SEARCH · Search NASA
Results for “APU”
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.
Unleashing the Power of Physically Unified CPU-GPU Memory in AMD MI300A APU in HPC Applications
Explore the source record for details and available documents.
Experimental review of the performances of protective coatings for interconnects in solid oxide fuel cells
Ferritic stainless-steel interconnects are used in solid oxide fuel cells; however, coatings are required to improve their performance. Although several types of coatings have been proposed, they have been scarcely investigated under similar conditions. This study compares the characteristics of uncoated Crofer 22 APU and eight different coatings on Crofer 22 APU for up to 3,000 hours at 800°C. The coatings were deposited at various research laboratories around the world, and the experiments were performed at Chalmers University of Technology, Sweden. Cross-sections of the samples were analysed using scanning electron microscopy and energy-dispersive x-ray spectroscopy. The (Co,Mn)-based coated steels showed more than 50-fold lower chromium evaporation and at least 3 times thinner Cr 2 O 3 scale thickness compared to uncoated steel. The coated steel samples showed lower area-specific resistance (ASR) values than the uncoated steel after 3,000 hours of exposure, irrespective of the coating thickness, composition and deposition method.
Performance Impact and Trade-Offs for Tuning Key Architectural Parameters on CPU+GPU Systems
In this work, we performed an initial design space exploration of an accelerated processing unit (APU)—a hybrid CPU+GPU architecture that integrates both compute units (CUs) and memory into a unified system. This integration aims to reduce data movement, enhance memory locality, and improve energy efficiency by enabling the CPU and GPU to share memory directly. This effort focused on the interplay of key design components—cache line size, the number of CUs, and main memory technology—and the trade-offs of each configuration were analyzed. This paper highlights the various configurations’ impact on memory accesses, data reuse, and power utilization. The results provide valuable insights that can be leveraged to optimize APU architectures for high-performance and energy-efficient computing and thus create a balanced architecture. This optimization can be achieved by adopting dynamic cache management, runtime CU scaling, and advanced memory integration, highlighting the potential of APUs to address critical challenges in compute, data movement, and memory power consumption.
LANL PVC bag-out-bag replacement [Slides]
The US material at 8 months aging, at room temperature and exposed to gamma at 80 degrees C has been completed. The FTIR data does not show any discernible difference between the US APU-ether material and the British material. There are some differences and the curves of the FTIR, but the behavior shows that the US material is behaving comparable to the British material. The British Blue sample exposed to gamma at 70 degrees C did have an observable change in the contact angle. This was the only notable change between the two different types of polymer the TGA on both the US and the British had no observable difference. Mechanical testing samples (dog-bones) are being prepared for tensile, shear, and puncture testing. This data will be provided in later meetings. Mass spec testing is needed to measure the headspace gases that are being seen during the aging study to better health characterize the radiolytic degradation of the APU ether samples, this is also ongoing.
Optimal management of electric hotel loads in mild hybrid heavy duty truck
Here, the problem of engine idling for heavy-duty trucks has been under study for decades with Auxiliary Power Units (APUs) and Truck Stop Electrification (TSE) as the most compelling solutions. With the electrification of trucks approaching feasibility in terms of cost-effective technology, hybridization offers another “degree of freedom” to tackle the problem. Here, this work aims at exploiting a battery pack of a 48 V mild-hybrid heavy-duty truck to store sufficient onboard energy for powering the auxiliary loads during the hoteling. This problem is not trivial, as the battery packs typically cannot recover the entire energy required through regeneration alone; hence an optimal energy management strategy needs to be employed to charge the battery through the engine during drive operation. This strategy optimizes powertrain performances among the four modes: (i) Engine of Coasting (EOC), (ii) Regeneration by braking, (iii) Regeneration by engine, and (iv) engine idling. This paper presents the development of a Dynamic Programming (DP) framework that employs a multi-objective cost function to minimize the fuel consumption and maximize the regeneration using the above-mentioned four modes. A typical heavy-duty truck drive cycle is used to represent the drive phase, with mandatory hoteling stops as per regulations. A comprehensive powertrain model is developed using validated components’ model. The DP employs two state variables: battery State-of-Charge (SOC) and engine mode, and three control inputs: (i) the engine ON–OFF state, (ii) clutch engagement state, and (iii) power request at the Electric Machine (EM) for calculating optimal SOC trajectory. The framework also tackles rapid engine ON–OFF scenarios to avoid the challenges associated with DP and the compromises in fuel cost with those approaches. Finally, the effectiveness of the proposed framework is tested for potential fuel savings on two different battery packs by performing the full cycle simulations. The results show 6.47% of fuel consumption reduction as compared to traditional APU-based heavy-duty truck.
MFC 5.0: An exascale many-physics flow solver
Many problems of interest in engineering, medicine, and the fundamental sciences rely on high-fidelity flow simulation, making performant computational fluid dynamics solvers a mainstay of the open-source software community. Previous work MFC 3.0 was made a published, documented, and open-source solver via Bryngelson et al. Comp. Phys. Comm. (2021) with numerous physical features, numerical methods, and scalable infrastructure. MFC 5.0 is a significant update to MFC 3.0, featuring a broad set of well-established and novel physical models and numerical methods, as well as the introduction of GPU and APU (or superchip) acceleration. Here, we exhibit state-of-the-art performance and ideal scaling on the first two exascale supercomputers, OLCF Frontier and LLNL El Capitan. Combined with MFC’s single-accelerator performance, MFC achieves exascale computation in practice, and achieved the largest-to-date public CFD simulation at 200 trillion grid points as a 2025 ACM Gordon Bell Prize finalist. New physical features include the immersed boundary method, N-fluid phase change, Euler–Euler and Euler–Lagrange sub-grid bubble models, fluid-structure interaction, hypo- and hyper-elastic materials, chemically reacting flow, two-material surface tension, magnetohydrodynamics (MHD), and more. Numerical techniques now represent the current state-of-the-art, including general relaxation characteristic boundary conditions, WENO variants, Strang splitting for stiff sub-grid flow features, and low Mach number treatments. Weak scaling to tens of thousands of GPUs on OLCF Summit and Frontier and LLNL El Capitan achieves efficiencies within 5% of ideal to over 90% of their respective system sizes. Strong scaling results for a 16-times increase in device count show parallel efficiencies over 90% on OLCF Frontier. MFC’s software stack has undergone further improvements, including continuous integration, which ensures code resilience and correctness through over 300 regression tests; metaprogramming, which reduces code length while maintaining performance portability; and code generation for computing chemical reactions
Performance of stainless steel interconnects with (Mn,Co) 3 O 4 -Based coating for solid oxide electrolysis
Mixed transition-metal oxide coatings are commonly applied to stainless steel interconnects for solid oxide cell stacks. Such coatings reduce oxidation and Cr evaporation rates, leading to improved degradation rate and stack lifetime. Here, the ChromLok™ MCO-based composition (Mn,Co) 3 O 4 is applied to Crofer 22 APU stainless steel and evaluated specifically for application in solid oxide electrolyzer stacks operating around 800 °C and utilizing oxygen-ion-conducting solid oxide cells. The MCO coating is found to decrease the stainless steel oxidation rate by about one order of magnitude, and decrease the Cr evaporation rate by fourfold. Furthermore, the coating also dramatically lowers the rate of area-specific resistance increase for stainless steel coupons oxidized for 500 h with constant current applied, from 33 mΩ*cm 2 kh -1 for an uncoated coupon to less than 4 mΩ*cm2 kh -1 for coated coupons. The coating is demonstrated on full-scale interconnects for single-cells, where the coating dramatically reduces degradation rate, and for a stack, which displays stable operation for 700 h.
Dynamic oxidation of (Mn,Co) 3 O 4 -Coated interconnects for solid oxide electrolysis cells
Solid oxide electrolysis cell stacks are expected to experience dynamic conditions when using renewable electricity derived from wind or solar power. To address this scenario, (Mn,Co) 3 O 4 (MCO)-coated Crofer 22 APU interconnect coupons are subjected to thermal cycling, and intermittent current density, with gas compositions relevant to high-temperature electrolysis (elevated steam:hydrogen ratio and oxygen content). Defects are also intentionally introduced in the MCO coating, to assess whether the difference in oxidation properties of the adjacent coated and uncoated (defective) surfaces causes sufficient stress to damage the protective oxide scale. Specimens with defects are subjected to oxidation for 1000 h at 800 °C, or thermal cycling. Before thermal cycling, some specimens are pre-oxidized to create a thick oxide scale to mimic the scale thickness expected after ~30 kh operation. In all cases, the coating and oxide scale remain well adhered with no cracking observed. Area-specific resistance (ASR) is monitored in both single-atmosphere and dual-atmosphere conditions, and the ASR is stable and is not impacted by dynamic cycling of the current density. Finally, this work provides confidence that the MCO coated interconnect will function as needed in dynamic operation conditions, even with coating defects.
Simulating many-engine spacecraft: Exceeding 1 quadrillion degrees of freedom via information geometric regularization
We present an optimized implementation of the recently proposed information geometric regularization (IGR) for unprecedented scale simulation of compressible fluid flows applied to multi-engine spacecraft boosters. We improve upon state-of-the-art computational fluid dynamics (CFD) techniques in terms of computational cost, memory footprint, and energy-to-solution metrics. Unified memory on coupled CPU–GPU or APU platforms increases problem size with negligible overhead. Mixed half/single-precision storage and computation are used on well-conditioned numerics. We simulate flow at 200 trillion grid points and 1 quadrillion degrees of freedom, exceeding the current record by a factor of 20. A factor of 4 wall-time speedup is achieved over optimized baselines. Ideal weak scaling is observed on OLCF Frontier, LLNL El Capitan, and CSCS Alps using the full systems. Strong scaling is near ideal at extreme conditions, including 80% efficiency on CSCS Alps with an 8 node baseline and stretching to the full system.
Metal-Supported Ceria Electrolyte-based SOFC Stack for Scalable, Low-Cost, High-Efficiency and Robust Stationary Power Systems (Final Report)
This is the end of project report for the DOE Project DE-FE-0027844 “Metal-Supported Ceria Electrolyte-based SOFC Stack for Scalable, Low-Cost, High-Efficiency and Robust Stationary Power Systems”. The objectives of the project were to develop several technologies required for the mass adoption of Solid Oxide Fuel Cell systems for medium to large power generation applications. The key achievements of this project were: 1) Development of fuel cells with complete internal fuel reforming capability. 2) Fuel cell power scale up with production of cells with larger active areas. 3) Assembly of an integrated 5 kW modular stack platform capable of installation into medium to large power systems. 4) Demonstration of a 10 kW Fuel Cell Power System utilizing the integrated 5 kW modular stack platform. 5) Cost modelling to show system cost of $1,500/kW are achievable at production volumes. 6) Development of cathode air poison getters for reduced degradation and extended fuel cell life. 7) Assessment of fuel poison on anode degradation, particularly poisons frequently produced from the process of gasification of coal. The Cummins led Voice of Customer (VOC) work identified several significant markets for natural gas powered fuel cell systems, where the end customer is interested in reducing energy consumption and emissions, or increasing reliability of electrical supply. Data centers and truck Auxiliary Power Units (APU) were highlighted as 2 key future markets for fuel cell systems.
Milestone Report: Co-Scheduling Applications on Current and Future Architectures
This report evaluates the viability and performance of co-scheduling scientific applications and workloads between CPUs and GPUs on current and future hardware architectures. This study was performed by collecting hardware independent and memory performance characteristics of 6 ECP mini-applications through Byfl and NSight Systems, by measuring the applications’ co-scheduled performance on current node architectures, and by gathering memory bandwidth utilization data with Intel Advisor and nvprof to begin reasoning about the feasibility of co-scheduling applications on future APU architectures.
Fine-Grained Power and Energy Attribution on AMD GPU/APU-Based Exascale Nodes
Modern exascale GPU- and APU-based systems provide multiple power and energy sensors, but differences in scope, update rate, timing, and filtering complicate the attribution of short-lived accelerator activity. This paper presents a methodology to characterize and correct these effects on Cray EX systems with AMD Instinct MI250X GPUs (Frontier) and MI300A APUs (Portage). Using controlled square-wave workloads, we quantify update intervals, delay, aliasing, and variability across up to 512 GPUs and 480 APUs with on-chip (rocm-smi/amd-smi) and off-chip Cray Power Management sensors. We reconstruct power from cumulative energy counters to achieve faster response times, validate it against on-chip, off-chip, and node-level sensors, and integrate the resulting streams into a Score-P/PAPI-based tool for time-aligned, phase-level attribution. Applied to rocHPL, rocHPL-MxP, and HPG-MxP, the method separates energy savings due to reduced runtime from changes in power. Mixed precision reduces node energy on Frontier by 79% for rocHPL-MxP and 31% for HPG-MxP, with similar trends on Portage. These results provide portable guidance for sensor validation and power-aware optimization on current and future exascale systems.