Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “distributed memory parallel”

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 757 records · Page 42

Persistent Octrees for Parallel Mesh Refinement Through Non-Volatile Byte-Addressable Memory

We report that octree-based mesh adaptation has enabled simulations of complex physical phenomena. Existing meshing algorithms were proposed with the assumption that computer memory is volatile. Consequently, for failure recovery, the in-core algorithms need to save memory states as snapshots with slow file I/Os. The out-of-core algorithms store octants on disks for persistence. However, neither of them was designed to leverage unique characteristics of non-volatile byte-addressable memory (NVBM). In this paper, we propose a novel data structure Distributed Persistent Merged octree (DPM-octree) for both meshing and in-memory storage of persistent octrees using NVBM. It is a multi-version data structure and can recover from failures using its earlier persistent version stored in NVBM. In addition, we design a feature-directed sampling approach to help dynamically transform the DPM-octree layout for reducing NVBM-induced memory write latency. DPM-octree uses parity trees which are created using erasure coding and stored in NVBM to support low-latency in-memory octant recovery after data loss. DPM-octree has been successfully integrated with Gerris software for simulation of fluid dynamics. Finally, our experimental results with real-world scientific workloads show that DPM-octree scales up to 1.1 billion mesh elements with 1000 processors on the Titan supercomputer.

97 MATHEMATICS AND COMPUTING↗

A message passing kernel for the hypercluster parallel processing test bed

A Message-Passing Kernel (MPK) for the Hypercluster parallel-processing test bed is described. The Hypercluster is being developed at the NASA Lewis Research Center to support investigations of parallel algorithms and architectures for computational fluid and structural mechanics applications. The Hypercluster resembles the hypercube architecture except that each node consists of multiple processors communicating through shared memory. The MPK efficiently routes information through the Hypercluster, using a message-passing protocol when necessary and faster shared-memory communication whenever possible. The MPK also interfaces all of the processors with the Hypercluster operating system (HYCLOPS), which runs on a Front-End Processor (FEP). This approach distributes many of the I/O tasks to the Hypercluster processors and eliminates the need for a separate I/O support program on the FEP.

Blech, Richard A.↗

Scalable Work Stealing

Irregular and dynamic parallel applications pose significant challenges to achieving scalable performance on large-scale multicore clusters. These applications often require ongoing, dynamic load balancing in order to maintain efficiency. While effective at small scale, centralized load balancing schemes quickly become a bottleneck on large-scale clusters. Work stealing is a popular approach to distributed dynamic load balancing; however its performance on large-scale clusters is not well understood. Prior work on work stealing has largely focused on shared memory machines. In this work we investigate the design and scalability of work stealing on modern distributed memory systems. We demonstrate high efficiency and low overhead when scaling to 8,192 processors for three benchmark codes: a producer-consumer benchmark, the unbalanced tree search benchmark, and a multiresolution analysis kernel.

99 GENERAL AND MISCELLANEOUS↗

Sequence length scaling in vision transformers for scientific images on frontier

Vision Transformers (ViTs) are pivotal for foundational models in scientific imagery, including Earth science applications, due to their capability to process large sequence lengths. While transformers for text have inspired scaling sequence lengths in ViTs, adapting these for ViTs introduces unique challenges. We develop distributed sequence parallelism for ViTs, enabling them to handle up to 1M tokens. Our approach, leveraging DeepSpeed-Ulysses and Long-Sequence-Segmentation with model sharding, is the first to apply sequence parallelism in ViT training, achieving a 94% batch scaling efficiency on 2,048 AMD-MI250X GPUs. Evaluating sequence parallelism in ViTs, particularly in models up to 10B parameters, highlighted substantial bottlenecks. We countered these with hybrid sequence, pipeline, and flash attention strategies, to scale beyond single GPU memory limits. Our method significantly enhances climate modeling accuracy by 20% in temperature predictions, marking the first training of a vision transformer model to convergence with a sequence length of 188K tokens, using full self-attention.

Tsaris, Aristeidis (aris) [ORNL] (ORCID:0000000277↗

ELSI: A unified software interface for Kohn–Sham electronic structure solvers

Solving the electronic structure from a generalized or standard eigenproblem is often the bottleneck in large scale calculations based on Kohn-Sham density-functional theory. This problem must be addressed by essentially all current electronic structure codes, based on similar matrix expressions, and by high-performance computation. We here present a unified software interface, ELSI, to access different strategies that address the Kohn-Sham eigenvalue problem. Currently supported algorithms include the dense generalized eigensolver library ELPA, the orbital minimization method implemented in libOMM, and the pole expansion and selected inversion (PEXSI) approach with lower computational complexity for semilocal density functionals. The ELSI interface aims to simplify the implementation and optimal use of the different strategies, by offering (a) a unified software framework designed for the electronic structure solvers in Kohn-Sham density-functional theory; (b) reasonable default parameters for a chosen solver; (c) automatic conversion between input and internal working matrix formats, and in the future (d) recommendation of the optimal solver depending on the specific problem. As a result, comparative benchmarks are shown for system sizes up to 11,520 atoms (172,800 basis functions) on distributed memory supercomputing architectures.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Exact signed distance fields using parallel Fast Sweeping Method

Signed distance fields are often used in multiphysics simulations to track material interfaces. We present a simple methodology based on the fast sweeping method to generate the exact signed distance from triangular meshes and linear paths on Cartesian grids. The methodology propagates the closest primitive to the boundary to the rest of the domain following the characteristics. A local upwind criterion is used to decide between the new and existing closest primitive at each grid point while capturing the correct sign of the global function. The methodology has optimal computational complexity and runs efficiently in distributed-memory architectures. We include 2D and 3D test cases along with a resolution study up to 0.512 trillion zones and 1,000 computer cores. The solution strategy can also be applied to other types of meshes or collections of primitives.

97 MATHEMATICS AND COMPUTING↗

On Data Transfers Over Wide-Area Dedicated Connections

Dedicated wide-area network connections are employed in big data and high-performance computing scenarios, since the absence of cross-traffic promises to make it easier to analyze and optimize data transfers over them. However, nonlinear transport dynamics and end-system complexity due to multi-core hosts and distributed file systems make these tasks surprisingly challenging. We present an overview of methods to analyze memory and disk file transfers using extensive measurements over 10 Gbps physical and emulated connections with 0–366 ms round trip times (RTTs). For memory transfers, we derive performance profiles of TCP and UDT throughput as a function of RTT, which show concave regions in contrast to entirely convex regions predicted by previous models. These highly desirable concave regions can be expanded by utilizing large buffers and more parallel flows. We also present Poincar´e maps and Lyapunov exponents of TCP and UDT throughputtraces that indicate complex throughput dynamics. For disk file transfers, we show that throughput can be optimized using a combination of parallel I/O and network threads under direct I/O mode. Our initial throughput measurements of Lustre filesystems mounted over long-haul connections using LNet routers show convex profiles indicative of I/O limits.

Rao, Nageswara S.↗

Efficient partitioning and assignment on programs for multiprocessor execution

The general problem studied is that of segmenting or partitioning programs for distribution across a multiprocessor system. Efficient partitioning and the assignment of program elements are of great importance since the time consumed in this overhead activity may easily dominate the computation, effectively eliminating any gains made by the use of the parallelism. In this study, the partitioning of sequentially structured programs (written in FORTRAN) is evaluated. Heuristics, developed for similar applications are examined. Finally, a model for queueing networks with finite queues is developed which may be used to analyze multiprocessor system architectures with a shared memory approach to the problem of partitioning. The properties of sequentially written programs form obstacles to large scale (at the procedure or subroutine level) parallelization. Data dependencies of even the minutest nature, reflecting the sequential development of the program, severely limit parallelism. The design of heuristic algorithms is tied to the experience gained in the parallel splitting. Parallelism obtained through the physical separation of data has seen some success, especially at the data element level. Data parallelism on a grander scale requires models that accurately reflect the effects of blocking caused by finite queues. A model for the approximation of the performance of finite queueing networks is developed. This model makes use of the decomposition approach combined with the efficiency of product form solutions.

Standley, Hilda M.↗

The PISCES 2 parallel programming environment

PISCES 2 is a programming environment for scientific and engineering computations on MIMD parallel computers. It is currently implemented on a flexible FLEX/32 at NASA Langley, a 20 processor machine with both shared and local memories. The environment provides an extended Fortran for applications programming, a configuration environment for setting up a run on the parallel machine, and a run-time environment for monitoring and controlling program execution. This paper describes the overall design of the system and its implementation on the FLEX/32. Emphasis is placed on several novel aspects of the design: the use of a carefully defined virtual machine, programmer control of the mapping of virtual machine to actual hardware, forces for medium-granularity parallelism, and windows for parallel distribution of data. Some preliminary measurements of storage use are included.

Pratt, Terrence W.↗

Fast analysis of molecular dynamics trajectories with graphics processing units-Radial distribution function histogramming

The calculation of radial distribution functions (RDFs) from molecular dynamics trajectory data is a common and computationally expensive analysis task. The rate limiting step in the calculation of the RDF is building a histogram of the distance between atom pairs in each trajectory frame. Here we present an implementation of this histogramming scheme for multiple graphics processing units (GPUs). The algorithm features a tiling scheme to maximize the reuse of data at the fastest levels of the GPU's memory hierarchy and dynamic load balancing to allow high performance on heterogeneous configurations of GPUs. Several versions of the RDF algorithm are presented, utilizing the specific hardware features found on different generations of GPUs. We take advantage of larger shared memory and atomic memory operations available on state-of-the-art GPUs to accelerate the code significantly. The use of atomic memory operations allows the fast, limited-capacity on-chip memory to be used much more efficiently, resulting in a fivefold increase in performance compared to the version of the algorithm without atomic operations. The ultimate version of the algorithm running in parallel on four NVIDIA GeForce GTX 480 (Fermi) GPUs was found to be 92 times faster than a multithreaded implementation running on an Intel Xeon 5550 CPU. On this multi-GPU hardware, the RDF between two selections of 1,000,000 atoms each can be calculated in 26.9 s per frame. The multi-GPU RDF algorithms described here are implemented in VMD, a widely used and freely available software package for molecular dynamics visualization and analysis.

97 MATHEMATICS AND COMPUTING↗

Integrating PGAS and MPI-based Graph Analysis

This project demonstrates that Chapel programs can interface with MPI-based libraries written in C++ without storing multiple copies of shared data. Chapel is a language for productive parallel computing using global address spaces (PGAS). We identified two approaches to interface Chapel code with the MPI-based Grafiki and Trilinos libraries. The first uses a single Chapel executable to call a C function that interacts with the C++ libraries. The second uses the mmap function to allow separate executables to read and write to the same block of memory on a node. We also encapsulated the second approach in Docker/Singularity containers to maximize ease of use. Comparisons of the two approaches using shared and distributed memory installations of Chapel show that both approaches provide similar scalability and performance.

97 MATHEMATICS AND COMPUTING↗

Programmability of the HPCS Languages: A Case Study with a Quantum Chemistry Kernel (Extended Version)

As high-end computer systems present users with rapidly increasing numbers of processors, possibly also incorporating attached co-processors, programmers are increasingly challenged to express the necessary levels of concurrency with the dominant parallel programming model, Fortran+MPI+OpenMP (or minor variations). In this paper, we examine the languages developed under the DARPA High-Productivity Computing Systems (HPCS) program (Chapel, Fortress, and X10) as representatives of a different parallel programming model which might be more effective on coming high-performance systems. The application used in this study is the Hartree-Fock method from quantum chemistry, which combines access to distributed data with a task-parallel algorithm and is characterized by significant irregularity in the computational tasks. We present several different implementation strategies for load balancing of the task parallel computation, as well as distributed array operations, in each of the three languages. We conclude that the HPCS languages provide a wide variety of mechanisms for expressing parallelism, which can be combined at multiple levels, making them quite expressive for this problem.

97 MATHEMATICS AND COMPUTING↗

Mads.jl

Mads.jl (Model analysis and decision support in Julia) is a code that streamlines the process of using data and models for analysis and decision support. It is based on another open-source code developed at LANL and written in C/C++ (MADS; http://mads.lanl.gov; LA-CC-11- 035). Mads.jl can work with external models of arbitrary complexity as well as built-in models of flow and transport in porous media. It enables a number of data- and model-based analyses including model calibration, sensitivity analysis, uncertainty quantification, and decision analysis. The code also can use a series of alternative adaptive computational techniques for Bayesian sampling, Monte Carlo, and Bayesian Information-Gap Decision Theory. The code is implemented in the Julia programming language, and has high-performance (parallel) and memory management capabilities. The code uses a series of third party modules developed by others. The code development will also include contributions to the existing third party modules written in Julia; this contributions will be important for the efficient implementation of the algorithm used by Mads.jl. The code also uses a series of LANL developed modules that are developed by Dan O'Malley; these modules will be also a part of the Mads.jl release. Mads.jl will be released under GPL V3 license. The code will be distributed as a Git repo at gitlab.com and github.com. Mads.jl manual and documentation will be posted at madsjulia.lanl.gov.

Vesselinov, Velimir↗

Mads.jl

Mads.jl (Model analysis and decision support in Julia) is a code that streamlines the process of using data and models for analysis and decision support. It is based on another open-source code developed at LANL and written in C/C++ (MADS; http://mads.lanl.gov; LA-CC-11- 035). Mads.jl can work with external models of arbitrary complexity as well as built-in models of flow and transport in porous media. It enables a number of data- and model-based analyses including model calibration, sensitivity analysis, uncertainty quantification, and decision analysis. The code also can use a series of alternative adaptive computational techniques for Bayesian sampling, Monte Carlo, and Bayesian Information-Gap Decision Theory. The code is implemented in the Julia programming language, and has high-performance (parallel) and memory management capabilities. The code uses a series of third party modules developed by others. The code development will also include contributions to the existing third party modules written in Julia; this contributions will be important for the efficient implementation of the algorithm used by Mads.jl. The code also uses a series of LANL developed modules that are developed by Dan O'Malley; these modules will be also a part of the Mads.jl release. Mads.jl will be released under GPL V3 license. The code will be distributed as a Git repo at gitlab.com and github.com. Mads.jl manual and documentation will be posted at madsjulia.lanl.gov.

Vesselinov, Velimir↗

Partitioning problems in parallel, pipelined and distributed computing

The problem of optimally assigning the modules of a parallel program over the processors of a multiple computer system is addressed. A Sum-Bottleneck path algorithm is developed that permits the efficient solution of many variants of this problem under some constraints on the structure of the partitions. In particular, the following problems are solved optimally for a single-host, multiple satellite system: partitioning multiple chain structured parallel programs, multiple arbitrarily structured serial programs and single tree structured parallel programs. In addition, the problems of partitioning chain structured parallel programs across chain connected systems and across shared memory (or shared bus) systems are also solved under certain constraints. All solutions for parallel programs are equally applicable to pipelined programs. These results extend prior research in this area by explicitly taking concurrency into account and permit the efficient utilization of multiple computer architectures for a wide range of problems of practical interest.

Bokhari, S.↗

Parallel solution of high-order numerical schemes for solving incompressible flows

A new parallel numerical scheme for solving incompressible steady-state flows is presented. The algorithm uses a finite-difference approach to solving the Navier-Stokes equations. The algorithms are scalable and expandable. They may be used with only two processors or with as many processors as are available. The code is general and expandable. Any size grid may be used. Four processors of the NASA LeRC Hypercluster were used to solve for steady-state flow in a driven square cavity. The Hypercluster was configured in a distributed-memory, hypercube-like architecture. By using a 50-by-50 finite-difference solution grid, an efficiency of 74 percent (a speedup of 2.96) was obtained.

Milner, Edward J.↗

Forward-Weighted CADIS Method for Variance Reduction of Monte Carlo Reactor Analyses

Current state-of-the-art tools and methods used to perform 'real' commercial reactor analyses use high-fidelity transport codes to produce few-group parameters at the assembly level for use in low-order methods applied at the core level. Monte Carlo (MC) methods, which allow detailed and accurate modeling of the full geometry and energy details and are considered the 'gold standard' for radiation transport solutions, are playing an ever-increasing role in correcting and/or verifying the several-decade-old methodology used in current practice. However, the prohibitive computational requirements associated with obtaining fully converged system-wide solutions restrict the role of MC to benchmarking deterministic results at a limited number of state-points for a limited number of relevant quantities. A goal of current research at Oak Ridge National Laboratory (ORNL) is to change this paradigm by enabling the direct use of MC for full-core reactor analyses. The most significant of the many technical challenges that must be overcome is the slow non-uniform convergence of system-wide MC estimates and the memory requirements associated with detailed solutions throughout a reactor (problems involving hundreds of millions of different material and tally regions due to fuel irradiation, temperature distributions, and the needs associated with multi-physics code coupling). To address these challenges, research has focused on development in the following two areas: (1) a hybrid deterministic/MC method for determining high-precision fluxes throughout the problem space in k-eigenvalue problems and (2) an efficient MC domain-decomposition algorithm that partitions the problem phase space onto multiple processors for massively parallel systems, with statistical uncertainty estimation. The focus of this paper is limited to the first area mentioned above. It describes the FW-CADIS method applied to variance reduction of MC reactor analyses and provides initial results for calculating group-wise fluxes throughout a generic 2-D pressurized water reactor (PWR) quarter core model.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Applications Performance on NAS Intel Paragon XP/S - 15#

The Numerical Aerodynamic Simulation (NAS) Systems Division received an Intel Touchstone Sigma prototype model Paragon XP/S- 15 in February, 1993. The i860 XP microprocessor with an integrated floating point unit and operating in dual -instruction mode gives peak performance of 75 million floating point operations (NIFLOPS) per second for 64 bit floating point arithmetic. It is used in the Paragon XP/S-15 which has been installed at NAS, NASA Ames Research Center. The NAS Paragon has 208 nodes and its peak performance is 15.6 GFLOPS. Here, we will report on early experience using the Paragon XP/S- 15. We have tested its performance using both kernels and applications of interest to NAS. We have measured the performance of BLAS 1, 2 and 3 both assembly-coded and Fortran coded on NAS Paragon XP/S- 15. Furthermore, we have investigated the performance of a single node one-dimensional FFT, a distributed two-dimensional FFT and a distributed three-dimensional FFT Finally, we measured the performance of NAS Parallel Benchmarks (NPB) on the Paragon and compare it with the performance obtained on other highly parallel machines, such as CM-5, CRAY T3D, IBM SP I, etc. In particular, we investigated the following issues, which can strongly affect the performance of the Paragon: a. Impact of the operating system: Intel currently uses as a default an operating system OSF/1 AD from the Open Software Foundation. The paging of Open Software Foundation (OSF) server at 22 MB to make more memory available for the application degrades the performance. We found that when the limit of 26 NIB per node out of 32 MB available is reached, the application is paged out of main memory using virtual memory. When the application starts paging, the performance is considerably reduced. We found that dynamic memory allocation can help applications performance under certain circumstances. b. Impact of data cache on the i860/XP: We measured the performance of the BLAS both assembly coded and Fortran coded. We found that the measured performance of assembly-coded BLAS is much less than what memory bandwidth limitation would predict. The influence of data cache on different sizes of vectors is also investigated using one-dimensional FFTs. c. Impact of processor layout: There are several different ways processors can be laid out within the two-dimensional grid of processors on the Paragon. We have used the FFT example to investigate performance differences based on processors layout.

Saini, Subhash↗