Search NASA⌕ Search

SEARCH · Search NASA

Results for “parallel systems”

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 433 records · Page 24

Leveraging Pre-Built Catalogs and Object-Level Scheduling to Eliminate I/O Bottlenecks in HPC Environments

Modern High-Performance Computing (HPC) environments face mounting challenges due to the shift from large to small file datasets, along with an increasing number of users and parallelized applications. As HPC systems rely on Parallel File Systems (PFS), such as Lustre for data processing, performance bottlenecks stemming from Object Storage Target (OST) contention have become a significant concern. Existing solutions, such as LADS with its object-level scheduling approach, fall short in large-scale HPC environments due to their inability to effectively address metadata I/O bottlenecks and the growing number of I/O processes. This study highlights the pressing need for a comprehensive solution that tackles both OST contention and metadata I/O challenges in diverse HPC workloads. To address these challenges, we propose SwiftLoad, an object-level I/O scheduling framework that leverages a metadata catalog to enhance the performance and efficiency of parallel HPC utilities. The adoption of the metadata catalog mitigates the metadata I/O bottlenecks that commonly occur in HPC utilities, a challenge that is particularly pronounced in object-level I/O scheduling. SwiftLoad addresses OST contention and the uneven distribution of I/O processes across different OSTs through mathematical modeling and incorporates a Loader Configuration Module to regulate the number of I/O processes. Evaluated with two representative utilities—data deduplication profiling and data augmentation—SwiftLoad achieved performance improvements of up to 5.63x and 11.0x, respectively, on a production supercomputer.

HPC↗

High performance remote sensing data analysis using parallel computation

This paper examines the JPL/Caltech parallel processing system designed for rapid processing and transfer of large quantities of data from remote sensing instruments flown on NASA missions. Two remote sensing analysis applications that use this processing system are described: (1) an analysis system for retrieval of atmospheric parameters (such as species abundance, atmospheric temperature, and water vapor profiles) from data obtained by a Fourier transform IR spectrometer and (2) a prototype airborne SAR processing system. It is shown that a parallel processing system such as the JPL/Caltech system can offer supercomputer computational capability and high-volume data throughput and still be cost-effective.

Patterson, Jean E.↗

Enabling Parallel Execution of System-level Simulations in SAM

This report summarizes the recent code updates related to “element ghosting” in SAM to enable the parallel execution of system-level simulations using multiple processors/cores. Unlike typical MOOSE-based applications, for system-level simulations, SAM mostly deals with a collection of discrete small pieces of meshes, and the connection of physics on these meshes are realized by using “connector” types of components/code structures, such as conjugate heat transfer and flow junctions. The required code implementation is to correctly mark the necessary ghost elements for each type of such components/code structures; thus, the lower-level libraries can correctly perform the necessary data transfer between processors (CPUs) when executed in parallel mode. After the code updates, SAM can now run system-level simulations in the parallel mode. The parallel execution capability was then tested with an ABTR input model with 23k DOFs. Significant speedup was demonstrated when the optimal number of CPUs were used in parallel mode. Future systematic studies on parallelization performance using additional test cases covering different physics/scenarios will be needed to provide additional insights into the scalability of SAM parallelization.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Parallelization of Rocket Engine System Software (Press)

The main goal is to assess parallelization requirements for the Rocket Engine Numeric Simulator (RENS) project which, aside from gathering information on liquid-propelled rocket engines and setting forth requirements, involve a large FORTRAN based package at NASA Lewis Research Center and TDK software developed by SUBR/UWF. The ultimate aim is to develop, test, integrate, and suitably deploy a family of software packages on various aspects and facets of rocket engines using liquid-propellants. At present, all project efforts by the funding agency, NASA Lewis Research Center, and the HBCU participants are disseminated over the internet using world wide web home pages. Considering obviously expensive methods of actual field trails, the benefits of software simulators are potentially enormous. When realized, these benefits will be analogous to those provided by numerous CAD/CAM packages and flight-training simulators. According to the overall task assignments, Hampton University's role is to collect all available software, place them in a common format, assess and evaluate, define interfaces, and provide integration. Most importantly, the HU's mission is to see to it that the real-time performance is assured. This involves source code translations, porting, and distribution. The porting will be done in two phases: First, place all software on Cray XMP platform using FORTRAN. After testing and evaluation on the Cray X-MP, the code will be translated to C + + and ported to the parallel nCUBE platform. At present, we are evaluating another option of distributed processing over local area networks using Sun NFS, Ethernet, TCP/IP. Considering the heterogeneous nature of the present software (e.g., first started as an expert system using LISP machines) which now involve FORTRAN code, the effort is expected to be quite challenging.

Cezzar, Ruknet↗

Molecular solid-state inverter-converter system

A modular approach for aerospace electrical systems has been developed, using lightweight high efficiency pulse width modulation techniques. With the modular approach, a required system is obtained by paralleling modules. The modular system includes the inverters and converters, a paralleling system, and an automatic control and fault-sensing protection system with a visual annunciator. The output is 150 V dc, or a low distortion three phase sine wave at 120 V, 400 Hz. Input power is unregulated 56 V dc. Each module is rated 2.5 kW or 3.6 kVA at 0.7 power factor.

Birchenough, A. G.↗

Lessons Learned in Deploying the World s Largest Scale Lustre File System

The Spider system at the Oak Ridge National Laboratory's Leadership Computing Facility (OLCF) is the world's largest scale Lustre parallel file system. Envisioned as a shared parallel file system capable of delivering both the bandwidth and capacity requirements of the OLCF's diverse computational environment, the project had a number of ambitious goals. To support the workloads of the OLCF's diverse computational platforms, the aggregate performance and storage capacity of Spider exceed that of our previously deployed systems by a factor of 6x - 240 GB/sec, and 17x - 10 Petabytes, respectively. Furthermore, Spider supports over 26,000 clients concurrently accessing the file system, which exceeds our previously deployed systems by nearly 4x. In addition to these scalability challenges, moving to a center-wide shared file system required dramatically improved resiliency and fault-tolerance mechanisms. This paper details our efforts in designing, deploying, and operating Spider. Through a phased approach of research and development, prototyping, deployment, and transition to operations, this work has resulted in a number of insights into large-scale parallel file system architectures, from both the design and the operational perspectives. We present in this paper our solutions to issues such as network congestion, performance baselining and evaluation, file system journaling overheads, and high availability in a system with tens of thousands of components. We also discuss areas of continued challenges, such as stressed metadata performance and the need for file system quality of service alongside with our efforts to address them. Finally, operational aspects of managing a system of this scale are discussed along with real-world data and observations.

99 GENERAL AND MISCELLANEOUS↗

Tuning HDF5 for Lustre File Systems

HDF5 is a cross-platform parallel I/O library that is used by a wide variety of HPC applications for the flexibility of its hierarchical object-database representation of scientific data. We describe our recent work to optimize the performance of the HDF5 and MPI-IO libraries for the Lustre parallel file system. We selected three different HPC applications to represent the diverse range of I/O requirements, and measured their performance on three different systems to demonstrate the robustness of our optimizations across different file system configurations and to validate our optimization strategy. We demonstrate that the combined optimizations improve HDF5 parallel I/O performance by up to 33 times in some cases running close to the achievable peak performance of the underlying file system and demonstrate scalable performance up to 40,960-way concurrency.

97 MATHEMATICS AND COMPUTING↗

Collectively loading programs in a multiple program multiple data environment

Techniques are disclosed for loading programs efficiently in a parallel computing system. In one embodiment, nodes of the parallel computing system receive a load description file which indicates, for each program of a multiple program multiple data (MPMD) job, nodes which are to load the program. The nodes determine, using collective operations, a total number of programs to load and a number of programs to load in parallel. The nodes further generate a class route for each program to be loaded in parallel, where the class route generated for a particular program includes only those nodes on which the program needs to be loaded. For each class route, a node is selected using a collective operation to be a load leader which accesses a file system to load the program associated with a class route and broadcasts the program via the class route to other nodes which require the program.

97 MATHEMATICS AND COMPUTING↗

Collectively loading programs in a multiple program multiple data environment

Techniques are disclosed for loading programs efficiently in a parallel computing system. In one embodiment, nodes of the parallel computing system receive a load description file which indicates, for each program of a multiple program multiple data (MPMD) job, nodes which are to load the program. The nodes determine, using collective operations, a total number of programs to load and a number of programs to load in parallel. The nodes further generate a class route for each program to be loaded in parallel, where the class route generated for a particular program includes only those nodes on which the program needs to be loaded. For each class route, a node is selected using a collective operation to be a load leader which accesses a file system to load the program associated with a class route and broadcasts the program via the class route to other nodes which require the program.

97 MATHEMATICS AND COMPUTING↗

Multi-Node Program Fuzzing on High Performance Computing Resources

Significant effort is placed on tuning the internal parameters of fuzzers to explore the state space, measured as coverage, of binaries. In this work, we investigate the effects of the external environment on the resulting coverage after fuzzing two binaries with AFL for 24 hours. Parameters such as scaling to multiple nodes, node saturation, and parallel file system type on HPC resources are controlled in order to maximize coverage. It will be shown that employing a parallel file system such as IBM's General Parallel File System offers an advantage for fuzzing operations, since it contains enhancements for performance optimization. When combined with scaling to two and four nodes, while simultaneously restricting the number of coordinated AFL tasks per node on the low end (10-50% of available physical cores), coverage may be enhanced within a shorter period of time. Thus, controlling the external environment is a useful effort.

97 MATHEMATICS AND COMPUTING↗

Optimizing Input/Output Using Adaptive File System Policies

Parallel input/output characterization studies and experiments with flexible resource management algorithms indicate that adaptivity is crucial to file system performance. In this paper we propose an automatic technique for selecting and refining file system policies based on application access patterns and execution environment. An automatic classification framework allows the file system to select appropriate caching and pre-fetching policies, while performance sensors provide feedback used to tune policy parameters for specific system environments. To illustrate the potential performance improvements possible using adaptive file system policies, we present results from experiments involving classification-based and performance-based steering.

Madhyastha, Tara M.↗

System-wide power management control via clock distribution network

An apparatus, method and computer program product for automatically controlling power dissipation of a parallel computing system that includes a plurality of processors. A computing device issues a command to the parallel computing system. A clock pulse-width modulator encodes the command in a system clock signal to be distributed to the plurality of processors. The plurality of processors in the parallel computing system receive the system clock signal including the encoded command, and adjusts power dissipation according to the encoded command.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Towards Native Execution of Deep Learning on a Leadership-Class HPC System

Large parallel machines generally offer the best parallel performance with "native execution" that is achieved using codes developed with the optimized compilers, communication libraries, and runtimes offered on the machines. In this paper, we report and analyze performance results from native execution of deep learning on a leadership-class high-performance computing (HPC) system. Using our new code called DeepEx, we present a study of the parallel speed up and convergence rates of learning achieved with native parallel execution. In the trade-off between computational parallelism and synchronized convergence, we first focus on maximizing parallelism while still obtaining convergence. Scaling results are reported from execution on up to 15,000 GPUs using two scientific data sets from atom microscopy and protein folding applications, and also using the popular ImageNet data set. In terms of the traditional measure of parallel speed up, excellent scaling is observed up to 12,000 GPUs. Additionally, accounting for convergence rates of deep learning accuracy or error, a deep learning-specific metric called "learning speed up" is also tracked. The performance results indicate the need to evaluate parallel deep learning execution in terms of learning speed up, and point to additional directions for improved exploitation of high-end HPC systems.

Yoginath, Srikanth↗

Synchronous parallel kinetic Monte Carlo for continuum diffusion-reaction systems

A novel parallel kinetic Monte Carlo (kMC) algorithm formulated on the basis of perfect time synchronicity is presented. The algorithm is intended as a generalization of the standard n-fold kMC method, and is trivially implemented in parallel architectures. In its present form, the algorithm is not rigorous in the sense that boundary conflicts are ignored. We demonstrate, however, that, in their absence, or if they were correctly accounted for, our algorithm solves the same master equation as the serial method. We test the validity and parallel performance of the method by solving several pure diffusion problems (i.e. with no particle interactions) with known analytical solution. We also study diffusion-reaction systems with known asymptotic behavior and find that, for large systems with interaction radii smaller than the typical diffusion length, boundary conflicts are negligible and do not affect the global kinetic evolution, which is seen to agree with the expected analytical behavior. Our method is a controlled approximation in the sense that the error incurred by ignoring boundary conflicts can be quantified intrinsically, during the course of a simulation, and decreased arbitrarily (controlled) by modifying a few problem-dependent simulation parameters.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Cloud object store for checkpoints of high performance computing applications using decoupling middleware

Cloud object storage is enabled for checkpoints of high performance computing applications using a middleware process. A plurality of files, such as checkpoint files, generated by a plurality of processes in a parallel computing system are stored by obtaining said plurality of files from said parallel computing system; converting said plurality of files to objects using a log structured file system middleware process; and providing said objects for storage in a cloud object storage system. The plurality of processes may run, for example, on a plurality of compute nodes. The log structured file system middleware process may be embodied, for example, as a Parallel Log-Structured File System (PLFS). The log structured file system middleware process optionally executes on a burst buffer node.

97 MATHEMATICS AND COMPUTING↗

Network support for system initiated checkpoints

A system, method and computer program product for supporting system initiated checkpoints in parallel computing systems. The system and method generates selective control signals to perform checkpointing of system related data in presence of messaging activity associated with a user application running at the node. The checkpointing is initiated by the system such that checkpoint data of a plurality of network nodes may be obtained even in the presence of user applications running on highly parallel computers that include ongoing user messaging activity.

97 MATHEMATICS AND COMPUTING↗

Network support for system initiated checkpoints

A system, method and computer program product for supporting system initiated checkpoints in parallel computing systems. The system and method generates selective control signals to perform checkpointing of system related data in presence of messaging activity associated with a user application running at the node. The checkpointing is initiated by the system such that checkpoint data of a plurality of network nodes may be obtained even in the presence of user applications running on highly parallel computers that include ongoing user messaging activity.

Chen, Dong↗