Search NASA⌕ Search

SEARCH · Search NASA

Results for “AUTO Program”

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 91 records · Page 5

Critical Points Based Register-Concurrency Autotuning for GPUs

The unprecedented prevalence of GPGPU is largely attributed to its abundant on-chip register resources, which allow massively concurrent threads and extremely fast context switch. However, due to internal memory capacity constraints, there is a tradeoff between the per-thread register usage and the overall concurrency. This becomes a design problem in terms of performance tuning, since the performance “sweet spot” which can be significantly affected by these two factors is generally unknown beforehand. In this paper, we propose an effective autotuning solution to quickly and efficiently select the optimal number of registers perthread for delivering the best GPU performance. Experiments on three generations of GPUs (Nvidia Fermi, Kepler and Maxwell) demonstrate that our simple strategy can achieve an average of 10% performance improvement while a max of 50% over the original version without modifying the user program. Additionally, to reduce local cache misses due to register spilling and further improve performance, we explore three optimization schemes (i.e. bypass L1 for global memory access, enlarge local L1 cache and spill into shared memory) and discuss their impact on performance on a Kepler GPU.

auto-tuning↗

Synthesis and thermoluminescence studies of γ-irradiated Dy{sup 3+} doped SrGd{sub 2}O{sub 4} phosphor

Highlights: • SrGd2O4:Dy3+ phosphors have been prepared via homogeneous precipitation cum auto-combustion method for the first time. • TL characteristics of γ-irradiated SrGd2O4: Dy3+ phosphors were studied for various Dy3+ contents and heating rates. • Linear TL profile, less fading and high reproducibility of the present phosphor system reveals its applicability in TL dosimetry. - Abstract: Dy{sup 3+} doped SrGd{sub 2}O{sub 4} phosphors were synthesized via homogeneous precipitation cum auto-combustion method for the first time. Orthorhombic phase formation and granular morphology of SrGd{sub 2}O{sub 4}: Dy{sup 3+} samples were investigated through X-ray diffraction (XRD) and microscopic studies. Thermoluminescence (TL) characteristics of SrGd{sub 2}O{sub 4}: Dy{sup 3+} phosphors were taken after exposure of {sup 60}Co γ-rays. The TL glow curve of as-prepared phosphors was well-resolved into five peaks 114 °C, 161 °C, 198 °C, 248 °C, and 293 °C by using computerized glow curve deconvolution program. The effect of impurity (Dy{sup 3+}) contents and heating rates for the TL glow curves were also studied. The linear behaviour of as-formed SrGd{sub 2}O{sub 4}: Dy{sup 3+} samples are observed within a wide range of dose of 80 Gy–2 kGy. Fading characteristics of SrGd{sub 2}O{sub 4}: Dy{sup 3+} phosphor was studied over the duration of 30 days. A comparative study was done to estimate different trapping parameters by employing Chen’s peak shape method and CGCD simulation of TL glow curves.

36 MATERIALS SCIENCE↗

Investigating Application of LiDAR for Nuclear Power Plants

Many evaluation, assessment, and modeling tasks at nuclear power plants require spatial in-formation; this often requires physical visits to locations within the facility because the 2D or 3D schematics and current models do not contain enough detail or do not capture as-built and real-world conditions. These visits require extensive manual labor for not only the requesting party, but also support groups, such as security. Light Detection and Ranging (LiDAR) mapping is trying to solve that problem by providing very detailed 3D models for low costs. However, the use of these models can be very limited because either component reference information is missing and too costly to add or there is no way to extract specific spatial data needed for other tools. This report outlines two main efforts. First, to reduce the effort of “Tagging” data in large 3D models, a general Application Programming Interface (API) was developed to import a variety of existing plant database information into a 3D-visualization engine. Filters allow the user to have only zone-specific items listed; then, they can simply click and assign the information to a specific spot or component in the 3D model. The second part of the work is the development of an interface for importing pieces from the3D-LiDAR model into other systems needed for modeling and simulation, outlined around fire modeling. This interface allows for the retrieval of item location and boundaries, enabling the auto generation of models for varying tools.

3D Modelling↗

Critical and Californium Source-Driven Noise Analysis Subcritical Measurements with an Unreflected Cylindrical Tank of Mixed Uranium-Plutonium Nitrate Solution

The reported experiments were performed by Oak Ridge National Laboratory (ORNL) at the Battelle Northwest Laboratory’s (now Pacific Northwest National Laboratory) critical experiments facility at Hanford, Washington in 1981 and used 16 days of critical facility time, not including 10 days for setup and removal of ORNL equipment. These measurements were to assess the capability of the Cf source-driven noise analysis (CSDNA) method to measure the subcriticality (keff) of mixed U-Pu nitrate solutions. In addition to the CSNDA measurements, measurements were also performed near delayed criticality where CSDNA measurements cannot be performed. This report documents the experiments that were not reported at that time by presenting the ORNL experimental results and any online analysis performed during and shortly after the measurements. The mixed nitrate solution had a U concentration at 188 grams per liter (g/L), a Pu concentration of 280 g/L, free acid normality of 2.80, H ion molarity of 5.9 and a specific gravity of 1.754 g/cm 3 , a 240 Pu isotopic content of 7.981 wt. %, and a 235 U isotopic content of 0.724 wt. %. The stainless-steel tank for the solution had an inside diameter of 35.38 cm, an outside diameter of 35.53 cm, a height of 56.72 cm, and bottom thickness of 0.9525 cm. A Zircaloy pipe with a 3.1496 cm outside diameter, a 2.7788 cm inside diameter, and bottom thickness of 0.635 cm was available for insertion of the Cf source in the center of the fissile solution. The Cf source was also located at the outside surface of the tank (solution height varied from 10 to 53 cm) and in the center of the solution (solution height varied from 10 to 60.7 cm). The CSDNA measurements were not analyzed online to determine the subcritical neutron multiplication factors. At all subcritical states, the break frequency noise analysis data was fitted to obtain the prompt neutron decay constant. The neutron multiplication factors were determined for the two configurations of the measurements near delayed criticality. The subcritical neutron multiplication factors from the CNSDA measurements can be obtained with further analysis. However, the near delayed critical configuration, the prompt neutron decay constants, the count rates, and the measured cross and auto power spectral densities can be calculated directly for benchmarking. Much of data presented in this report are from ORNL notes—not in the ORNL logbooks. For the final benchmark analysis, the data from the Battelle Northwest Laboratory (which operated the critical facility in 1981) critical facility logbook should be consulted and be incorporated where appropriate. The purpose of this report is to document the experimental information for the measurements performed so that at a later date researchers could perform the required uncertainty and calculational analyses and documentation to use these data for an International Criticality Safety Benchmark Program (ICSBEP) or Nuclear Energy Agency benchmark. The data from these measurements are available from the ORNL Records Management Services Department, and the logbook is available from ICSBEP at Idaho National Laboratory. Preparation of the present report is part of a larger cooperative effort between Idaho National Laboratory (INL) and Oak Ridge National Laboratory (ORNL) to document more than 15 undocumented critical and subcritical experiments enumerated in ORNL/TM-2019/18 and performed by ORNL at ORCEF and other USDOE critical experiments facilities using more than 500 operational days of critical facility time.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

Fuel Property Characterization and Prediction

This presentation describes the progress made over the last year on the Fuel Property Characterization and Prediction tasks as part of the CoOptima program under the Vehicle Technologies Office. The tasks reviewed are: Fuel Property Blending Model (F.1.2.2 and F.1.2.3), the Heat of Vaporization Measurement for Full Boiling Range Fuels (F.1.3.2), and the Flow Reactor Autoignition Kinetic Mechanism Development and Validation (F.2.1.4) tasks.

33 ADVANCED PROPULSION SYSTEMS↗

Next Generation RivGen Power System: Kvichak River, AK Overwinter Ice Study

The University of Alaska Fairbanks (UAF) Alaska Hydrokinetic Energy Research Center was tasked with developing a real-time data telemetry / remote power generation system to monitor frazil ice conditions in the Kvichak River in support of the U.S. Department of Energy funded "Next Generation MHK River Power System Optimized for Performance, Durability and Survivability" project. A real-time telemetry system was requested because of the short time span between the end of the frazil ice season when the instruments would be recovered, limited vessel availability and the project end-date. To meet the project objectives, UAF designed and assembled a remote power/real-time data telemetry system that included an auto start propane generator, a small PV array, a small battery bank and line-of-sight radios as well as two sonar systems to monitor river velocity and water column acoustic backscatter strength. Both sonars included internal batteries for powering the instruments in case of failure of the shore based power system. The sonars, deployed in ~5 m of water on the bed of the Kvichak River, adjacent to the Village of Igiugig, Alaska were tethered to shore via a waterproof armored cable that conveyed power to the subsurface instruments and data from the instruments to the shore based telemetry system. The instruments were programmed to record data internally as well as to transmit data serially over the cables to the shore based system. The system was in-place between November, 2016 and June, 2017. While the real-time data telemetry system was not successful and the remote power generation power system was only partially successful, the system design included sufficient redundant power in the form of internal instrument batteries to enable the collection of nearly three months of overlapping velocity and backscatter data (from November through February) and a record of acoustic backscatter strength spanning the entire ~150 day frazil ice season between November, 2016 and ~April, 2017. This submission includes the overwinter ice study plan, raw data, and report.

16 TIDAL AND WAVE POWER↗

Automatic Thread-Level Parallelization in the Chombo AMR Library

The increasing on-chip parallelism has some substantial implications for HPC applications. Currently, hybrid programming models (typically MPI+OpenMP) are employed for mapping software to the hardware in order to leverage the hardware?s architectural features. In this paper, we present an approach that automatically introduces thread level parallelism into Chombo, a parallel adaptive mesh refinement framework for finite difference type PDE solvers. In Chombo, core algorithms are specified in the ChomboFortran, a macro language extension to F77 that is part of the Chombo framework. This domain-specific language forms an already used target language for an automatic migration of the large number of existing algorithms into a hybrid MPI+OpenMP implementation. It also provides access to the auto-tuning methodology that enables tuning certain aspects of an algorithm to hardware characteristics. Performance measurements are presented for a few of the most relevant kernels with respect to a specific application benchmark using this technique as well as benchmark results for the entire application. The kernel benchmarks show that, using auto-tuning, up to a factor of 11 in performance was gained with 4 threads with respect to the serial reference implementation.

97 MATHEMATICS AND COMPUTING↗

Performance of solar leasing for low- and middle-income customers in Connecticut

Policymakers are increasingly interested in expanding access to rooftop photovoltaic systems. This study analyzes the financial performance of a Connecticut Green Bank (CGB) solar leasing program, run in partnership with PosiGen, that targets low- and moderate-income customers. We show that this program has successfully reached underserved customers and has reasonable repayment rates given the credit characteristics of the participants. The CGB/PosiGen program reaches many more underserved customers than other PV financing programs in Connecticut.For example, the majority of CGB/PosiGen participants (58%) live in census tracts that have a median income of less than 80% of the area median income (AMI). In contrast, only 9% of participants in the other CGB solar financing programs live in these census tracts. Furthermore, the majority of CGB/PosiGen participants (56%) have FICO scores that would generally be considered non-prime (<670), whereas only 2% of participants in the other programs have similarly low scores. Credit, not income, is the primary factor that explains participants’ financial performance. Overall, we find that delinquency and annualized losses are higher for PosiGen (2.3% and 0.9%) than for other CGB programs (1.4% and 0.1%). Across the CGB programs, lower credit scores are associated with higher rates of delinquency and loss. Therefore, participants’ lower credit scores explain much of the program’s higher rates of delinquency and annualized losses. PosiGen leases perform competitively with market-rate solar and non-solar leases and loans. When compared to securities backed by market-rate PV loans and leases with similar amounts of seasoning, we find that the PosiGen leases have higher delinquency rates but comparable gross loss rates. The similarity in losses is notable given that loss rates for PosiGen leases were higher than those of the other CGB leases and loans. Rather than the PosiGen leases having high loss rates, other CGB leases and loans have unusually low loss rates. We also compare PosiGen to non-solar benchmarks, including indices of auto and consumer loans. We find that the PosiGen leases have significantly less delinquency than non-prime auto loans and have performance comparable to many consumer loans. The U.S. Department of Energy’s Solar Energy Technologies Office supported this research.

14 SOLAR ENERGY↗

Boosting RDataFrame performance with transparent bulk event processing

RDataFrame is ROOT’s high-level interface for Python and C++ data analysis. Since it first became available, RDataFrame adoption has grown steadily and it is now poised to be a major component of analysis software pipelines for LHC Run 3 and beyond. Thanks to its design inspired by declarative programming principles, RDataFrame enables the development of highperformance, highly parallel analyses without requiring expert knowledge of multi-threading and I/O: user logic is expressed in terms of self-contained, small computation kernels tied together by a high-level API. This design completely decouples analysis logic from its actual execution, and opens several interesting avenues for workflow optimization. In particular, in this work we explore the benefits of moving internal data processing from an event-by-event to a bulkby-bulk loop. This refactoring dramatically reduces the framework’s runtime overheads; in collaboration with the I/O layer it improves data access patterns; it exposes information that optimizing compilers might use to auto-vectorize the invocation of user-defined computations; finally, while existing user-facing interfaces remain unaffected, it becomes possible to additionally offer interfaces that explicitly expose bulks of events, useful e.g. for the injection of GPU kernels into the analysis workflow. In order to inform similar future R&D, design challenges will be presented, as well as an investigation of the relevant timememory trade-off backed by novel performance benchmarks.

Guiraud, Enrico↗

Domain-Specific Type-Safe APIs for Hierarchical Scientific Data with Modern C++

General-purpose library application programming interfaces (APIs) for self-describing hierarchical scientific data storage, such as the HDF5 and NetCDF libraries, are traditionally of runtime nature. Runtime errors for entry existence and data types are typically caught later in the development process of higher-level application-specific APIs. In this paper, we propose exploiting modern C++ metaprogramming features to add compile-time type-safety to improve the interaction with a well-defined metadata-rich scientific schema in domain-specific hierarchical datasets. We tackle two aspects of common use: (i) direct data access, (ii) flexible “in-memory” index models for efficient search and data processing. The proposed APIs use C++17’s template type auto deduction features, C++11’s enum class for type-safety and C-style preprocessor macros for generative templated code. We showcase the pros and cons of our initial work on the standard NeXus schema used for annotating and storing experimental neutron scattering data at several facilities around the world on top of HDF5. Extendable compile-time type-safe APIs are a desirable feature that could be indexed by any modern integrated development environment (IDE). Hence, such APIs can help ease the learning curve for domain scientists using a less error-prone software interaction to enhance the findability of their data without resorting to a domain-specific language (DSL).

Godoy, William↗

Towards Exascale: High Performance Visualization and Analytics – Project Status Report

The LBNL Visualization research program focuses on a key set of problems facing our community as we move towards the exascale regime of computing. First, we aim to better understand how to effectively take advantage of an evolving architectural landscape where there are an increasing number of cores per processor by examining algorithmic reformulation and alternative methods of programming these increasingly complex platforms. Second, as the gap widens between computational and I/O capacity, it is increasingly imperative to perform more visual data exploration and analysis operations concurrent with the simulation to avoid increasingly expensive I/O. Our approach is to examine several interrelated issues aimed at minimizing the memory footprint, data movement, and design/implementation issues related to deploying tools that can have broad applicability across many different simulation codes and science domains. Third, increasing computational power produces larger and more complex scientific data, and our research program focuses on new technologies aimed at enabling scientific data understanding with ever-larger data. Here, we explore how multi-resolution data representation techniques can reduce data movement, and study novel approaches that focus visualization and analysis processing on subsets of data that are “scientifically interesting,” thereby also minimizing data movement while accelerating scientific knowledge discovery.

97 MATHEMATICS AND COMPUTING↗

The CompTox Chemistry Dashboard: a community data resource for environmental chemistry

Despite an abundance of online databases providing access to chemical data, there is increasing demand for highquality, structure-curated, open data to meet the various needs of the environmental sciences and computational toxicology communities. The U.S. Environmental Protection Agency’s (EPA) web-based CompTox Chemistry Dashboard is addressing these needs by integrating diverse types of relevant domain data through a cheminformatics layer, built upon a database of curated substances linked to chemical structures. These data include physicochemical, environmental fate and transport, exposure, usage, in vivo toxicity, and in vitro bioassay data, surfaced through an integration hub with link-outs to additional EPA data and public domain online resources. Batch searching allows for direct chemical identifer (ID) mapping and downloading of multiple data streams in several diferent formats. This facilitates fast access to available structure, property, toxicity, and bioassay data for collections of chemicals (hundreds to thousands at a time). Advanced search capabilities are available to support, for example, non-targeted analysis and identifcation of chemicals using mass spectrometry. The contents of the chemistry database, presently containing ~ 760,000 substances, are available as public domain data for download. The chemistry content underpinning the Dashboard has been aggregated over the past 15 years by both manual and auto-curation techniques within EPA’s DSSTox project. DSSTox chemical content is subject to strict quality controls to enforce consistency among chemical substance-structure identifers, as well as list curation review to ensure accurate linkages of DSSTox substances to chemical lists and associated data. The Dashboard, publicly launched in April 2016, has expanded considerably in content and user trafc over the past year. It is continuously evolving with the growth of DSSTox into high-interest or data-rich domains of interest to EPA, such as chemicals on the Toxic Substances Control Act listing, while providing the user community with a fexible and dynamic web-based platform for integration, processing, visualization and delivery of data and resources. The Dashboard provides support for a broad array of research and regulatory programs across the worldwide community of toxicologists and environmental scientists.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Perovskite oxides for redox oxidative cracking of n -hexane under a cyclic redox scheme

Steam cracking of naphtha is a commercially proven technology for light olefin production and the primary source of ethylene in the Europe and Asia-Pacific markets. However, its significant energy consumption and high CO 2 intensity (up to 2 tons CO 2 /ton C 2 H 4 ), stemming from endothermic cracking reactions and complex product separations, make this state-of-the-art process increasingly undesirable from an environmental standpoint. We propose a redox oxidative cracking (ROC) approach as an alternative pathway for naphtha conversion. Enabled by perovskite oxide-based redox catalysts, the ROC process converts naphtha (represented by n-hexane) in an auto-thermal, cyclic redox mode. The addition of 20 wt.% Na 2 WO 4 to SrMnO 3 and CaMnO 3 created highly selective redox catalysts capable of achieving enhanced olefin yields from n-hexane oxy-cracking. This was largely attributed to the redox catalysts’ high activity, selectivity, and stability towards selective hydrogen combustion (SHC) under a redox mode. Na 2 WO 4 /CaMnO 3 demonstrated significantly higher olefin yield (55–58%) when compared to that from thermal cracking (34%) at 725 °C and 4500 h –1 . CO x yield as low as 1.7% was achieved along with complete combustion of H 2 over 25 cycles. Similarly, Na 2 WO 4 /SrMnO 3 achieved 41% olefin yield, 0.4% CO x yield, and 73% H 2 combustion at this condition. Oxygen-temperature-programmed desorption (O 2 -TPD) indicated that Na 2 WO 4 hindered gaseous oxygen release from CaMnO 3 . Low-energy ion scattering (LEIS) and X-ray photoelectron spectroscopy (XPS) revealed an outermost perovskite surface layer covered by Na 2 WO 4 , which suppressed near-surface Mn and alkaline earth metal cations. The formation of non-selective surface oxygen species was also inhibited. XPS analysis further confirmed that promotion of SrMnO 3 with Na 2 WO 4 suppressed surface Sr species by 90%, with a similar effect also observed on CaMnO 3 . Furthermore, these findings point to the promoting effect of Na 2 WO 4 and the potential of promoted SrMnO 3 and CaMnO 3 as selective redox catalysts for efficient production of light olefins from naphtha via the ROC process.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

A Pattern Based Algorithmic Autotuner for Graph Processing on GPUs

This paper presents Gswitch, a pattern-based algorithmic autotuning system that dynamically switches to the suitable optimization variants with negligible overhead. Its novelty is a small set of algorithmic patterns that enables configurable assembling of algorithm variants. The fast transition of Gswitch is based on a machine learning model trained from 644 real graphs from the network repository. In addition, Gswitch provides succinct programming interface which hides all low-level tuning details. We evaluate Gswitch for typical graph algorithms (BFS, CC, PR, SSSP, and BC) on Nvidia Kepler and Pascal GPUs. The results show that Gswitch runs up to 10× faster than the best configuration of the state-of-the-art programmable GPU-based graph processing libraries on ten representative graphs. Gswitch wins on 92.4% cases of 644 graph data which is the largest dataset evaluation reported to date.

HPC, graph algorithms, GPUs, auto tuning↗

TPR investigations on the reducibility of Cu supported on Al{sub 2}O{sub 3}, zeolite Y and SAPO-5

Reducibility of Cu supported on Al{sub 2}O{sub 3}, zeolite Y and silicoaluminophosphate SAPO-5 has been investigated in dependence on the Cu content using a method combining conventional temperature programmed reduction (TPR) by hydrogen with reoxidation in N{sub 2}O followed by a second the so-called surface-TPR (s-TPR). The method enables discrimination and a quantitative estimation of the Cu oxidation states +2, +1 and 0. The quantitative results show that the initial oxidation state of Cu after calcination in air at 400 deg. C, independent on the nature of the support, is predominantly +2. Cu{sup 2+} supported on Al{sub 2}O{sub 3} is quantitatively reduced by hydrogen to metallic Cu{sup 0}. Comparing the TPR of the samples calcined in air and that of samples additionally pre-treated in argon reveals that in zeolite Y and SAPO-5 Cu{sup 2+} cations are stabilized as weakly and strongly forms. In both systems, strongly stabilized Cu{sup 2+} ions are not auto-reduced by pre-treatment in argon at 650 deg. C, but are reduced in hydrogen to form Cu{sup +}. The weakly stabilized Cu{sup 2+} ions, in contrast, may be auto-reduced by pre-treatment in argon at 650 deg. C forming Cu{sup +} but are reduced in hydrogen to metallic Cu{sup 0}. - Graphical Abstract: TPR, TPR-(act), s-TPR and s-TPR-(act) profiles of (1.36)Cu/SAPO-5 (A), (4.55)Cu/SAPO-5 (B) and (9.19)Cu/SAPO-5 (C) samples. The intensities of TCD signals in (A) and (B) are multiplied by 4.5 and 2, respectively. Highlights: > Cu supported on SAPO-5, alumina and zeolite Y was investigated by TPR and s-TPR. > Cu oxidation states can be discriminated and quantitatively determined. > In zeolite Y and SAPO-5 Cu{sup 2+} cations are stabilized at weak and strong forms. > Strongly stabilized Cu{sup 2+} cannot be auto-reduced in argon at 650 deg. C.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Teuchos C++ memory management classes, idioms, and related topics, the complete reference : a comprehensive strategy for safe and efficient memory management in C++ for high performance computing.

The ubiquitous use of raw pointers in higher-level code is the primary cause of all memory usage problems and memory leaks in C++ programs. This paper describes what might be considered a radical approach to the problem which is to encapsulate the use of all raw pointers and all raw calls to new and delete in higher-level C++ code. Instead, a set of cooperating template classes developed in the Trilinos package Teuchos are used to encapsulate every use of raw C++ pointers in every use case where it appears in high-level code. Included in the set of memory management classes is the typical reference-counted smart pointer class similar to boost::shared ptr (and therefore C++0x std::shared ptr). However, what is missing in boost and the new standard library are non-reference counted classes for remaining use cases where raw C++ pointers would need to be used. These classes have a debug build mode where nearly all programmer errors are caught and gracefully reported at runtime. The default optimized build mode strips all runtime checks and allows the code to perform as efficiently as raw C++ pointers with reasonable usage. Also included is a novel approach for dealing with the circular references problem that imparts little extra overhead and is almost completely invisible to most of the code (unlike the boost and therefore C++0x approach). Rather than being a radical approach, encapsulating all raw C++ pointers is simply the logical progression of a trend in the C++ development and standards community that started with std::auto ptr and is continued (but not finished) with std::shared ptr in C++0x. Using the Teuchos reference-counted memory management classes allows one to remove unnecessary constraints in the use of objects by removing arbitrary lifetime ordering constraints which are a type of unnecessary coupling [23]. The code one writes with these classes will be more likely to be correct on first writing, will be less likely to contain silent (but deadly) memory usage errors, and will be much more robust to later refactoring and maintenance. The level of debug-mode runtime checking provided by the Teuchos memory management classes is stronger in many respects than what is provided by memory checking tools like Valgrind and Purify while being much less expensive. However, tools like Valgrind and Purify perform a number of types of checks (like usage of uninitialized memory) that makes these tools very valuable and therefore complement the Teuchos memory management debug-mode runtime checking. The Teuchos memory management classes and idioms largely address the technical issues in resolving the fragile built-in C++ memory management model (with the exception of circular references which has no easy solution but can be managed as discussed). All that remains is to teach these classes and idioms and expand their usage in C++ codes. The long-term viability of C++ as a usable and productive language depends on it. Otherwise, if C++ is no safer than C, then is the greater complexity of C++ worth what one gets as extra features? Given that C is smaller and easier to learn than C++ and since most programmers don't know object-orientation (or templates or X, Y, and Z features of C++) all that well anyway, then what really are most programmers getting extra out of C++ that would outweigh the extra complexity of C++ over C? C++ zealots will argue this point but the reality is that C++ popularity has peaked and is becoming less popular while the popularity of C has remained fairly stable over the last decade22. Idioms like are advocated in this paper can help to avert this trend but it will require wide community buy-in and a change in the way C++ is taught in order to have the greatest impact. To make these programs more secure, compiler vendors or static analysis tools (e.g. klocwork23) could implement a preprocessor-like language similar to OpenMP24 that would allow the programmer to declare (in comments) that certain blocks of code should be ''pointer-free'' or allow smaller blocks to be 'pointers allowed'. This would significantly improve the robustness of code that uses the memory management classes described here.

99 GENERAL AND MISCELLANEOUS↗

Agentic AI vs ML-Based Autotuning: A Comparative Study for Loop Reordering Optimization

High Performance Computing (HPC) applications rely heavily on code optimizations to achieve good performance on modern CPU and GPU architectures. Traditional Machine Learning auto-tuning approaches have demonstrated success in exploring high-dimensional spaces, but they often require expensive compile-run evaluations and lack adaptability for large HPC applications. The recent advances in Large Language Models (LLMs) and Agentic AI systems raise intriguing questions about the potential of these approaches to address specific optimization methodologies. This work aims to answer an essential question for the HPC community: “How Agentic AI Systems Compare to Traditional ML Autotuning Techniques?” To address this question, we present a comparative analysis between a traditional ML-based optimization approach and an Agentic AI system, evaluating their respective capabilities and limitations for loop-level optimization. In addition, we introduced a new Agentic AI system named LoopGen-AI using three different Large Language Models: GPT-4.1, Claude 4.0, and Gemini 2.5. A key finding is that LoopGen-AI achieves competitive per-formance with only a few program runs, the reasoning logs from the agents revealed that their decisions rely heavily on the combination of semantic understanding of the target kernel with dynamic feedback from the environment, highlighting a promising new dimension in performance tuning. In contrast, ML-based autotuners focus on statistical exploration, and require orders of magnitude more runs to reach peak performance. Additionally, our analysis shows that prompt engineering, particularly using Persona + Context Manager patterns, significantly impacts the effectiveness of Agentic AI. Our results indicate that while Agentic AI systems are not yet a complete replacement for ML-based autotuners, it can effectively complement traditional methods.

Rosas, Miguel Romero↗

Fuel Property Characterization and Prediction

The Fuel Property Characterization and Prediction effect with the Co-Optimization of Fuels and Engines (Co-Optima) initiative is focused on the measurement or prediction of key fuel properties. Efforts under the Vehicle Technologies Office (VTO) have focused on understanding how fuel evaporation phenomena can impact auto-ignition, mixing, and pollutant formation by studying the most important azeotropic interactions between alcohols and gasoline components utilizing a novel Differential Scanning Calorimetry/Thermogravimetric Analysis instrument coupled to a Mass Spectrometer (DSC/TGA/MS). In another effort, Nuclear Magnetic Resonance (NMR) measurements were utilized to understand how oxygenate clusters and networks contribute to vapor pressure effects. This is critical for predicting fuel droplet evaporation and effects on combustion. A method was developed on the Advanced Fuel Ignition Delay Analyzer (AFIDA) to rapidly measure the Research Octane Number (RON) and Sensitivity (S) of samples utilizing small volumes, which will aid in the pace of development of new fuels having targeted properties. Additionally, the AFIDA was employed to develop a method for fast measurement of phi-sensitivity in a bench-scale test that will allow studies of molecular structure effects on mechanism kinetics of this poorly understood kinetic phenomena. A flow reactor was also utilized to investigate the mechanistic basis of phi-sensitivity by identifying and quantitating key species in autoignition mechanisms that can explain phi-sensitivity. The outcomes of these efforts will help ensure that the Co-Optima program can advance the underlying science needed to develop biomass-derived-fuel and advanced engine technologies that will work in tandem to achieve Co-Optima's efficiency, environmental, and economic goals.

33 ADVANCED PROPULSION SYSTEMS↗