Results for “parallel machines” — Page 25

NASA NTRS2006

Implementing Scientific Simulation Codes Highly Tailored for Vector Architectures Using Custom Configurable Computing Machines

The motivation for this work comes from an observation that amidst the push for Massively Parallel (MP) solutions to high-end computing problems such as numerical physical simulations, large amounts of legacy code exist that are highly optimized for vector supercomputers. Because re-hosting legacy code often requires a complete re-write of the original code, which can be a very long and expensive effort, this work examines the potential to exploit reconfigurable computing machines in place of a vector supercomputer to implement an essentially unmodified legacy source code. Custom and reconfigurable computing resources could be used to emulate an original application's target platform to the extent required to achieve high performance. To arrive at an architecture that delivers the desired performance subject to limited resources involves solving a multi-variable optimization problem with constraints. Prior research in the area of reconfigurable computing has demonstrated that designing an optimum hardware implementation of a given application under hardware resource constraints is an NP-complete problem. The premise of the approach is that the general issue of applying reconfigurable computing resources to the implementation of an application, maximizing the performance of the computation subject to physical resource constraints, can be made a tractable problem by assuming a computational paradigm, such as vector processing. This research contributes a formulation of the problem and a methodology to design a reconfigurable vector processing implementation of a given application that satisfies a performance metric. A generic, parametric, architectural framework for vector processing implemented in reconfigurable logic is developed as a target for a scheduling/mapping algorithm that maps an input computation to a given instance of the architecture. This algorithm is integrated with an optimization framework to arrive at a specification of the architecture parameters that attempts to minimize execution time, while staying within resource constraints. The flexibility of using a custom reconfigurable implementation is exploited in a unique manner to leverage the lessons learned in vector supercomputer development. The vector processing framework is tailored to the application, with variable parameters that are fixed in traditional vector processing. Benchmark data that demonstrates the functionality and utility of the approach is presented. The benchmark data includes an identified bottleneck in a real case study example vector code, the NASA Langley Terminal Area Simulation System (TASS) application.

Rutishauser, David↗

DOE OSTI2022

FutureTense

Protective vaccines and reliable diagnostics are essential tools for controlling viral diseases. However, the efficacy of these tools can be diminished by mutations in viral genomes. The delay between the emergence of new viral strains and the redesign of vaccines and diagnostics allows for continued viral transmission. Is it possible to address this challenge by computationally predicting viral genome sequence evolution? Can we “future-proof” vaccines and diagnostics by targeting both current and anticipated future sequence variants? While predicting viral evolution is still an unsolved, “grand challenge” problem in biology, the large, and rapidly growing, number of SARS-CoV-2 genome sequences provide an opportunity to quantify the ability of machine learning to predict viral genome sequence evolution. Towards this end, we have developed a simple computational model for predicting viral evolution at the level of individual nucleotides. The key metric for quantifying the per-base, prediction accuracy for viral evolution is the Mann-Whitney U statistic (or, equivalently, the area under the receiver operator curve). Since the Mann-Whitney U statistic is not a differentiable function, existing deep leaning packages (like Pytorch and Keras/TensorFlow) are not useful, as they require that the accuracy metric/objective function be analytically differentiable with respect to the model parameters. To overcome this challenge, we have implemented custom software, “FutureTense”, that can train a machine learning model by maximizing the non-differentiable Mann-Whitney U statistic. This software trains a machine learning model by exploring along the direction of the discrete gradient of the Mann-Whitney U statistic in the model parameter space. Parallel computing and genome sequence-specific optimizations are used to accelerate model training. The resulting machine learning model learns the observed high C->U mutation rates in the SARS-CoV-2 genome (which are potentially induced by host defenses) and provides prediction accuracies that are significantly better than one would expect from random chance. While predicting viral evolution is still quite far from a solved problem, the surprising performance of this simple model gives hope that the accuracy of predicting viral genome evolution can be further increased by more sophisticated approaches.

Gans, Jason↗

NASA NTRS1994

Characterizing parallel file-access patterns on a large-scale multiprocessor

Rapid increases in the computational speeds of multiprocessors have not been matched by corresponding performance enhancements in the I/O subsystem. To satisfy the large and growing I/O requirements of some parallel scientific applications, we need parallel file systems that can provide high-bandwidth and high-volume data transfer between the I/O subsystem and thousands of processors. Design of such high-performance parallel file systems depends on a thorough grasp of the expected workload. So far there have been no comprehensive usage studies of multiprocessor file systems. Our CHARISMA project intends to fill this void. The first results from our study involve an iPSC/860 at NASA Ames. This paper presents results from a different platform, the CM-5 at the National Center for Supercomputing Applications. The CHARISMA studies are unique because we collect information about every individual read and write request and about the entire mix of applications running on the machines. The results of our trace analysis lead to recommendations for parallel file system design. First the file system should support efficient concurrent access to many files, and I/O requests from many jobs under varying load conditions. Second, it must efficiently manage large files kept open for long periods. Third, it should expect to see small requests predominantly sequential access patterns, application-wide synchronous access, no concurrent file-sharing between jobs appreciable byte and block sharing between processes within jobs, and strong interprocess locality. Finally, the trace data suggest that node-level write caches and collective I/O request interfaces may be useful in certain environments.

Purakayastha, Apratim↗

DOE OSTI2024

Importance of Higher Fidelity Model Geometries during Optimization of Critical Experiments

PARADIGM, PARallel Approach of Differential and InteGral Measurements, is a cross-collaborative effort at Los Alamos National Laboratory between nuclear data theorists, differential and integral experimenters, as well as machine learning statisticians to tackle uncertainties in the intermediate region of 239 Pu. In essence, the idea behind PARADIGM is to remove the linear conceptualization of the nuclear data pipeline, shown in Figure 1, and replace it with a far more parallelized approach. The novel approach leverages machine learning to guide which differential measurements and integral experiments will result in the largest decrease in uncertain ties for a nuclide reaction pair in a given energy range. The concept builds off earlier work, EUCLID, which focused on the fast region of 239 Pu. The practical benefit of having evaluation, differential measurement, and integral experiment personnel in collaboration with machine learning is to represent the entire nuclear data in one snapshot. This enable large reduction in the time to deliver improved nuclear data, which using the PARADIGM approach could be done in 3 years. A general outline of PARADIGM and specific topics are available in other papers. The discussion here will pertain directly to the integral experiment design. More specifically, the process of taking a rough design and transforming it into a finalized neutronic model will be discussed.

97 MATHEMATICS AND COMPUTING↗

NASA NTRS1992

Automatic data partitioning on distributed memory multicomputers

Distributed-memory parallel computers are increasingly being used to provide high levels of performance for scientific applications. Unfortunately, such machines are not very easy to program. A number of research efforts seek to alleviate this problem by developing compilers that take over the task of generating communication. The communication overheads and the extent of parallelism exploited in the resulting target program are determined largely by the manner in which data is partitioned across different processors of the machine. Most of the compilers provide no assistance to the programmer in the crucial task of determining a good data partitioning scheme. A novel approach is presented, the constraints-based approach, to the problem of automatic data partitioning for numeric programs. In this approach, the compiler identifies some desirable requirements on the distribution of various arrays being referenced in each statement, based on performance considerations. These desirable requirements are referred to as constraints. For each constraint, the compiler determines a quality measure that captures its importance with respect to the performance of the program. The quality measure is obtained through static performance estimation, without actually generating the target data-parallel program with explicit communication. Each data distribution decision is taken by combining all the relevant constraints. The compiler attempts to resolve any conflicts between constraints such that the overall execution time of the parallel program is minimized. This approach has been implemented as part of a compiler called Paradigm, that accepts Fortran 77 programs, and specifies the partitioning scheme to be used for each array in the program. We have obtained results on some programs taken from the Linpack and Eispack libraries, and the Perfect Benchmarks. These results are quite promising, and demonstrate the feasibility of automatic data partitioning for a significant class of scientific application programs with regular computations.

Gupta, Manish↗

DOE OSTI2026

Fast and Accurate Intersections on a Sphere

We introduce a fast, high-precision algorithm for calculating intersections between great circle arcs and lines of constant latitude on the unit sphere. We first propose a simplified intersection point formula with improved speed and numerical robustness over the ones traditionally implemented in geoscience software. We then show how algorithms based on the concept of error-free transformations (EFT) can be applied to evaluate this formula within a relative error bound that is on the order of machine precision. Here, we demonstrate that, with a vectorized and parallelized implementation, this enhanced accuracy is achieved with no compute time overhead compared to a direct calculation in hardware floating point, making our algorithm suitable for performance-sensitive applications like regridding of high-resolution climate data. In contrast, evaluating our formula using high-precision data types like quadruple precision and arbitrary precision, or using the robust intersection computation routines from the Computational Geometry Algorithms Library, leads to significant computational overhead, especially since these alternatives inhibit vectorization. More generally, our work demonstrates how EFT techniques can be combined and extended to implement nontrivial geometric calculations with high accuracy and speed.

Environmental sciences↗

DOE OSTI2022

The Portals 4.3 Network Programming Interface

This report presents a specification for the Portals 4 network programming interface. Portals 4 is intended to allow scalable, high-performance network communication between nodes of a parallel computing system. Portals 4 is well suited to massively parallel processing and embedded systems. Portals 4 represents an adaption of the data movement layer developed for massively parallel processing platforms, such as the 4500-node Intel TeraFLOPS machine. Sandia's Cplant cluster project motivated the development of Version 3.0, which was later extended to Version 3.3 as part of the Cray Red Storm machine and XT line. Version 4 is targeted to the next generation of machines employing advanced network interface architectures that support enhanced offload capabilities.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2024

Tensor Extraction of Latent Features (TELF)

Tensor ELF is a user-friendly parallel tensor decomposition Python toolbox that includes a suite of machine learning algorithms for CPU and GPU architectures for the analysis of sparse and dense data including utility tools for pre-processing and post-processing.

Eren, Maksim↗

NASA NTRS1985

Reordering computations for parallel execution

The computations are reordered in the SOR algorithm to maintain the same asymptotic rate of convergence as the rowwise ordering to obtain parallelism at different levels. A parallel program is written to illustrate these ideas and actual machines for implementation of this program are discussed.

Adams, L.↗

NASA NTRS1990

Finite-State Codes

New class of codes with desirable properties constructed from block codes. Finite-state encoder is machine that has finite number of internal states. Receives k parallel information symbols and transmits n parallel code symbols at each clock pulse. Finite-state codes include both block and convolutional codes. Theory of finite-state codes used to study properties of error-correcting codes in general and to design new codes to various specifications.

Pollara, F.↗

DOE OSTI2022

Systems and methods for powder bed additive manufacturing anomaly detection

Detection and classification of anomalies for powder bed metal additive manufacturing. Anomalies, such as recoater blade impacts, binder deposition issues, spatter generation, and some porosities, are surface-visible at each layer of the building process. A multi-scaled parallel dynamic segmentation convolutional neural network architecture provides additive manufacturing machine and imaging system agnostic pixel-wise semantic segmentation of layer-wise powder bed image data. Learned knowledge is easily transferrable between different additive manufacturing machines. The anomaly detection can be conducted in real-time and provides accurate and generalizable results.

Scime, Luke R.↗

DOE OSTI2021

TQEA: Temporal Quantum Error Analysis

For the last few years quantum computers have captured researchers’ interest in solving complicated problems in finance, chemistry, machine learning, and many other NP problems. Quantum computing benefits from parallel computation, entanglement, and calculation speedup compared to classical computing. Even though the operations and number of qubits in quantum computers are limited, previous research showed that using quantum computers can be beneficial. These benefits gained researchers’ interest to work on the improvement of quantum computers and overcome the limitations by building more advanced quantum computers with fault tolerance capability, more physical qubits, and better operational gates. In practice, Quantum applications would be translated into Quantum Circuits and each circuit can contain one or many entanglements (s), single qubits operations, and readouts. The importance of running Quantum applications with controlled errors or without errors is critical to the desired result from Quantum readout. In this paper, we study the characteristics of superconducting quantum machines along with the previously proposed methods of Quantum error detection/correction and in order to build a better error model we propose a temporal method for analyzing the the behavior of qubits errors (T1 and T2), frequency of operation, entanglement errors, and readout errors. In order to detect and predict the behavior of each machine, we use simple exponential smoothing analysis on heuristic calibration data. Finally, based on observation, we show that based on frequency, qubits, readout and single operation errors we can build a temporal based model to study and predict the outcome of runs on each Quantum machine.

Baheri, Betis↗

DOE OSTI2021

High-Level Synthesis of Parallel Specifications Coupling Static and Dynamic Controllers

The increased need for efficient ways to implement domain-specific accelerators is driving design methodologies towards the use of abstractions higher than the Register Transfer Level (RTL). In this scenario, High Level Synthesis (HLS) plays a significant role by enabling the automatic generation of custom hardware accelerators starting from high level descriptions (e.g., C code). Conventional HLS tools exploit parallelism mostly at the Instruction Level (ILP). They statically schedule the input specifications, and build centralized Finite State Machine (FSM) controllers. However, aggressive exploitation of ILP in many applications has diminishing returns and, usually, centralized approaches do not efficiently exploit coarser parallelism because FSMs are inherently serial. In this paper we present a HLS framework able to synthesize applications that, beside ILP, also expose Task Level Parallelism (TLP). An application can expose TLP through annotations that identify the parallel functions (i.e., tasks). To generate accelerators that efficiently execute concur- rent tasks, we need to solve several issues: devise a mechanism to support concurrent execution flows, exploit memory parallelism, and manage synchronization. To support concurrent execution flows, we introduce a novel adaptive controller. The adaptive controller is composed of a set of interacting control elements that independently manage the execution of a single operation or function call. These control elements check dependencies and resource constraints at runtime, enabling as soon as possible execution. To support parallel access to shared memories and synchronization, we introduce a novel Hierarchical Memory Interface (HMI). With respect to previous solutions, the proposed interface supports multi-ported memories and atomic memory operations, which commonly occur in parallel programming. Our framework can generate the hardware implementation of C functions by employing two different approaches, depending on its characteristics. If a function exposes TLP, then the framework generates hardware implementations based on the adaptive controller. Otherwise, the framework implements the function by exploiting a more conventional FSM approach, which is optimized for ILP exploitation. We evaluate our framework on a set of parallel applications, and show substantial performance improvements (average speedup of 4.7) with limited area over- heads (average area increase of 5.48 times).

Castellana, Vito G.↗

DOE OSTI2021

High-Level Synthesis of Parallel Specifications Coupling Static and Dynamic Controllers

The increased need for efficient ways to implement domain-specific accelerators is driving design methodologies towards the use of abstractions higher than the Register Transfer Level (RTL). In this scenario, High Level Synthesis (HLS) plays a significant role by enabling the automatic generation of custom hardware accelerators starting from high level descriptions (e.g., C code). Conventional HLS tools exploit parallelism mostly at the Instruction Level (ILP). They statically schedule the input specifications, and build centralized Finite State Machine (FSM) controllers. However, aggressive exploitation of ILP in many applications has diminishing returns and, usually, centralized approaches do not efficiently exploit coarser parallelism because FSMs are inherently serial. In this paper we present a HLS framework able to synthesize applications that, beside ILP, also expose Task Level Parallelism (TLP). An application can expose TLP through annotations that identify the parallel functions (i.e., tasks). To generate accelerators that efficiently execute concur- rent tasks, we need to solve several issues: devise a mechanism to support concurrent execution flows, exploit memory parallelism, and manage synchronization. To support concurrent execution flows, we introduce a novel adaptive controller. The adaptive controller is composed of a set of interacting control elements that independently manage the execution of a single operation or function call. These control elements check dependencies and resource constraints at runtime, enabling as soon as possible execution. To support parallel access to shared memories and synchronization, we introduce a novel Hierarchical Memory Interface (HMI). With respect to previous solutions, the proposed interface supports multi-ported memories and atomic memory operations, which commonly occur in parallel programming. Our framework can generate the hardware implementation of C functions by employing two different approaches, depending on its characteristics. If a function exposes TLP, then the framework generates hardware implementations based on the adaptive controller. Otherwise, the framework implements the function by exploiting a more conventional FSM approach, which is optimized for ILP exploitation. We evaluate our framework on a set of parallel applications, and show substantial performance improvements (average speedup of 4.7) with limited area over- heads (average area increase of 5.48 times).

Castellana, Vito G.↗

NASA NTRS2004

Using the GeoFEST Faulted Region Simulation System

GeoFEST (the Geophysical Finite Element Simulation Tool) simulates stress evolution, fault slip and plastic/elastic processes in realistic materials, and so is suitable for earthquake cycle studies in regions such as Southern California. Many new capabilities and means of access for GeoFEST are now supported. New abilities include MPI-based cluster parallel computing using automatic PYRAMID/Parmetis-based mesh partitioning, automatic mesh generation for layered media with rectangular faults, and results visualization that is integrated with remote sensing data. The parallel GeoFEST application has been successfully run on over a half-dozen computers, including Intel Xeon clusters, Itanium II and Altix machines, and the Apple G5 cluster. It is not separately optimized for different machines, but relies on good domain partitioning for load-balance and low communication, and careful writing of the parallel diagonally preconditioned conjugate gradient solver to keep communication overhead low. Demonstrated thousand-step solutions for over a million finite elements on 64 processors require under three hours, and scaling tests show high efficiency when using more than (order of) 4000 elements per processor. The source code and documentation for GeoFEST is available at no cost from Open Channel Foundation. In addition GeoFEST may be used through a browser-based portal environment available to approved users. That environment includes semi-automated geometry creation and mesh generation tools, GeoFEST, and RIVA-based visualization tools that include the ability to generate a flyover animation showing deformations and topography. Work is in progress to support simulation of a region with several faults using 16 million elements, using a strain energy metric to adapt the mesh to faithfully represent the solution in a region of widely varying strain.

Geophyical Finite Element Simulation Tool (GeoFEST↗

DOE OSTI2024

Refining HPCToolkit for application performance analysis at exascale

As part of the US Department of Energy’s Exascale Computing Project (ECP), Rice University has been refining its HPCToolkit performance tools to better support measurement and analysis of applications executing on exascale supercomputers. To efficiently collect performance measurements of GPU-accelerated applications, HPCToolkit employs novel non-blocking data structures to communicate performance measurements between tool threads and application threads. To attribute performance information in detail to source lines, loop nests, and inlined call chains, HPCToolkit performs parallel analysis of large CPU and GPU binaries involved in the execution of an exascale application to rapidly recover mappings between machine instructions and source code. To analyze terabytes of performance measurements gathered during executions at exascale, HPCToolkit employs distributed-memory parallelism, multithreading, sparse data structures, and out-of-core streaming analysis algorithms. To support interactive exploration of profiles up to terabytes in size, HPCToolkit’s hpcviewer graphical user interface uses out-of-core methods to visualize performance data. The result of these efforts is that HPCToolkit now supports collection, analysis, and presentation of profiles and traces of GPU-accelerated applications at exascale. These improvements have enabled HPCToolkit to efficiently measure, analyze and explore terabytes of performance data for executions using as many as 64K MPI ranks and 64K GPU tiles on ORNL’s Frontier supercomputer. HPCToolkit’s support for measurement and analysis of GPU-accelerated applications has been employed to study a collection of open-science applications developed as part of ECP. This paper reports on these experiences, which provided insight into opportunities for tuning applications, strengths and weaknesses of HPCToolkit itself, as well as unexpected behaviors in executions at exascale.

Adhianto, Laksono↗

NASA NTRS1990

Execution time supports for adaptive scientific algorithms on distributed memory machines

Optimizations are considered that are required for efficient execution of code segments that consists of loops over distributed data structures. The PARTI (Parallel Automated Runtime Toolkit at ICASE) execution time primitives are designed to carry out these optimizations and can be used to implement a wide range of scientific algorithms on distributed memory machines. These primitives allow the user to control array mappings in a way that gives an appearance of shared memory. Computations can be based on a global index set. Primitives are used to carry out gather and scatter operations on distributed arrays. Communications patterns are derived at runtime, and the appropriate send and receive messages are automatically generated.

Berryman, Harry↗

NASA NTRS1990

Execution time support for scientific programs on distributed memory machines

Optimizations are considered that are required for efficient execution of code segments that consists of loops over distributed data structures. The PARTI (Parallel Automated Runtime Toolkit at ICASE) execution time primitives are designed to carry out these optimizations and can be used to implement a wide range of scientific algorithms on distributed memory machines. These primitives allow the user to control array mappings in a way that gives an appearance of shared memory. Computations can be based on a global index set. Primitives are used to carry out gather and scatter operations on distributed arrays. Communications patterns are derived at runtime, and the appropriate send and receive messages are automatically generated.

Berryman, Harry↗