Search NASA⌕ Search

SEARCH · Search NASA

Results for “parallel computer architecture”

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 325 records · Page 18

A Highly Parallel Implementation of K-Means for Multithreaded Architecture

We present a parallel implementation of the popular k-means clustering algorithm for massively multithreaded computer systems, as well as a parallelized version of the KKZ seed selection algorithm. We demonstrate that as system size increases, sequential seed selection can become a bottleneck. We also present an early attempt at parallelizing k-means that highlights critical performance issues when programming massively multithreaded systems. For our case studies, we used data collected from electric power simulations and run on the Cray XMT.

99 GENERAL AND MISCELLANEOUS↗

Parallel Evolutionary Optimization for Neuromorphic Network Training

One of the key impediments to the success of current neuromorphic computing architectures is the issue of how best to program them. Evolutionary optimization (EO) is one promising programming technique; in particular, its wide applicability makes it especially attractive for neuromorphic architectures, which can have many different characteristics. In this paper, we explore different facets of EO on a spiking neuromorphic computing model called DANNA. We focus on the performance of EO in the design of our DANNA simulator, and on how to structure EO on both multicore and massively parallel computing systems. We evaluate how our parallel methods impact the performance of EO on Titan, the U.S.'s largest open science supercomputer, and BOB, a Beowulf-style cluster of Raspberry Pi's. We also focus on how to improve the EO by evaluating commonality in higher performing neural networks, and present the result of a study that evaluates the EO performed by Titan.

97 MATHEMATICS AND COMPUTING↗

Parallel Evolutionary Optimization for Neuromorphic Network Training

One of the key impediments to the success of current neuromorphic computing architectures is the issue of how best to program them. Evolutionary optimization (EO) is one promising programming technique; in particular, its wide applicability makes it especially attractive for neuromorphic architectures, which can have many different characteristics. In this paper, we explore different facets of EO on a spiking neuromorphic computing model called DANNA. We focus on the performance of EO in the design of our DANNA simulator, and on how to structure EO on both multicore and massively parallel computing systems. We evaluate how our parallel methods impact the performance of EO on Titan, the U.S.'s largest open science supercomputer, and BOB, a Beowulf-style cluster of Raspberry Pi's. We also focus on how to improve the EO by evaluating commonality in higher performing neural networks, and present the result of a study that evaluates the EO performed by Titan.

Schuman, Catherine D↗

Massively parallel Wang Landau sampling on multiple GPUs

Wang Landau sampling is implemented on the Graphics Processing Unit (GPU) with the Compute Unified Device Architecture (CUDA). Performances on three different GPU cards, including the new generation Fermi architecture card, are compared with that on a Central Processing Unit (CPU). The parameters for massively parallel Wang Landau sampling are tuned in order to achieve fast convergence. For simulations of the water cluster systems, we obtain an average of over 50 times speedup for a given workload.

99 GENERAL AND MISCELLANEOUS↗

Architecture and method for a burst buffer using flash technology

A parallel supercomputing cluster includes compute nodes interconnected in a mesh of data links for executing an MPI job, and solid-state storage nodes each linked to a respective group of the compute nodes for receiving checkpoint data from the respective compute nodes, and magnetic disk storage linked to each of the solid-state storage nodes for asynchronous migration of the checkpoint data from the solid-state storage nodes to the magnetic disk storage. Each solid-state storage node presents a file system interface to the MPI job, and multiple MPI processes of the MPI job write the checkpoint data to a shared file in the solid-state storage in a strided fashion, and the solid-state storage node asynchronously migrates the checkpoint data from the shared file in the solid-state storage to the magnetic disk storage and writes the checkpoint data to the magnetic disk storage in a sequential fashion.

97 MATHEMATICS AND COMPUTING↗

High Energy Electromagnetic Particle Transportation on the GPU

We present massively parallel high energy electromagnetic particle transportation through a finely segmented detector on a Graphics Processing Unit (GPU). Simulating events of energetic particle decay in a general-purpose high energy physics (HEP) detector requires intensive computing resources, due to the complexity of the geometry as well as physics processes applied to particles copiously produced by primary collisions and secondary interactions. The recent advent of hardware architectures of many-core or accelerated processors provides the variety of concurrent programming models applicable not only for the high performance parallel computing, but also for the conventional computing intensive application such as the HEP detector simulation. The components of our prototype are a transportation process under a non-uniform magnetic field, geometry navigation with a set of solid shapes and materials, electromagnetic physics processes for electrons and photons, and an interface to a framework that dispatches bundles of tracks in a highly vectorized manner optimizing for spatial locality and throughput. Core algorithms and methods are excerpted from the Geant4 toolkit, and are modified and optimized for the GPU application. Program kernels written in C/C++ are designed to be compatible with CUDA and OpenCL and with the aim to be generic enough for easy porting to future programming models and hardware architectures. To improve throughput by overlapping data transfers with kernel execution, multiple CUDA streams are used. Issues with floating point accuracy, random numbers generation, data structure, kernel divergences and register spills are also considered. Performance evaluation for the relative speedup compared to the corresponding sequential execution on CPU is presented as well.

Canal, P.↗

Enabling machine learning-ready HPC ensembles with Merlin

With the growing complexity of computational and experimental facilities, many scientific researchers are turning to machine learning (ML) techniques to analyze large scale ensemble data. With complexities such as multi-component workflows, heterogeneous machine architectures, parallel file systems, and batch scheduling, care must be taken to facilitate this analysis in a high performance computing (HPC) environment. Here, we present Merlin, a workflow framework to enable large ML-friendly ensembles of scientific HPC simulations. By augmenting traditional HPC with distributed compute technologies, Merlin aims to lower the barrier for scientific subject matter experts to incorporate ML into their analysis. As a producer–consumer workflow model, Merlin enables multi-machine, cross-batch job, dynamically allocated yet persistent workflows capable of utilizing surge-compute resources. Key features of Merlin are a flexible HPC-centric interface, low per-task overhead, multi-tiered fault recovery, and a hierarchical sampling algorithm that allows for $\mathscr{O}$(N) task execution and $\mathscr{O}$(N ln N) task queuing to ensembles of millions of tasks. In addition to Merlin’s design, we test the algorithm’s performance in an HPC center and demonstrate the ability to enqueue 40 million simulations in 100 s, with a 30 millisecond per-task overhead that is independent of ensemble size. Finally, we describe some example applications that Merlin has enabled on leadership-class HPC resources, such as the ML-augmented optimization of nuclear fusion experiments and the calibration of infectious disease models to study the progression of and possible mitigation strategies for COVID-19.

97 MATHEMATICS AND COMPUTING↗

Galactos: computing the anisotropic 3-point correlation function for 2 billion galaxies

The nature of dark energy and the complete theory of gravity are two central questions currently facing cosmology. A vital tool for addressing them is the 3-point correlation function (3PCF), which probes deviations from a spatially random distribution of galaxies. However, the 3PCF's formidable computational expense has prevented its application to astronomical surveys comprising millions to billions of galaxies. We present Galactos, a high-performance implementation of a novel, O( N 2 ) algorithm that uses a load-balanced k-d tree and spherical harmonic expansions to compute the anisotropic 3PCF. Our implementation is optimized for the Intel Xeon Phi architecture, exploiting SIMD parallelism, instruction and thread concurrency, and significant L1 and L2 cache reuse, reaching 39% of peak performance on a single node. Galactos scales to the full Cori system, achieving 9.8 PF (peak) and 5.06 PF (sustained) across 9636 nodes, making the 3PCF easily computable for all galaxies in the observable universe.

79 ASTRONOMY AND ASTROPHYSICS↗

HYPPO: A Surrogate-Based Multi-Level Parallelism Tool for Hyperparameter Optimization

We present a new software, HYPPO, that enables the automatic tuning of hyperparameters of various deep learning (DL) models. Unlike other hyperparameter optimization (HPO) methods, HYPPO uses adaptive surrogate models and directly accounts for uncertainty in model predictions to find accurate and reliable models that make robust predictions. Using asynchronous nested parallelism, we are able to significantly alleviate the computational burden of training complex architectures and quantifying the uncertainty. HYPPO is implemented in Python and can be used with both TensorFlow and PyTorch libraries. We demonstrate various software features on time-series prediction and image classification problems as well as a scientific application in computed tomography image reconstruction. Finally, we show that (1) we can reduce by an order of magnitude the number of evaluations necessary to find the most optimal region in the hyperparameter space and (2) we can reduce by two orders of magnitude the throughput for such HPO process to complete.

adaptation models↗

Efficiency and Scalability of Barrier Synchronization on NoC Based Many-core Architectures

Interconnects based on Networks-on-Chip are an appealing solution to address future microprocessor designs where, very likely, hundreds of cores will be connected on a single chip. A fundamental role in highly parallelized applications running on many-core architectures will be played by barrier primitives used to synchronize the execution of parallel processes. This paper focuses on the analysis of the efficiency and scalability of different barrier implementations in many-core architectures based on NoCs. Several message passing barrier implementations based on four algorithms (all-to-all, master-slave, butterfly and tree) have been implemented and evaluated for a single-chip target architecture composed of a variable number of cores (from 4 to 128) and different network topologies (mesh, torus, ring, clustered-ring and fat-tree). Using a cycle-accurate simulator, we show the scalability of each barrier for every NoC topology, analyzing and comparing theoretical with real behaviors. We observed that some barrier algorithms, when implemented in hardware or software, show a different scaling behavior with respect to those theoretically expected. We evaluate the efficiency of each combination topology-barrier, demonstrating that, in many cases, simple network topologies can be more efficient than complex and highly connected topologies.

97 MATHEMATICS AND COMPUTING↗

Reconstruction for Liquid Argon TPC Neutrino Detectors Using Parallel Architectures [Slides]

Neutrinos are particles that interact rarely, so identifying them requires large detectors which produce lots of data. Processing this data with the computing power available is becoming more difficult as the detectors increase in size to reach their physics goals. In liquid argon time projection chambers (TPCs) the charged particles from neutrino interactions produce ionization electrons which drift in an electric field towards a series of collection wires, and the signal on the wires is used to reconstruct the interaction. The MicroBooNE detector currently collecting data at Fermilab has 8000 wires, and planned future experiments like DUNE will have 100 times more, which means that the time required to reconstruct an event will scale accordingly. Modernization of liquid argon TPC reconstruction code, including vectorization, parallelization and code portability to GPUs, will help to mitigate these challenges. The liquid argon TPC hit finding algorithm within the L arSoft framework used across multiple experiments has been vectorized and parallelized. This increases the speed of the algorithm on the order of ten times within a stand alone version on Intel architectures. This new version has been incorporated back into LarSoft so that it can be generally used. These methods will also be applied to other low level reconstruction of the wire signals such as the deconvolution. The applications and performance of this modernized liquid argon TPC wire reconstruction will be presented.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Multithreaded sparse matrix-matrix multiplication for many-core and GPU architectures

Sparse matrix-matrix multiplication is a key kernel that has applications in several domains such as scientific computing and graph analysis. Several algorithms have been studied in the past for this foundational kernel. In this work, we develop parallel algorithms for sparse matrix-matrix multiplication with a focus on performance portability across different high performance computing architectures. The performance of these algorithms depend on the data structures used in them. We compare different types of accumulators in these algorithms and demonstrate the performance difference between these data structures. Furthermore, we develop a meta-algorithm, kkSpGEMM, to choose the right algorithm and data structure based on the characteristics of the problem. We show performance comparisons on three architectures and demonstrate the need for the community to develop two phase sparse matrix-matrix multiplication implementations for efficient reuse of the data structures involved.

97 MATHEMATICS AND COMPUTING↗

Heterogeneous scalable framework for multiphase flows

Two categories of challenges confront the developer of computational spray models: those related to the computation and those related to the physics. Regarding the computation, the trend towards heterogeneous, multi- and many-core platforms will require considerable re-engineering of codes written for the current supercomputing platforms. Regarding the physics, accurate methods for transferring mass, momentum and energy from the dispersed phase onto the carrier fluid grid have so far eluded modelers. Significant challenges also lie at the intersection between these two categories. To be competitive, any physics model must be expressible in a parallel algorithm that performs well on evolving computer platforms. This work created an application based on a software architecture where the physics and software concerns are separated in a way that adds flexibility to both. The develop spray-tracking package includes an application programming interface (API) that abstracts away the platform-dependent parallelization concerns, enabling the scientific programmer to write serial code that the API resolves into parallel processes and threads of execution. The project also developed the infrastructure required to provide similar API s to other application. The API allow object-oriented Fortran applications direct interaction with Trilinos to support memory management of distributed objects in central processing units (CPU) and graphic processing units (GPU) nodes for applications using C++.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Ensemble Sampling vs. Time Sampling in Molecular Dynamics Simulations of Thermal Conductivity

In this report we compare time sampling and ensemble averaging as two different methods available for phase space sampling. For the comparison, we calculate thermal conductivities of solid argon and silicon structures, using equilibrium molecular dynamics. We introduce two different schemes for the ensemble averaging approach, and show that both can reduce the total simulation time as compared to time averaging. It is also found that velocity rescaling is an efficient mechanism for phase space exploration. Although our methodology is tested using classical molecular dynamics, the ensemble generation approaches may find their greatest utility in computationally expensive simulations such as first principles molecular dynamics. For such simulations, where each time step is costly, time sampling can require long simulation times because each time step must be evaluated sequentially and therefore phase space averaging is achieved through sequential operations. On the other hand, with ensemble averaging, phase space sampling can be achieved through parallel operations, since each ensemble is independent. For this reason, particularly when using massively parallel architectures, ensemble sampling can result in much shorter simulation times and exhibits similar overall computational effort.

74 ATOMIC AND MOLECULAR PHYSICS↗

Multigrid methods with space–time concurrency

Here, we consider the comparison of multigrid methods for parabolic partial differential equations that allow space–time concurrency. With current trends in computer architectures leading towards systems with more, but not faster, processors, space–time concurrency is crucial for speeding up time-integration simulations. In contrast, traditional time-integration techniques impose serious limitations on parallel performance due to the sequential nature of the time-stepping approach, allowing spatial concurrency only. This paper considers the three basic options of multigrid algorithms on space–time grids that allow parallelism in space and time: coarsening in space and time, semicoarsening in the spatial dimensions, and semicoarsening in the temporal dimension. We develop parallel software and performance models to study the three methods at scales of up to 16K cores and introduce an extension of one of them for handling multistep time integration. We then discuss advantages and disadvantages of the different approaches and their benefit compared to traditional space-parallel algorithms with sequential time stepping on modern architectures.

97 MATHEMATICS AND COMPUTING↗

IB: a Monte Carlo Simulation Tool for Neutron Scattering Instrument Design under Parallel Virtual Machine

IB is a Monte Carlo simulation tool for aiding neutron scattering instrument designs. It is written in C++ and implemented under Parallel Virtual Machine. The program has a few basic components, or modules, that can be used to build a virtual neutron scattering instrument. More complex components, such as neutron guides and multichannel beam benders, can be constructed using the grouping technique unique to IB. Users can specify a collection of modules as a group. For example, a neutron guide can be constructed by grouping four neutron mirrors together that make up the four sides of the guide. IB s simulation engine ensures that neutrons entering a group will be properly operated upon by all members of the group. For simulations that require higher computer speed, the program can be run in parallel mode under the PVM architecture. Initially, the program was written for designing instruments on pulsed neutron sources, it has since been used to simulate reactor based instruments as well.

Monte Carlo Simulations↗

Event recognition using signal spectrograms in long pulse experiments

As discharge duration increases, real-time complex analysis of the signal becomes more important. In this context, data acquisition and processing systems must provide models for designing experiments which use event oriented plasma control. One example of advanced data analysis is signal classification. The off-line statistical analysis of a large number of discharges provides information to develop algorithms for the determination of the plasma parameters from measurements of magnetohydrodinamic waves, for example, to detect density fluctuations induced by the Alfven cascades using morphological patterns. The need to apply different algorithms to the signals and to address different processing algorithms using the previous results necessitates the use of an event-based experiment. The Intelligent Test and Measurement System platform is an example of architecture designed to implement distributed data acquisition and real-time processing systems. The processing algorithm sequence is modeled using an event-based paradigm. The adaptive capacity of this model is based on the logic defined by the use of state machines in SCXML. The Intelligent Test and Measurement System platform mixes a local multiprocessing model with a distributed deployment of services based on Jini.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

PETSc/TAO developments for GPU-based early exascale systems

The Portable Extensible Toolkit for Scientific Computation (PETSc) library provides scalable solvers for nonlinear time-dependent differential and algebraic equations and for numerical optimization via the Toolkit for Advanced Optimization (TAO). PETSc is used in dozens of scientific fields and is an important building block for many simulation codes. During the U.S. Department of Energy’s Exascale Computing Project, the PETSc team has made substantial efforts to enable efficient utilization of the massive fine-grain parallelism present within exascale compute nodes and to enable performance portability across exascale architectures. We recap some of the challenges that designers of numerical libraries face in such an endeavor, and then discuss the many developments we have made, which include the addition of new GPU backends, features supporting efficient on-device matrix assembly, better support for asynchronicity and GPU kernel concurrency, and new communication infrastructure. In conclusion, we evaluate the performance of these developments on some pre-exascale systems as well as the early exascale systems Frontier and Aurora, using compute kernel, communication layer, solver, and mini-application benchmark studies, and then close with a few observations drawn from our experiences on the tension between portable performance and other goals of numerical libraries.

Exascale Computing Project (ECP)↗