 ns.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

FY 2014 Sodium-Supercritical CO 2 Interactions in the SNAKE Experiment Facility

One appealing feature of the supercritical carbon dioxide Brayton cycle energy conversion system is the small footprint that the hardware requires, which is in part due to the use of compact diffusion-bonded heat exchangers such as Printed Circuit Heat Exchanger (PCHEs) or Hybrid Heat Exchangers (H 2 X) as the heat source heat exchanger (sodium-to-CO 2 ) as well as the recuperator and cooler modules. Although PCHEs and H 2 Xs have a high degree of structural integrity, the potential for leaks to develop between the sodium and CO 2 coolant channels in the heat exchanger cannot be ruled out, and this would lead to discharge of high pressure CO 2 into the secondary coolant circuit. Due to the robustness of the PCHE design, catastrophic failure leading to CO 2 jet blowdown into the secondary sodium loop is not deemed likely. Rather, small cracks (or micro-leaks) may develop in which CO 2 will bleed into the secondary system at a relatively low rate and chemically react with the sodium. The goal of the sodium-CO 2 interaction tests is to gain a fundamental understanding of sodium-CO 2 interactions under prototypical conditions of compact diffusion-bonded heat exchanger failure, a fundamental understanding of self-plugging if it occurs, and the development of one-dimensional phenomenological models for the interactions between high-pressure CO 2 issuing into liquid sodium from a micro-leak across a stainless steel pressure boundary. These models will be validated using experiment data. Therefore, an experiment program at Argonne was initiated in Fiscal Year 2010 to investigate the reaction behavior between sodium and CO 2 under micro-leak conditions. Several reports have described the facility scaling rationale and design. Assembly of the SNAKE (S-CO 2 , Na Kinetics Experiment) began in Fiscal Year 2011 and was completed in July 2012. The sodium inventory in this facility consists of approximately 44 lbs (20 kg; ~21 L/5.5 gal).

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Extended Low Load Boiler Operation to Improve Performance and Economics of an Existing Coal Fired Power Plant (Final Report)

The overall goal is to improve the performance and economics of existing coal fired power plants by extending low load boiler operation to lower loads than is currently achievable. The objective of this program is to develop and validate sensor hardware and analytical algorithms to lower plant operating expenses (OPEX) for the currently operating pulverized coal utility boiler fleet. Coal fired utility boilers are increasingly under grid dispatch pressure. In some cases, the coal fired cost of generation is noncompetitive with respect to natural gas generation and subsidized renewable sources. To remain profitable and remain fully compliant with existing environmental regulations, the installed coal fired fleet must find technologies which allow it to move into a more flexible cyclic load dispatch model. Today the installed coal fired utility fleet must be cost of generation competitive, fully emissions compliant, and responsive to the variability inherent in renewable energy generation sources. In the Phase I of the project, GE Steam Power, Inc. (GE) performed modeling of different operating scenarios for low load operation using an existing full plant dynamic model developed for a 660MW steam power plant. Sensors and analytic algorithms to enable a stable and steady coal supply for low load pulverizer operation were identified and tested at the Pulverizer Development Facility (PDF) at GE’s Clean Energy Center in Bloomfield, Connecticut. Sensors and analytic algorithms to enable stable combustion for low load operation were identified and tested at the 15 MWth Industrial Scale Burner facility (ISBF) at GE’s Clean Energy Center. A concept was developed to test the sensors and control algorithms, down selected after testing, at a full-scale coal fired power plant. A budget estimate was then developed, and the concept was implemented at an existing utility power plant. The specific objectives of the experimental work were to: • Identify and select sensors and analytic algorithms for monitoring coal pulverizer operation at lower loads to provide stable operation and appropriate coal fineness at lower coal throughput; Identify and select sensors and analytic algorithms for a Boiler Flame Stability Monitor to better balance air and fuel at each burner. This enables a reduction in a coal boiler’s safe low load power level while maintaining stable flame characteristics; Develop a concept in Phase I for low load operation of a full-scale power plant and develop a budget estimate for testing and execute the test plan at an existing plant in Phase II; Validate the capability of the extended low load boiler system to extend the minimum load operating point in a safe and reliable manner on an existing full-scale utility boiler. At the completion of this experimental study, GE has developed a set of sensors and analytic algorithms, down selected after testing, that have the potential to enable safe low load operation of a utility boiler. GE has also identified a host site for testing these identified sensors and analytic algorithms. GE has generated a full set of deliverables that provide sufficient information to proceed with the next step of testing at a host site. This includes a potential host site and budget estimate for concept testing at host site. In the Phase II of the project, a series of field tests were completed to validate the extended low load boiler operation, which consisted of detailed engineering, installation, commissioning, and testing the additional sensors and analytics for the coal-fired combustion system on an existing full-scale utility boiler. The optimization work has been supported by the host plant and endorsed by their engineering and operation staff.

01 COAL, LIGNITE, AND PEAT↗

Experiences with SYCL on AMD GPUs with Kokkos

With the recent diversification of the hardware landscape in the high-performance computing (HPC) community, performance-portability solutions are becoming more and more important. One of the most popular choices is Kokkos, which recently became a Linux Foundation project. Most of its development is supported by the US Department of Energy and the French Alternative Energies and Atomic Energy Commission. Kokkos is implemented as a C++ library with multiple backends to support CPUs as well as various GPU architectures. These backends include OpenMP, CUDA, HIP, and also SCYL. This approach enables users to leverage the preferred vendor toolchain for the respective platform (e.g. CUDA, ROCm, OneAPI). The SYCL backend is used to target Intel GPUs, in particular to support the Aurora exascale supercomputer. However, SYCL itself also offers a large degree of portability, and in fact Kokkos’ CI for SYCL has been running on NVIDIA hardware due to a lack of access to Intel GPUs. In this report, we describe our experience with using Kokkos SYCL backend on AMD GPUs targeting the Frontier supercomputer at Oak Ridge National Laboratory. The two major SYCL implementations are DPC++ and AdaptiveCpp. While the Kokkos SYCL backend has been implemented using the former, the latter was the first implementation to target AMD GPUs. We will discuss the experience with both of these SYCL implementations in terms of functionality and performance. Using Kokkos to evaluate SYCL toolchains has a number of benefits. Kokkos’ use of SYCL is fairly complex, exercising features such as graphs, relocatable device functions, atomics – including for non-arithmetic types, as well as pinned and page migratable memory allocations. Kokkos also needs to implement capabilities such as Kokkos’ hierarchical parallelism that are not a straight-forward mapping to SYCL capabilities. Furthermore, a large number of libraries and applications that represent diverse use cases are implemented in Kokkos, providing readily available test cases for a toolchain evaluation. Preliminary results show that support for AMD GPUs in DPC++ is much less mature than for NVIDIA GPUs or Intel GPUs. While the situation has improved significantly over the last year, we still encounter many runtime failures, dispatching problems, and code generation issues. With AdaptiveCpp the challenges arise even earlier in the evaluation process. Since Kokkos’ SYCL implementation is largely focused on supporting Intel GPUs, we opted to leverage SYCL extensions which are available in DPC++ but not in AdaptiveCpp. Furthermore, AdaptiveCpp appears to be less conformant with the SYCL2020 standard which Kokkos relies on. In some cases, we are able to work around the lack of feature support, in other cases we have to disable certain Kokkos capabilities to evaluate the toolchain. Our evaluation will leverage Kokkos’ unit tests to establish basic functionality and feature completeness. We then use simple benchmarks for components of a CG implementation as a measure of usability and performance of the SYCL toolchains.

97 MATHEMATICS AND COMPUTING↗

A novel vibration-based fault diagnostic algorithm for gearboxes under speed fluctuations without rotational speed measurement

The localized failures of gears introduce cyclic-transient impulses in the measured gearbox vibration signals. These impulses are usually identified from the sidebands around gear-mesh harmonics through the spectral analysis of cyclo-stationary signals. However, in practice, several high-powered applications of gearboxes like wind turbines are intrinsically characterized by nonstationary processes that blur the measured vibration spectra of a gearbox and deteriorate the efficacy of spectral diagnostic methods. Although order-tracking techniques have been proposed to improve the performance of spectral diagnosis for nonstationary signals measured in such applications, the required hardware for the measurement of rotational speed of these machines is often unavailable in industrial settings. Moreover, existing tacho-less order-tracking approaches are usually limited by the high time-frequency resolution requirement, which is a prerequisite for the precise estimation of the instantaneous frequency. To address such issues, a novel fault-signature enhancement algorithm is proposed that can alleviate the spectral smearing without the need of rotational speed measurement. This proposed tacho-less diagnostic technique resamples the measured acceleration signal of the gearbox based on the optimal warping path evaluated from the fast dynamic time-warping algorithm, which aligns a filtered shaft rotational harmonic signal with respect to a reference signal assuming a constant shaft rotational speed estimated from the approximation of operational speed. The effectiveness of this method is validated using both simulated signals from a fixed-axis gear pair under nonstationary conditions and experimental measurements from a 750-kW planetary wind turbine gearbox on a dynamometer test rig. Lastly, the results demonstrate that the proposed algorithm can identify fault information from typical gearbox vibration measurements carried out in a resource-constrained industrial environment.

17 WIND ENERGY↗

Profiling the BLAST bioinformatics application for load balancing on high-performance computing clusters

Abstract Background The Basic Local Alignment Search Tool (BLAST) is a suite of commonly used algorithms for identifying matches between biological sequences. The user supplies a database file and query file of sequences for BLAST to find identical sequences between the two. The typical millions of database and query sequences make BLAST computationally challenging but also well suited for parallelization on high-performance computing clusters. The efficacy of parallelization depends on the data partitioning, where the optimal data partitioning relies on an accurate performance model. In previous studies, a BLAST job was sped up by 27 times by partitioning the database and query among thousands of processor nodes. However, the optimality of the partitioning method was not studied. Unlike BLAST performance models proposed in the literature that usually have problem size and hardware configuration as the only variables, the execution time of a BLAST job is a function of database size, query size, and hardware capability. In this work, the nucleotide BLAST application BLASTN was profiled using three methods: shell-level profiling with the Unix “time” command, code-level profiling with the built-in “profiler” module, and system-level profiling with the Unix “gprof” program. The runtimes were measured for six node types, using six different database files and 15 query files, on a heterogeneous HPC cluster with 500+ nodes. The empirical measurement data were fitted with quadratic functions to develop performance models that were used to guide the data parallelization for BLASTN jobs. Results Profiling results showed that BLASTN contains more than 34,500 different functions, but a single function, RunMTBySplitDB, takes 99.12% of the total runtime. Among its 53 child functions, five core functions were identified to make up 92.12% of the overall BLASTN runtime. Based on the performance models, static load balancing algorithms can be applied to the BLASTN input data to minimize the runtime of the longest job on an HPC cluster. Four test cases being run on homogeneous and heterogeneous clusters were tested. Experiment results showed that the runtime can be reduced by 81% on a homogeneous cluster and by 20% on a heterogeneous cluster by re-distributing the workload. Discussion Optimal data partitioning can improve BLASTN’s overall runtime 5.4-fold in comparison with dividing the database and query into the same number of fragments. The proposed methodology can be used in the other applications in the BLAST+ suite or any other application as long as source code is available.

59 BASIC BIOLOGICAL SCIENCES↗

SU-E-T-638: Evaluation and Comparison of Landauer Microstar (OSLD) Readers

Purpose: To evaluate and compare characteristic performance of a new Landauer nanodot Reader with the previous model. Methods: In order to calibrate and test the reader, a set of nanodots were irradiated using a Varian Truebeam Linac. Solid water slabs and bolus were used in the process of irradiation. Calibration sets of nanodots were irradiated for radiation dose ranges: 0 to 10 and 20 to 1000 cGy, using 6MV photons. Additionally, three sets of nanodots were each irradiated using 6MV, 10MV and 15MV beams. For each beam energy, and selected dose in the range of 3 to 1000 cGy, a pair of nanodots was irradiated and three readings were obtained with both readers. Results: The analysis shows that for 3 photon beam energies and selected ranges of dose, the calculated absorbed dose agrees well with the expected value. The results illustrate that the new Microstar II reader is a highly consistent system and that the repeated readings provide results with a reasonably small standard deviation. For all practical purposes, the response of system is linear for all radiation beam energies. Conclusion: The Microstar II nanodot reader is consistent, accurate, and reliable. The new hardware design and corresponding software contain several advantages over the previous model. The automatic repeat reading mechanism, that helps improve reproducibility and reduce processing time, and the smaller unit size that renders ease of transport, are two of such features. Present study shows that for high dose ranges a polynomial calibration equation provides more consistent results. A 3rd order polynomial calibration curve was used to analyze the readings of dosimeters exposed to high dose range radiation. It was observed that the results show less error compared to those calculated by using linear calibration curves, as provided by Landauer system software for all dose ranges.

60 APPLIED LIFE SCIENCES↗

WE-G-18A-02: Calibration-Free Combined KV/MV Short Scan CBCT

Purpose: To combine orthogonal kilo-voltage (kV) and Mega-voltage (MV) projection data for short scan cone-beam CT to reduce imaging time on current radiation treatment systems, using a calibration-free gain correction method. Methods: Combining two orthogonal projection data sets for kV and MV imaging hardware can reduce the scan angle to as small as 110° (90°+fan) such that the total scan time is ∼18 seconds, or within a breath hold. To obtain an accurate reconstruction, the MV projection data is first linearly corrected using linear regression using the redundant data from the start and end of the sinogram, and then the combined data is reconstructed using the FDK method. To correct for the different changes of attenuation coefficients in kV/MV between soft tissue and bone, the forward projection of the segmented bone and soft tissue from the first reconstruction in the redundant region are added to the linear regression model. The MV data is corrected again using the additional information from the segmented image, and combined with kV for a second FDK reconstruction. We simulated polychromatic 120 kVp (conventional a-Si EPID with CsI) and 2.5 MVp (prototype high-DQE MV detector) projection data with Poisson noise using the XCAT phantom. The gain correction and combined kV/MV short scan reconstructions were tested with head and thorax cases, and simple contrast-to-noise ratio measurements were made in a low-contrast pattern in the head. Results: The FDK reconstruction using the proposed gain correction method can effectively reduce artifacts caused by the differences of attenuation coefficients in the kV/MV data. The CNRs of the short scans for kV, MV, and kV/MV are 5.0, 2.6 and 3.4 respectively. The proposed gain correction method also works with truncated projections. Conclusion: A novel gain correction and reconstruction method was developed to generate short scan CBCT from orthogonal kV/MV projections. This work is supported by NIH Grant 5R01CA138426-05.

60 APPLIED LIFE SCIENCES↗