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 991 records · Page 55

Processor-In-Memory (PIM) Based Architectures for PetaFlops Potential Massively Parallel Processing

The report summarizes the work performed at the University of Notre Dame under a NASA grant from July 15, 1995 through July 14, 1996. Researchers involved in the work included the PI, Dr. Peter M. Kogge, and three graduate students under his direction in the Computer Science and Engineering Department: Stephen Dartt, Costin Iancu, and Lakshmi Narayanaswany. The organization of this report is as follows. Section 2 is a summary of the problem addressed by this work. Section 3 is a summary of the project's objectives and approach. Section 4 summarizes PIM technology briefly. Section 5 overviews the main results of the work. Section 6 then discusses the importance of the results and future directions. Also attached to this report are copies of several technical reports and publications whose contents directly reflect results developed during this study.

Kogge, Peter M.↗

An object-oriented approach for parallel self adaptive mesh refinement on block structured grids

Self-adaptive mesh refinement dynamically matches the computational demands of a solver for partial differential equations to the activity in the application's domain. In this paper we present two C++ class libraries, P++ and AMR++, which significantly simplify the development of sophisticated adaptive mesh refinement codes on (massively) parallel distributed memory architectures. The development is based on our previous research in this area. The C++ class libraries provide abstractions to separate the issues of developing parallel adaptive mesh refinement applications into those of parallelism, abstracted by P++, and adaptive mesh refinement, abstracted by AMR++. P++ is a parallel array class library to permit efficient development of architecture independent codes for structured grid applications, and AMR++ provides support for self-adaptive mesh refinement on block-structured grids of rectangular non-overlapping blocks. Using these libraries, the application programmers' work is greatly simplified to primarily specifying the serial single grid application and obtaining the parallel and self-adaptive mesh refinement code with minimal effort. Initial results for simple singular perturbation problems solved by self-adaptive multilevel techniques (FAC, AFAC), being implemented on the basis of prototypes of the P++/AMR++ environment, are presented. Singular perturbation problems frequently arise in large applications, e.g. in the area of computational fluid dynamics. They usually have solutions with layers which require adaptive mesh refinement and fast basic solvers in order to be resolved efficiently.

Lemke, Max↗

Global Arrays Parallel Programming Toolkit

The two predominant classes of programming models for parallel computing are distributed memory and shared memory. Both shared memory and distributed memory models have advantages and shortcomings. Shared memory model is much easier to use but it ignores data locality/placement. Given the hierarchical nature of the memory subsystems in modern computers this characteristic can have a negative impact on performance and scalability. Careful code restructuring to increase data reuse and replacing fine grain load/stores with block access to shared data can address the problem and yield performance for shared memory that is competitive with message-passing. However, this performance comes at the cost of compromising the ease of use that the shared memory model advertises. Distributed memory models, such as message-passing or one-sided communication, offer performance and scalability but they are difficult to program. The Global Arrays toolkit attempts to offer the best features of both models. It implements a shared-memory programming model in which data locality is managed by the programmer. This management is achieved by calls to functions that transfer data between a global address space (a distributed array) and local storage. In this respect, the GA model has similarities to the distributed shared-memory models that provide an explicit acquire/release protocol. However, the GA model acknowledges that remote data is slower to access than local data and allows data locality to be specified by the programmer and hence managed. GA is related to the global address space languages such as UPC, Titanium, and, to a lesser extent, Co-Array Fortran. In addition, by providing a set of data-parallel operations, GA is also related to data-parallel languages such as HPF, ZPL, and Data Parallel C. However, the Global Array programming model is implemented as a library that works with most languages used for technical computing and does not rely on compiler technology for achieving parallel efficiency. It also supports a combination of task- and data-parallelism and is available as an extension of the message passing (MPI) model. The GA model exposes to the programmer the hierarchical memory of modern high-performance computer systems, and by recognizing the communication overhead for remote data transfer, it promotes data reuse and locality of reference. Virtually all the scalable architectures possess non-uniform memory access characteristics that reflect their multi-level memory hierarchies. These hierarchies typically comprise processor registers, multiple levels of cache, local memory, and remote memory. Over time, both the number of levels and the cost (in processor cycles) of accessing deeper levels has been increasing. It is important for any scalable programming model to address memory hierarchy since it is critical to the efficient execution of scalable applications.

ARMCI↗

Fast and Scalable Implementations of Influence Maximization Algorithms

The Influence Maximization problem has been extensively studied in the past decade because of its practical applications in finding the key influencers in social networks. Due to the hardness of the underlying problem, existing algorithms have tried to trade off practical efficiency with approximation guarantees. However, even such approximate solutions take several hours of compute time on modest sized real world inputs. On the other hand, there is a lack of effective parallel and distributed algorithms to solve this problem. In this paper, we present efficient parallel algorithms for multithreaded and distributed systems to solve the influence maximization with approximation guarantee. Our algorithms extend state-of-the-art sequential approach based on computing reverse reachability sets. We present a detailed experimental evaluation, and analyze their performance and their sensitivity to input parameters, using real world inputs. Our experimental results demonstrate significant speedup on parallel architectures. We further show a speedup of up to 586X relative to the state-of-the-art sequential baseline using 1024 nodes of a supercomputer at far greater accuracy and twice the seed set size. To the best of our knowledge, this is the first effort in parallelizing the influence maximization operation at scale.

Minutoli, Marco↗

Accelerating Binarized Neural Networks via Bit-Tensor-Cores in Turing GPUs

Despite foreseeing tremendous speedups over conventional deep neural networks, the performance advantage of binarized neural networks (BNNs) has merely been showcased on general-purpose processors such as CPUs and GPUs. In fact, due to being unable to leverage bit-level-parallelism with a word-based architecture, GPUs have been criticized for extremely low utilization (1%) when executing BNNs. Consequently, the latest tensorcores in NVIDIA Turing GPUs start to experimentally support bit computation. In this work, we look into this brand new bit computation capability and characterize its unique features. We show that the stride of memory access can significantly affect performance delivery and a data-format co-design is highly desired to support the tensorcores for achieving superior performance than existing software solutions without tensorcores. We realize the tensorcore-accelerated BNN design, particularly the major functions for fully-connect and convolution layers — bit matrix multiplication and bit convolution. Evaluations on two NVIDIA Turing GPUs show that, with ResNet-18, our BTC-BNN design can process ImageNet at a rate of 5.6K images per second, 77% faster than state-of-the-art. Our BNN approach is released on https://github.com/pnnl/TCBNN.

Li, Ang↗

From Physics Model to Results: An Optimizing Framework for Cross-Architecture Code Generation

Starting from a high-level problem description in terms of partial differential equations using abstract tensor notation, the Chemora framework discretizes, optimizes, and generates complete high performance codes for a wide range of compute architectures. Chemora extends the capabilities of Cactus, facilitating the usage of large-scale CPU/GPU systems in an efficient manner for complex applications, without low-level code tuning. Chemora achieves parallelism through MPI and multi-threading, combining OpenMP and CUDA. Optimizations include high-level code transformations, efficient loop traversal strategies, dynamically selected data and instruction cache usage strategies, and JIT compilation of GPU code tailored to the problem characteristics. The discretization is based on higher-order finite differences on multi-block domains. Chemora's capabilities are demonstrated by simulations of black hole collisions. This problem provides an acid test of the framework, as the Einstein equations contain hundreds of variables and thousands of terms.

Blazewicz, Marek↗

The ijk forms of factorization methods. II - Parallel systems

The paper considers the 'ijk forms' of LU and Cholesky factorization on certain parallel computers. This extends an earlier analysis for vector computers. Attention is restricted to local memory systems with processors that may or may not have vector capability. Special attention is given to bus architectures but qualitative analyses are given for other interconnection systems.

Ortega, J. M.↗

A novel VLSI processor architecture for supercomputing arrays

Design of the processor element for general purpose massively parallel supercomputing arrays is highly complex and cost ineffective. To overcome this, the architecture and organization of the functional units of the processor element should be such as to suit the diverse computational structures and simplify mapping of complex communication structures of different classes of algorithms. This demands that the computation and communication structures of different class of algorithms be unified. While unifying the different communication structures is a difficult process, analysis of a wide class of algorithms reveals that their computation structures can be expressed in terms of basic IP,IP,OP,CM,R,SM, and MAA operations. The execution of these operations is unified on the PAcube macro-cell array. Based on this PAcube macro-cell array, we present a novel processor element called the GIPOP processor, which has dedicated functional units to perform the above operations. The architecture and organization of these functional units are such to satisfy the two important criteria mentioned above. The structure of the macro-cell and the unification process has led to a very regular and simpler design of the GIPOP processor. The production cost of the GIPOP processor is drastically reduced as it is designed on high performance mask programmable PAcube arrays.

Venkateswaran, N.↗

Session on High Speed Civil Transport Design Capability Using MDO and High Performance Computing

Since the inception of CAS in 1992, NASA Langley has been conducting research into applying multidisciplinary optimization (MDO) and high performance computing toward reducing aircraft design cycle time. The focus of this research has been the development of a series of computational frameworks and associated applications that increased in capability, complexity, and performance over time. The culmination of this effort is an automated high-fidelity analysis capability for a high speed civil transport (HSCT) vehicle installed on a network of heterogeneous computers with a computational framework built using Common Object Request Broker Architecture (CORBA) and Java. The main focus of the research in the early years was the development of the Framework for Interdisciplinary Design Optimization (FIDO) and associated HSCT applications. While the FIDO effort was eventually halted, work continued on HSCT applications of ever increasing complexity. The current application, HSCT4.0, employs high fidelity CFD and FEM analysis codes. For each analysis cycle, the vehicle geometry and computational grids are updated using new values for design variables. Processes for aeroelastic trim, loads convergence, displacement transfer, stress and buckling, and performance have been developed. In all, a total of 70 processes are integrated in the analysis framework. Many of the key processes include automatic differentiation capabilities to provide sensitivity information that can be used in optimization. A software engineering process was developed to manage this large project. Defining the interactions among 70 processes turned out to be an enormous, but essential, task. A formal requirements document was prepared that defined data flow among processes and subprocesses. A design document was then developed that translated the requirements into actual software design. A validation program was defined and implemented to ensure that codes integrated into the framework produced the same results as their standalone counterparts. Finally, a Commercial Off the Shelf (COTS) configuration management system was used to organize the software development. A computational environment, CJOPT, based on the Common Object Request Broker Architecture, CORBA, and the Java programming language has been developed as a framework for multidisciplinary analysis and Optimization. The environment exploits the parallelisms inherent in the application and distributes the constituent disciplines on machines best suited to their needs. In CJOpt, a discipline code is "wrapped" as an object. An interface to the object identifies the functionality (services) provided by the discipline, defined in Interface Definition Language (IDL) and implemented using Java. The results of using the HSCT4.0 capability are described. A summary of lessons learned is also presented. The use of some of the processes, codes, and techniques by industry are highlighted. The application of the methodology developed in this research to other aircraft are described. Finally, we show how the experience gained is being applied to entirely new vehicles, such as the Reusable Space Transportation System. Additional information is contained in the original.

Rehder, Joe↗

Towards Exascale Supernova Simulations with GenASiS

Core-collapse supernovae are worth simulating on leadership computing resources because they are the origin of many elements, are interesting targets for expensive observational programs, and provide valuable computational experience. The need to perform implicit neutrino transport addressing up to six dimensions of phase space makes simulations aimed at the explosion mechanism an exascale problem. The multiscale nature of the problem in space and time can be addressed with adaptive mesh refinement and implicit evolution. Several types of solvers - explicit and implicit, hyperbolic and possibly elliptic - must be deployed with operator splitting to address the multiphysics nature of the problem. While we have some ideas about how parallelism can be exploited, we look to applied mathematicians and computer scientists to work with us and provide libraries that deploy capabilities needed by our solvers on heterogeneous and communication-unfriendly computer architectures.

Cardall, Christian Y↗

The organic redox transistor for neuromorphic computing

Inspired by the in-memory computing architectures of biological systems, neuromorphic computing using crossbar arrays of artificial synapses based on non-volatile memory (NVM) devices with variable conductances has emerged as a new paradigm to enable massively parallel and ultra-low power computing hardware for data centric applications. Although inference has been demonstrated successfully using crossbars based on a variety of NMV technologies, efficient learning and scaling to large arrays (>10 6 elements) remains a challenge due to the synaptic elements' non-ideal electrical characteristics which degrades ANN accuracy. A further challenge is that in the conductive state memristors draw large currents >μA resulting in significant voltage drops in the interconnect wires and increased probability of failure in scaled arrays. We suggest the organic polymer redox transistor (RT) is an alternate approach that could solve many of these challenges, enabling both inference and parallel outer product updates, as recently demonstrated by Fuller et al. An RT consists of redox-active channel and gate electrodes in contact with a liquid or solid electrolyte. lon insertion through the electrolyte controls the channel electronic conductivity, while electron transfer through an external circuit maintains overall charge neutrality. Unlike a rechargeable battery, in the RT the voltage built-up across the electrolyte is kept to a minimum (typically <100 mV) by using the same material for the gate and channel. Elimination of the voltage offset simplifies integration of the RT into programmable arrays by enabling the use of various selectors. RTs based on inorganic and organic materials have been recently demonstrated with conductance tuning occurring at potentials of just a few mV and hundreds to thousands of linearly and symmetrically programmable conductance states, enabling near ideal accuracy in neural network simulations. Introduced in the 1980's, redox transistors with metallic gate electrodes and organic channel materials, also known as organic electrochemical transistors (OECTs), have been explored for a variety of applications such as chem- and bio-sensing, neural interfaces, and low cost printed circuits. A typical channel material for OECTs is the conducting polymer poly(3,4-ethylenedioxythiophene) doped with poly(styrene sulfonate) (PEDOT:PSS). PEDOT is a p-type semiconducting polymer with mobile positively charged polarons that hop chain-to-chain.

Talin, Albert Alec↗

Initial Characterization of Parallel NFS Implementations

Parallel NFS (pNFS) is touted as an emergent standard protocol for parallel I/O access in various storage environments. Several pNFS prototypes have been implemented for initial validation and protocol examination. Previous efforts have focused on realizing the pNFS protocol to expose the best bandwidth potential from underlying file and storage systems. In this presentation, we provide an initial characterization of two pNFS prototype implementations, lpNFS (a Lustre-based parallel NFS implementation) and spNFS (another reference implementation from Network Appliance, Inc.). We show that both lpNFS and spNFS can faithfully achieve the primary goal of pNFS, i.e., aggregating I/O bandwidth from many storage servers. However, they both face the challenge of scalable metadata management. Particularly, the throughput of sp-NFS metadata operations degrades significanlty with an increasing number of data servers. Even for the better-performing lpNFS, we discuss its architecture and propose a direct I/O request flow protocol to improve its performance.

97 MATHEMATICS AND COMPUTING↗

An efficient three-dimensional Poisson solver for SIMD high-performance-computing architectures

We present an algorithm that solves the three-dimensional Poisson equation on a cylindrical grid. The technique uses a finite-difference scheme with operator splitting. This splitting maps the banded structure of the operator matrix into a two-dimensional set of tridiagonal matrices, which are then solved in parallel. Our algorithm couples FFT techniques with the well-known ADI (Alternating Direction Implicit) method for solving Elliptic PDE's, and the implementation is extremely well suited for a massively parallel environment like the SIMD architecture of the MasPar MP-1. Due to the highly recursive nature of our problem, we believe that our method is highly efficient, as it avoids excessive interprocessor communication.

Cohl, H.↗

Fine-grained multithreading support for hybrid threaded MPI programming.

As high-end computing systems continue to grow in scale, recent advances in multi- and many-core architectures have pushed such growth toward more dense architectures, that is, more processing elements per physical node, rather than more physical nodes themselves. Although a large number of scientific applications have relied so far on an MPI-everywhere model for programming high-end parallel systems; this model may not be sufficient for future machines, given their physical constraints such as decreasing amounts of memory per processing element and shared caches. As a result, application and computer scientists are exploring alternative programming models that involve using MPI between address spaces and some other threaded model, such as OpenMP, Pthreads, or Intel TBB, within an address space. Such hybrid models require efficient support from an MPI implementation for MPI messages sent from multiple threads simultaneously. In this paper, we explore the issues involved in designing such an implementation. We present four approaches to building a fully thread-safe MPI implementation, with decreasing levels of critical-section granularity (from coarse-grain locks to fine-grain locks to lock-free operations) and correspondingly increasing levels of complexity. We present performance results that demonstrate the performance implications of the different approaches.

99 GENERAL AND MISCELLANEOUS↗

A parallel WLS state estimator on shared memory computers

This paper describes our experience with developing a parallel weighted- least-square (WLS) state estimation (SE) program for shared-memory parallel computers. Since the key computational kernel of the WLS algorithm based on the Newton-Raphson approach is a solver of sparse linear equations, a significant part of our effort was focused on selecting, implementing and evaluating this algorithm. An optimized shared memory version of the conjugate gradient (CG) algorithm was found to be competitive to state-of-the-art implementation of LU solvers for the SE problem on the SGI Altix, a shared memory architecture. We also ported the full SE algorithm including CG to the Cray MTA-2 shared memory multithreaded architecture and investigated its performance.

97 MATHEMATICS AND COMPUTING↗

Parallel computing for power system climate resiliency: Solving a large-scale stochastic capacity expansion problem with mpi-sppy

Here we propose a nodal stochastic generation and transmission expansion planning model that incorporates the output from high-resolution global climate models through load and generation availability scenarios. We implement our model in Pyomo and perform computational studies on a realistically-sized test case of the California electric grid in a high performance computing environment. We propose model reformulations and algorithm tuning to efficiently solve this large problem using a variant of the Progressive Hedging Algorithm. We utilize the parallelization capabilities and overall versatility of mpi-sppy, exploiting its hub-and-spoke architecture to concurrently obtain inner and outer bounds on an optimal expansion plan. Initial results show that instances with 360 representative days on a system with over 8,000 buses can be solved to within 5% of optimality in under 4 h of wall clock time, a first step towards solving a large-scale power system expansion planning problem across a wide range of climate-informed operational scenarios.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Unified Performance and Power Modeling of Scientific Workloads

It is expected that scientific applications executing on future large-scale HPC must be optimized not only in terms of performance, but also in terms of power consumption. As power and energy become increasingly constrained resources, researchers and developers must have access to tools that will allow for accurate prediction of both performance and power consumption. Reasoning about performance and power consumption in concert will be critical for achieving maximum utilization of limited resources on future HPC systems. To this end, we present a unified performance and power model for the Nek-Bone mini-application developed as part of the DOE's CESAR Exascale Co-Design Center. Our models consider the impact of computation, point-to-point communication, and collective communication

Modeling techniques↗

Data Acquisition System for Multi-Frequency Radar Flight Operations Preparation

A three-channel data acquisition system was developed for the NASA Multi-Frequency Radar (MFR) system. The system is based on a commercial-off-the-shelf (COTS) industrial PC (personal computer) and two dual-channel 14-bit digital receiver cards. The decimated complex envelope representations of the three radar signals are passed to the host PC via the PCI bus, and then processed in parallel by multiple cores of the PC CPU (central processing unit). The innovation is this parallelization of the radar data processing using multiple cores of a standard COTS multi-core CPU. The data processing portion of the data acquisition software was built using autonomous program modules or threads, which can run simultaneously on different cores. A master program module calculates the optimal number of processing threads, launches them, and continually supplies each with data. The benefit of this new parallel software architecture is that COTS PCs can be used to implement increasingly complex processing algorithms on an increasing number of radar range gates and data rates. As new PCs become available with higher numbers of CPU cores, the software will automatically utilize the additional computational capacity.

Leachman, Jonathan↗