Search NASA⌕ Search

SEARCH · Search NASA

Results for “GPFS”

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.

9 records
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Hvac: Removing I/O Bottleneck for Large-Scale Deep Learning Applications

Scientific communities are increasingly adopting deep learning (DL) models in their applications to accelerate scientific discovery processes. However, with rapid growth in the computing capabilities of HPC supercomputers, large-scale DL applications have to spend a significant portion of training time performing I/O to a parallel storage system. Previous research works have investigated optimization techniques such as prefetching and caching. Unfortunately, there exist non-trivial challenges to adopting the existing solutions on HPC supercomputers for large-scale DL training applications, which include non-performance and/or failures at extreme scale, lack of portability and generality in design, complex deployment methodology, and being limited to a specific application or dataset. To address these challenges, we propose High-Velocity AI Cache (HVAC), a distributed read-cache layer that targets and fully exploits the node-local storage or near node-local storage technology. HVAC seamlessly accelerates read I/O by aggregating node-local or near node-local storage, avoiding metadata lookups and file locking while preserving portability in the application code. We deploy and evaluate HVAC on 1,024 nodes (with over 6000 NVIDIA V100 GPUS) of the Summit supercomputer. In particular, we evaluate the scalability, efficiency, accuracy, and load distribution of HVAC compared to GPFS and XFS-on-NVMe. With four different DL applications, we observe an average 25 % performance improvement atop GPFS and 9% drop against XFS-on-NVMe, which scale linearly and are considered the performance upper bound. We envision HVAC as an important caching library for upcoming HPC supercomputers such as Frontier.

Khan, Awais↗
DOE OSTI2021

I/O Bottleneck Detection and Tuning: Connecting the Dots using Interactive Log Analysis

Using parallel file systems efficiently is a tricky problem due to inter-dependencies among multiple layers of I/O software, including high-level I/O libraries (HDF5, netCDF, etc.), MPI-IO, POSIX, and file systems (GPFS, Lustre, etc.). Profiling tools such as Darshan collect traces to help understand the I/O performance behavior. However, there are significant gaps in analyzing the collected traces and then applying tuning options offered by various layers of I/O software. Seeking to connect the dots between I/O bottleneck detection and tuning, we propose DXT Explorer, an interactive log analysis tool. In this paper, we present a case study using our interactive log analysis tool to identify and apply various I/O optimizations. We report an evaluation of performance improvement achieved for four I/O kernels extracted from science applications.

Bez, Jean Luca↗
DOE OSTI2023

Investigation of Lubricant Additive Interactions on Gasoline Particulate Filters

To understand how the composition of novel lubricant additives and their ash interact with gasoline particulate filters (GPFs), an accelerated aging protocol was conducted using three lubricant additive formulations and two GPF types. The additive packages (adpaks) consisted of Ca+Mg detergent in a 3:1 or 0:1 ratio and an anti-wear component—either zinc dialkyl dithiophosphate (ZDDP) or a novel phosphonium-phosphinate ionic liquid (IL) substitute. The particulate sampling captured amount/compositions of particulate matter (PM) generated, total particulate number, and size distribution. Five ash loadings were completed. GPF position and adpak composition affected the backpressure, ash composition, ash morphology, and captured mass. The particulate sampling indicated that the ash component consisted primarily of particles less than 50 nm in size and that the Mg-only adpak resulted in more particulate of 50–400 nm in size. Postmortem materials characterization indicated GPFs in the underfloor position had deeper penetration of ash into the walls compared to the close-coupled position. Additionally, the Mg-only adpak had a higher filter collection efficiency (>90%) and the ash particles consisted of a higher concentration of dense ash material. In contrast, four of the 3:1 Ca:Mg lubricant adpaks resulted in a collection efficiency of only 40–50%. Although the collection efficiency was higher with the Mg-only adpak, the ash layer in the GPF was not thicker, nor was the penetration into the wall more significant, and surprisingly the full useful life (FUL) backpressure was lower than with Ca:Mg adpaks. The higher density of the Mg-derived ash was the only detectable difference. A possible explanation of this observation is that Mg ash has a lower melting point and is more susceptible to densification during combustion or GPF regeneration. The substitution of IL in place of the ZDDP did not lead to any notable changes in collection efficiency or location of the ash.

42 ENGINEERING↗
DOE OSTI2024

Accelerating Application Bulk Synchronous Writes in HPC Environments

High-bandwidth storage tiers are becoming more common for their capability to absorb high-rate, bursty I/Os. Notably, the designs of these fast storage tiers differ from system to system. The variation of these layers and non-uniform methods of access can pose chal- lenges for applications seeking to run at multiple HPC facilities. Therefore, in this work, we present Spectral, a rapid-output ab- straction library to accelerate application, bulk-synchronous writes on HPC systems. We design Spectral to enable applications to use high-bandwidth storage, such as node-local storage and dis- tributed, write-caches (e.g., burst buffers) transparently without requiring modifications to the application or file system source code. The key idea is to allow applications to spend most of the time performing productive work and to not require any source code changes for maximum portability on different HPC archi- tectures. Spectral internally re-routes write-only files through available, high-performance I/O resources before ultimately mi- grating them to the shared global parallel file system. For instance, on Summit, Spectral transparently places application outputs on node-local storage and then utilizes asynchronous migration to the center-wide GPFS file system. We evaluate Spectral on the Summit HPC system (1024 nodes) using the IOR benchmark and real scientific applications. Spectral shows linear performance scaling, improving application write performance by over an order of magnitude when compared to GPFS.

Khan, Awais↗
DOE OSTI2022

Impacts of engine lubrication oil-derived ash on soot oxidative reactivity on a catalytic gasoline particulate filter

Characterizing soot oxidation kinetics is crucial for understanding how gasoline particulate filters (GPFs) perform both in terms of filtration efficiency and pressure drop. The most common method for measuring soot oxidative reactivity is thermogravimetric analysis (TGA). Because TGA is an offline method, it may inaccurately predict how soot oxidizes on a GPF, especially if a catalytic washcoat is present that may enhance oxidation rates through surface interactions as the soot loads on the filter. In this work, a novel in-situ soot oxidative reactivity measurement method was developed. The method involved loading a catalytic GPF under high temperature conditions conducive to soot oxidation and measuring the filtration efficiency of 100 nm particles. A correlation was made between filtration efficiency and loaded soot mass on the filter to allow calculation of the soot oxidation rate. The method was evaluated in experiments using a 2.0 L gasoline direct injection (GDI) engine with three oils of varying additive packages, including an oil with a high zinc dialkyldithiophosphate (ZDDP) concentration, a non-additive, pure poly-alpha olefin (PAO) oil, and an oil with a high concentration of calcium sulfonate. Calculated specific soot oxidation rates ranged from ~0.06–0.6 min –1 and showed strong dependence on GPF inlet gas temperature and lubrication oil additive type. The results clearly demonstrated a catalytic effect of calcium-containing ash particles. Reactivity of soot produced by the engine running with the high calcium containing lubrication oil was increased. Similarly, the results indicated that the catalytic washcoat on the filter increased soot oxidation rates on-filter, especially at low soot loading. Conversely, the ZDDP oil additive exhibited a reactivity inhibiting effect, resulting in lower soot oxidation rates. Finally, this work represents the first known in-situ soot oxidative reactivity measurements on a GPF and elucidates the effect of lubrication oil additives and catalytic washcoat on oxidation rate.

42 ENGINEERING↗
DOE OSTI2022

olcf/GLS

Clone of ls that includes file annotation based upon GPFS extended attributes

III, James Wynne↗
DOE OSTI2020

Enabling Lean and Stoichiometric Gasoline Direct Injection Engines through Mitigation of Nanoparticle Emissions

This project had the objective to efficiently reduce particulate mass (PM) and particulate number (PN) from lean and stoichiometric gasoline direction injection engines used in light duty vehicle applications. It also sought to use suspended particle instruments to measure the effective density, illustrating a pathway for new methods for accurately measuring soot mass at low concentration. The three-year effort took a systems level approach to evaluate fuel and lubricant impacts on GDI soot formation in lean and stoichiometric operation and to evaluate the impact of these factors on soot filtration aftertreatment in three-way catalyst-coated gasoline particulate filters (GPFs).

33 ADVANCED PROPULSION SYSTEMS↗
DOE OSTI2024

Collection of Disk Failure Events from Alpine, the Parallel File System for Summit Supercomputer

This dataset contains disk (HDD) failure events collected from the Alpine storage system of the Summit supercomputer, hosted at OLCF, spanning from January 4, 2019, to December 21, 2023 (a total of 4 years, 11 months, and 18 days), covering 89% of its operational lifetime. It includes 3,766 disk failure events, each recorded with its detection timestamp (in ISO 8601 format) and detailed by its location within the storage system - rack, enclosure, and drive slot number.

97 MATHEMATICS AND COMPUTING↗