Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “CRAY”

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 397 records · Page 22

Using Hybrid Model OpenSHMEM + CUDA to Implement the SHOC Benchmark Suite

This work describes the process of porting the Scalable HeterOgeneous Computing (SHOC) benchmark suite from the hybrid MPI + CUDA implementation to OpenSHMEM + CUDA. SHOC includes a wide variety of benchmark kernels used to measure accelerator performance in both single node and cluster configurations. The hybrid model implementation attempts to place all major computation on accelerator devices, and uses MPI to synchronize and aggregate results. In some cases, MPI Groups are used to gradually reduce the number of accelerators used for computation as the problem size drops. Porting this behavior to OpenSHMEM required implementing several synchronizing collective operations, and using SHMEM teams to replace MPI Group functionality. Benchmark results on a Cray XK7 system with one GPU per compute node show that SHMEM performance is equal to MPI performance in these hybrid tasks. These results and porting experience show that using OpenSHMEM for accelerator devices benefits from adding functionality for synchronization and teams, and would further benefit from adding support for communication within accelerator kernels.

97 MATHEMATICS AND COMPUTING↗

Balancing Performance and Portability with Containers in HPC: An OpenSHMEM Example

There is a growing interest in using Linux containers to streamline software development and application deployment. A container enables the user to bundle the salient elements of the software stack from an application’s perspective. In this paper, we discuss initial experiences in using the Open MPI implementation of OpenSHMEM with containers on HPC resources. We provide a brief overview of two container runtimes, Docker & Singularity, highlighting elements that are of interest for HPC users. The Docker platform offers a rich set of services that are widely used in enterprise environments, whereas Singularity is an emerging container runtime that is specifically written for use on HPC systems. We describe our procedure for container assembly and deployment that strives to maintain the portability of the container-based application. We show performance results for the Graph500 benchmark running along the typical continuum of development testbed up to full production supercomputer (ORNL’s Titan). The results show consistent performance between the native and Singularity (container) tests. The results also showed an unexplained drop in performance when using the Cray Gemini network with Open MPI’s OpenSHMEM, which was unrelated to the container usage.

Naughton III, Thomas↗

High-performance Computing Applied to Semantic Databases

To-date, the application of high-performance computing resources to Semantic Web data has largely focused on commodity hardware and distributed memory platforms. In this paper we make the case that more specialized hardware can offer superior scaling and close to an order of magnitude improvement in performance. In particular we examine the Cray XMT. Its key characteristics, a large, global shared-memory, and processors with a memory-latency tolerant design, offer an environment conducive to programming for the Semantic Web and have engendered results that far surpass current state of the art. We examine three fundamental pieces requisite for a fully functioning semantic database: dictionary encoding, RDFS inference, and query processing. We show scaling up to 512 processors (the largest configuration we had available), and the ability to process 20 billion triples completely in-memory.

99 GENERAL AND MISCELLANEOUS↗

GPU Acceleration of the Locally Selfconsistent Multiple Scattering Code for First Principles Calculation of the Ground State and Statistical Physics of Materials

The Locally Self-consistent Multiple Scattering (LSMS) code solves the first principles Density Functional theory Kohn-Sham equation for a wide range of materials with a special focus on metals, alloys and metallic nano-structures. It has traditionally exhibited near perfect scalability on massively parallel high performance computer architectures. We present our efforts to exploit GPUs to accelerate the LSMS code to enable first principles calculations of O(100,000) atoms and statistical physics sampling of finite temperature properties. Using the Cray XK7 system Titan at the Oak Ridge Leadership Computing Facility we achieve a sustained performance of 14.5PFlop/s and a speedup of 8.6 compared to the CPU only code.

Eisenbach, Markus↗

Runtime Techniques to Enable a Highly-Scalable Global Address Space Model for Petascale Computing

Over the past decade, the trajectory to the petascale has been built on increased complexity and scale of the underlying parallel architectures. Meanwhile, software developers have struggled to provide tools that maintain the productivity of computational science teams using these new systems. In this regard, Global Address Space (GAS) programming models provide a straightforward and easy to use addressing model, which can lead to improved productivity. However, the scalability of GAS depends directly on the design and implementation of the runtime system on the target petascale distributed-memory architecture. In this paper, we describe the design, implementation, and optimization of the Aggregate Remote Memory Copy Interface (ARMCI) runtime library on the Cray XT5 2.3 PetaFLOPs computer at Oak Ridge National Laboratory. We optimized our implementation with the flow intimation technique that we have introduced in this paper. Our optimized ARMCI implementation improves scalability of both the Global Arrays programming model and a real-world chemistry application—NWChem—from small jobs up through 180,000 cores.

Computer Science↗

Nekbone performance on GPUs with OpenACC and CUDA Fortran implementations

We present a hybrid GPU implementation and performance analysis of Nekbone, which represents one of the core kernels of the incompressible Navier–Stokes solver Nek5000. The implementation is based on OpenACC and CUDA Fortran for local parallelization of the compute-intensive matrix–matrix multiplication part, which significantly minimizes the modification of the existing CPU code while extending the simulation capability of the code to GPU architectures. Our discussion includes the GPU results of OpenACC interoperating with CUDA Fortran and the gather–scatter operations with GPUDirect communication. We demonstrate performance of up to 552 Tflops on 16, 384 GPUs of the OLCF Cray XK7 Titan.

Computer Science↗

Queue wait time prediction in high performance computing (HPC) systems

High Performance Computing (HPC) systems are critical enablers for groundbreaking scientific research across various domains. Efficient resource allocation, facilitated by job scheduling, is paramount for maximizing the utilization of HPC systems. However, the variability in wait times for queued jobs poses challenges for users, necessitating accurate job wait time estimation. This paper explores the influence of job characteristics, including job size (the number of nodes requested and walltime), the queue to which the job is submitted and other resource requirements, on job wait times in leadership-class HPC systems. Focusing on the Theta Cray XC40 and Polaris machines at Argonne National Laboratory, the study evaluates the performance of different supervised learning algorithms in predicting job wait times. It also evaluates the impact of data preprocessing, including outlier detection, Principal Component Analysis (PCA), and feature selection, on the performance of wait time prediction models. The findings reveal insights into the relationship between job characteristics and wait times, offering a foundation for optimizing resource allocation and enhancing user experience. The methodologies and tools developed in this study are adaptable to other leadership-class HPC systems, providing a valuable contribution to the broader HPC community aiming to improve job scheduling efficiency and user satisfaction.

Okafor, Nwamaka↗

A Hybrid MPI/OpenMP Approach for Parallel Groundwater Model Calibration on Multicore Computers

Groundwater model calibration is becoming increasingly computationally time intensive. We describe a hybrid MPI/OpenMP approach to exploit two levels of parallelism in software and hardware to reduce calibration time on multicore computers with minimal parallelization effort. At first, HydroGeoChem 5.0 (HGC5) is parallelized using OpenMP for a uranium transport model with over a hundred species involving nearly a hundred reactions, and a field scale coupled flow and transport model. In the first application, a single parallelizable loop is identified to consume over 97% of the total computational time. With a few lines of OpenMP compiler directives inserted into the code, the computational time reduces about ten times on a compute node with 16 cores. The performance is further improved by selectively parallelizing a few more loops. For the field scale application, parallelizable loops in 15 of the 174 subroutines in HGC5 are identified to take more than 99% of the execution time. By adding the preconditioned conjugate gradient solver and BICGSTAB, and using a coloring scheme to separate the elements, nodes, and boundary sides, the subroutines for finite element assembly, soil property update, and boundary condition application are parallelized, resulting in a speedup of about 10 on a 16-core compute node. The Levenberg-Marquardt (LM) algorithm is added into HGC5 with the Jacobian calculation and lambda search parallelized using MPI. With this hybrid approach, compute nodes at the number of adjustable parameters (when the forward difference is used for Jacobian approximation), or twice that number (if the center difference is used), are used to reduce the calibration time from days and weeks to a few hours for the two applications. This approach can be extended to global optimization scheme and Monte Carol analysis where thousands of compute nodes can be efficiently utilized.

54 ENVIRONMENTAL SCIENCES↗

Performance of fully-coupled algebraic multigrid preconditioners for large-scale VMS resistive MHD

Here, we explore the current performance and scaling of a fully-implicit stabilized unstructured finite element (FE) variational multiscale (VMS) capability for large-scale simulations of 3D incompressible resistive magnetohydrodynamics (MHD). The large-scale linear systems that are generated by a Newton nonlinear solver approach are iteratively solved by preconditioned Krylov subspace methods. The efficiency of this approach is critically dependent on the scalability and performance of the algebraic multigrid preconditioner. Our study considers the performance of the numerical methods as recently implemented in the second-generation Trilinos implementation that is 64-bit compliant and is not limited by the 32-bit global identifiers of the original Epetra-based Trilinos. The study presents representative results for a Poisson problem on 1.6 million cores of an IBM Blue Gene/Q platform to demonstrate very large-scale parallel execution. Additionally, results for a more challenging steady-state MHD generator and a transient solution of a benchmark MHD turbulence calculation for the full resistive MHD system are also presented. These results are obtained on up to 131,000 cores of a Cray XC40 and one million cores of a BG/Q system.

97 MATHEMATICS AND COMPUTING↗

Sassena — X-ray and neutron scattering calculated from molecular dynamics trajectories using massively parallel computers

Massively parallel computers now permit the molecular dynamics (MD) simulation of multi-million atom systems on time scales up to the microsecond. However, the subsequent analysis of the resulting simulation trajectories has now become a high performance computing problem in itself. Here, we present software for calculating X-ray and neutron scattering intensities from MD simulation data that scales well on massively parallel supercomputers. The calculation and data staging schemes used maximize the degree of parallelism and minimize the IO bandwidth requirements. The strong scaling tested on the Jaguar Petaflop Cray XT5 at Oak Ridge National Laboratory exhibits virtually linear scaling up to 7000 cores for most benchmark systems. Since both MPI and thread parallelism is supported, the software is flexible enough to cover scaling demands for different types of scattering calculations. The result is a high performance tool capable of unifying large-scale supercomputing and a wide variety of neutron/synchrotron technology.

Computer Science↗

Dynamic load balancing for petascale quantum Monte Carlo applications: The Alias method

Diffusion Monte Carlo is the most accurate widely used Quantum Monte Carlo method for the electronic structure of materials, but it requires frequent load balancing or population redistribution steps to maintain efficiency and avoid accumulation of systematic errors on parallel machines. The load balancing step can be a significant factor affecting performance, and will become more important as the number of processing elements increases. We propose a new dynamic load balancing algorithm, the Alias Method, and evaluate it theoretically and empirically. An important feature of the new algorithm is that the load can be perfectly balanced with each process receiving at most one message. It is also optimal in the maximum size of messages received by any process. We also optimize its implementation to reduce network contention, a process facilitated by the low messaging requirement of the algorithm. Empirical results on the petaflop Cray XT Jaguar supercomputer at ORNL showing up to 30% improvement in performance on 120,000 cores. The load balancing algorithm may be straightforwardly implemented in existing codes. The algorithm may also be employed by any method with many near identical computational tasks that requires load balancing.

97 MATHEMATICS AND COMPUTING↗

GPU acceleration of the Locally Selfconsistent Multiple Scattering code for first principles calculation of the ground state and statistical physics of materials

The Locally Self-consistent Multiple Scattering (LSMS) code solves the first principles Density Functional theory Kohn–Sham equation for a wide range of materials with a special focus on metals, alloys and metallic nano-structures. It has traditionally exhibited near perfect scalability on massively parallel high performance computer architectures. In this paper, we present our efforts to exploit GPUs to accelerate the LSMS code to enable first principles calculations of O(100,000) atoms and statistical physics sampling of finite temperature properties. We reimplement the scattering matrix calculation for GPUs with a block matrix inversion algorithm that only uses accelerator memory. Finally, using the Cray XK7 system Titan at the Oak Ridge Leadership Computing Facility we achieve a sustained performance of 14.5PFlop/s and a speedup of 8.6 compared to the CPU only code.

36 MATERIALS SCIENCE↗

An efficient and portable SIMD algorithm for charge/current deposition in Particle-In-Cell codes

In current computer architectures, data movement (from die to network) is by far the most energy consuming part of an algorithm (≈20pJ/word on-die to ≈10,000 pJ/word on the network). To increase memory locality at the hardware level and reduce energy consumption related to data movement, future exascale computers tend to use many-core processors on each compute nodes that will have a reduced clock speed to allow for efficient cooling. To compensate for frequency decrease, machine vendors are making use of long SIMD instruction registers that are able to process multiple data with one arithmetic operator in one clock cycle. SIMD register length is expected to double every four years. As a consequence, Particle-In-Cell (PIC) codes will have to achieve good vectorization to fully take advantage of these upcoming architectures. In this paper, we present a new algorithm that allows for efficient and portable SIMD vectorization of current/charge deposition routines that are, along with the field gathering routines, among the most time consuming parts of the PIC algorithm. Our new algorithm uses a particular data structure that takes into account memory alignment constraints and avoids gather/scat;ter instructions that can significantly affect vectorization performances on current CPUs. The new algorithm was successfully implemented in the 3D skeleton PIC code PICSAR and tested on Haswell Xeon processors (AVX2-256 bits wide data registers). Results show a factor of ×2 to ×2.5 speed-up in double precision for particle shape factor of orders 1–3. The new algorithm can be applied as is on future KNL (Knights Landing) architectures that will include AVX-512 instruction sets with 512 bits register lengths (8 doubles/16 singles). Program summary Program Title: vec_deposition Program Files doi:http://dx.doi.org/10.17632/nh77fv9k8c.1 Licensing provisions: BSD 3-Clause Programming language: Fortran 90 External routines/libraries: OpenMP > 4.0 Nature of problem: Exascale architectures will have many-core processors per node with long vector data registers capable of performing one single instruction on multiple data during one clock cycle. Data register lengths are expected to double every four years and this pushes for new portable solutions for efficiently vectorizing Particle-In-Cell codes on these future many-core architectures. One of the main hotspot routines of the PIC algorithm is the current/charge deposition for which there is no efficient and portable vector algorithm. Solution method: Here we provide an efficient and portable vector algorithm of current/charge deposition routines that uses a new data structure, which significantly reduces gather/scatter operations. Vectorization is controlled using OpenMP 4.0 compiler directives for vectorization which ensures portability across different architectures. Restrictions: Here we do not provide the full PIC algorithm with an executable but only vector routines for current/charge deposition. These scalar/vector routines can be used as library routines in your 3D Particle-In-Cell code. However, to get the best performances out of vector routines you have to satisfy the two following requirements: (1) Your code should implement particle tiling (as explained in the manuscript) to allow for maximized cache reuse and reduce memory accesses that can hinder vector performances. The routines can be used directly on each particle tile. (2) You should compile your code with a Fortran 90 compiler (e.g Intel, gnu or cray) and provide proper alignment flags and compiler alignment directives (more details in README file).

97 MATHEMATICS AND COMPUTING↗

Massively parallel and linear-scaling algorithm for second-order Moller–Plesset perturbation theory applied to the study of supramolecular wires

Here, we present a scalable cross-platform hybrid MPI/OpenMP/OpenACC implementation of the Divide–Expand–Consolidate (DEC) formalism with portable performance on heterogeneous HPC architectures. The Divide–Expand–Consolidate formalism is designed to reduce the steep computational scaling of conventional many-body methods employed in electronic structure theory to linear scaling, while providing a simple mechanism for controlling the error introduced by this approximation. Our massively parallel implementation of this general scheme has three levels of parallelism, being a hybrid of the loosely coupled task-based parallelization approach and the conventional MPI +X programming model, where X is either OpenMP or OpenACC. We demonstrate strong and weak scalability of this implementation on heterogeneous HPC systems, namely on the GPU-based Cray XK7 Titan supercomputer at the Oak Ridge National Laboratory. Using the “resolution of the identity second-order Moller–Plesset perturbation theory” (RI-MP2) as the physical model for simulating correlated electron motion, the linear-scaling DEC implementation is applied to 1-aza-adamantane-trione (AAT) supramolecular wires containing up to 40 monomers (2440 atoms, 6800 correlated electrons, 24 440 basis functions and 91 280 auxiliary functions). This represents the largest molecular system treated at the MP2 level of theory, demonstrating an efficient removal of the scaling wall pertinent to conventional quantum many-body methods.

97 MATHEMATICS AND COMPUTING↗

Large-scale GW calculations on pre-exascale HPC systems

The ab initio GW approach is a rigorous Green’s-function-based framework that can be employed to compute electronic excitation properties of a wide variety of materials such as extended systems, molecules, as well as confined and nanostructured materials with a very satisfactory accuracy. However, GW calculations on complex systems are often hindered by the high computational cost associated with the method. Here, in this study, we demonstrate how to significantly speedup GW calculations with a novel algorithm for the computationally intense kernel based on a non-blocking chunked cyclic communication scheme that minimizes latency in MPI messages, allows for overlapping of communication and computation, and improves cache usage in matrix multiplication operations. The optimized version of the code, implemented in the BerkeleyGW software package, is capable of scaling well to the full Cori computer at NERSC (Cray XC40) and achieves over 11 Peta FLOP/s of sustained performance. We showcase our work by performing large-scale GW calculations of defect structures on silicon, which require simulation cells containing over 1700 atoms, and which can now be efficiently executed in just a few minutes on large pre-exascale high-performance computing systems.

97 MATHEMATICS AND COMPUTING↗

GPU acceleration and performance of the particle-beam-dynamics code Elegant

Elegant is an accelerator physics and particle-beam dynamics code widely used for modeling and design of a variety of high-energy particle accelerators and accelerator-based systems. We discuss a recently developed version of the code that can take advantage of CUDA-enabled graphics processing units (GPUs) to achieve significantly improved performance for a large class of simulations that are important in practice. The GPU version is largely defined by a framework that simplifies implementations of the fundamental kernel types that are used by Elegant: particle operations, reductions, particle loss, histograms, array convolutions and random number generation. Accelerated performance on the Titan Cray XK-7 supercomputer is approximately 6–10 times better with the GPU than all the CPU cores associated with the same node count. In addition to performance, the maintainability of the GPU-accelerated version of the code was considered a key design objective. Accuracy with respect to the CPU implementation is also a core consideration. Finally, four different methods are used to ensure that the accelerated code faithfully reproduces the CPU results.

43 PARTICLE ACCELERATORS↗

Enabling Graph Mining in RDF Triplestores using SPARQL for Holistic In-situ Graph Analysis

The graph analysis is now considered as a promising technique to discover useful knowledge in data with a new perspective. We envi- sion that there are two dimensions of graph analysis: OnLine Graph Analytic Processing (OLGAP) and Graph Mining (GM) where each respectively focuses on subgraph pattern matching and automatic knowledge discovery in graph. Moreover, as these two dimensions aim to complementarily solve complex problems, holistic in-situ graph analysis which covers both OLGAP and GM in a single system is critical for minimizing the burdens of operating multiple graph systems and transferring intermediate result-sets between those systems. Nevertheless, most existing graph analysis systems are only capable of one dimension of graph analysis. In this work, we take an approach to enabling GM capabilities (e.g., PageRank, connected-component analysis, node eccentricity, etc.) in RDF triplestores, which are originally developed to store RDF datasets and provide OLGAP capability. More specifically, to achieve our goal, we implemented six representative graph mining algorithms using SPARQL. The approach allows a wide range of available RDF data sets directly applicable for holistic graph analysis within a system. For validation of our approach, we evaluate performance of our implementations with nine real-world datasets and three different computing environments - a laptop computer, an Amazon EC2 instance, and a shared-memory Cray XMT2 URIKA-GD graph-processing appliance. The experimen- tal results show that our implementation can provide promising and scalable performance for real world graph analysis in all tested environments. The developed software is publicly available in an open-source project that we initiated.

97 MATHEMATICS AND COMPUTING↗

Scaling Deep Learning workloads: NVIDIA DGX-1/Pascal and Intel Knights Landing

Deep Learning (DL) algorithms have become ubiquitous in data analytics. As a result, major computing vendors—including NVIDIA, Intel, AMD, and IBM—have architectural road maps influenced by DL workloads. Furthermore, several vendors have recently advertised new computing products as accelerating large DL workloads. Unfortunately, it is difficult for data scientists to quantify the potential of these different products. Here, this article provides a performance and power analysis of important DL workloads on two major parallel architectures: NVIDIA DGX-1 (eight Pascal P100 GPUs interconnected with NVLink) and Intel Knights Landing (KNL) CPUs interconnected with Intel Omni-Path or Cray Aries. Our evaluation consists of a cross section of convolutional neural net workloads: CifarNet, AlexNet, GoogLeNet, and ResNet50 topologies using the Cifar10 and ImageNet datasets. The workloads are vendor-optimized for each architecture. We use sequentially equivalent implementations to maintain iso-accuracy between parallel and sequential DL models. Our analysis indicates that although GPUs provide the highest overall performance, the gap can close for some convolutional networks; and the KNL can be competitive in performance/watt. We find that NVLink facilitates scaling efficiency on GPUs. However, its importance is heavily dependent on neural network architecture. Furthermore, for weak-scaling—sometimes encouraged by restricted GPU memory—NVLink is less important.

97 MATHEMATICS AND COMPUTING↗