Search NASA⌕ Search

SEARCH · Search NASA

Results for “OpenMP parallelization”

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.

299 records · Page 17

An MPI + $X$ implementation of contact global search using Kokkos

This paper describes an approach that seeks to parallelize the spatial search associated with computational contact mechanics. In contact mechanics, the purpose of the spatial search is to find “nearest neighbors,” which is the prelude to an imprinting search that resolves the interactions between the external surfaces of contacting bodies. In particular, we are interested in the contact global search portion of the spatial search associated with this operation on domain-decomposition-based meshes. Specifically, we describe an implementation that combines standard domain-decomposition-based MPI-parallel spatial search with thread-level parallelism (MPI-X) available on advanced computer architectures (those with GPU coprocessors). Our goal is to demonstrate the efficacy of the MPI-X paradigm in the overall contact search. Standard MPI-parallel implementations typically use a domain decomposition of the external surfaces of bodies within the domain in an attempt to efficiently distribute computational work. This decomposition may or may not be the same as the volume decomposition associated with the host physics. The parallel contact global search phase is then employed to find and distribute surface entities (nodes and faces) that are needed to compute contact constraints between entities owned by different MPI ranks without further inter-rank communication. Key steps of the contact global search include computing bounding boxes, building surface entity (node and face) search trees and finding and distributing entities required to complete on-rank (local) spatial searches. To enable source-code portability and performance across a variety of different computer architectures, we implemented the algorithm using the Kokkos hardware abstraction library. While we targeted development towards machines with a GPU accelerator per MPI rank, we also report performance results for OpenMP with a conventional multi-core compute node per rank. Results here demonstrate a 47 % decrease in the time spent within the global search algorithm, comparing the reference ACME algorithm with the GPU implementation, on an 18M face problem using four MPI ranks. As a result, while further work remains to maximize performance on the GPU, this result illustrates the potential of the proposed implementation.

97 MATHEMATICS AND COMPUTING↗

Parallel k-means++ for Multiple Shared-Memory Architectures

In recent years k-means++ has become a popular initialization technique for improved k-means clustering. To date, most of the work done to improve its performance has involved parallelizing algorithms that are only approximations of k-means++. In this paper we present a parallelization of the exact k-means++ algorithm, with a proof of its correctness. We develop implementations for three distinct shared-memory architectures: multicore CPU, high performance GPU, and the massively multithreaded Cray XMT platform. We demonstrate the scalability of the algorithm on each platform. In addition we present a visual approach for showing which platform performed k-means++ the fastest for varying data sizes.

data clustering↗

Optimizing the hypre solver for manycore and GPU architectures

The solution of large-scale combustion problems with codes such as Uintah on modern computer architectures requires the use of multithreading and GPUs to achieve performance. Uintah uses a low-Mach number approximation that requires iteratively solving a large system of linear equations. The Hypre iterative solver has solved such systems in a scalable way for Uintah, but the use of OpenMP with Hypre leads to at least slowdown due to OpenMP overheads. The proposed solution uses the MPI Endpoints within Hypre, where each team of threads acts as a different MPI rank. This approach minimizes OpenMP synchronization overhead and performs as fast or (up to 1.44) faster than Hypre's MPI-only version, and allows the rest of Uintah to be optimized using OpenMP. The profiling of the GPU version of Hypre shows the bottleneck to be the launch overhead of thousands of micro-kernels. The GPU performance was improved by fusing these micro-kernels and was further optimized by using Cuda-aware MPI, resulting in an overall speedup of 1.16—1.44 compared to the baseline GPU implementation. The above optimization strategies were published in the International Conference on Computational Science 2020 [1]. This work extends the previously published research by carrying out the second phase of communication-centered optimizations in Hypre to improve its scalability on large-scale supercomputers. Additionally, this includes an efficient non-blocking inter-thread communication scheme, communication-reducing patch assignment, and expression of logical communication parallelism to a new version of the MPICH library that utilizes the underlying network parallelism [2]. The above optimizations avoid communication bottlenecks previously observed during strong scaling and improve performance by up to 2 on 256 nodes of Intel Knight's Landing processor.

97 MATHEMATICS AND COMPUTING↗

Final Report: Correctness Tools for Petascale Computing

In the course of developing parallel programs for leadership computing systems, subtle programming errors often arise that are extremely difficult to diagnose without tools. To meet this challenge, University of Maryland, the University of Wisconsin—Madison, and Rice University worked to develop lightweight tools to help code developers pinpoint a variety of program correctness errors that plague parallel scientific codes. The aim of this project was to develop software tools that help diagnose program errors including memory leaks, memory access errors, round-off errors, and data races. Research at Rice University focused on developing algorithms and data structures to support efficient monitoring of multithreaded programs for memory access errors and data races. This is a final report about research and development work at Rice University as part of this project.

97 MATHEMATICS AND COMPUTING↗

Developing Ultrahigh-Resolution E3SM Land Model for GPU Systems

Designing and refactoring complex scientific code, such as the E3SM land model (ELM), for new computing architectures is challenging. This paper presents design strategies and technical approaches to develop a data-oriented, GPU-ready ELM model using compiler directives (OpenACC/OpenMP). We first analyze the datatypes and processes in the original ELM code. Then we present design considerations for ultrahigh-resolution ELM (uELM) development for massive GPU systems. These techniques include the global data-oriented simulation workflow, domain partition, code porting and data copy, memory reduction, parallel loop restructure and flattening, and race condition detection. We implemented the first version of uELM using OpenACC targeting the NVidia GPUs in the Summit supercomputer at Oak Ridge National Laboratory. During the implementation, we developed a software tool (named SPEL) to facilitate code generation, verification, and performance tuning using these techniques. The first uELM implementation for Nvidia GPUs on Summit delivered promising results: 1) over 98% of the ELM code was automatically generated and tuned by scripts. Most ELM modules had better computational performances than the original ELM code for CPUs. The GPU-ready uELM is more scalable than the CPU code on fully-loaded Summit nodes. Example profiling results from several modules are also presented to illustrate the performance improvements and race condition detection. The lessons learned and toolkit developed in the study are also suitable for further uELM deployment using OpenMP on the first US exascale computer, Frontier, equipped with AMD CPUs and GPUs.

Schwartz, Peter↗

Implementing Arbitrary/Common Concurrent Writes of CRCW PRAM

The Parallel Random Access Machines (PRAM) abstraction is the simplest and most elegant algorithmic model for the design and analysis of parallel algorithms. It consists of different models categorized based on the underlying memory access mode used, the most powerful of which is the Concurrent Read Concurrent Write (CRCW) model. A PRAM algorithm describes a series of rounds, each of which consists of a collection of operations that can be executed concurrently within the same time step. However, the lack of support for concurrent memory accesses and the prevalence of asynchronous programming models led to the belief that implementing CRCW PRAM algorithms is unattainable and prompted many to avoid this model except for theoretical studies of optimal performance.In this work, we study the arbitrary and common concurrent writes in the CRCW PRAM model and explore implementation challenges on general-purpose systems. Moreover, we examine current practices for implementing common/arbitrary concurrent writes and propose a new efficient lightweight and thread-safe method to implement concurrent writes through leveraging atomic instructions. To demonstrate the efficacy of our method, we developed OpenMP kernels for classical CRCW PRAM algorithms and provide experimental results and comparisons based on run time performance measured over the x86 multicore architecture. Our results show a performance speedup compared to current practices up to 4.5x across all our benchmarks.

Ghanim, Fady↗

Modeling of a chain of three plasma accelerator stages with the WarpX electromagnetic PIC code on GPUs

The fully electromagnetic particle-in-cell code WarpX is being developed by a team of the U.S. DOE Exascale Computing Project (with additional non-U.S. collaborators on part of the code) to enable the modeling of chains of tens to hundreds of plasma accelerator stages on exascale supercomputers, for future collider designs. The code is combining the latest algorithmic advances (e.g., Lorentz boosted frame and pseudo-spectral Maxwell solvers) with mesh refinement and runs on the latest computer processing unit and graphical processing unit (GPU) architectures. In this paper, we summarize the strategy that was adopted to port WarpX to GPUs, report on the weak parallel scaling of the pseudo-spectral electromagnetic solver, and then present solutions for decreasing the time spent in data exchanges from guard regions between subdomains. In Sec. IV, we demonstrate the simulations of a chain of three consecutive multi-GeV laser-driven plasma accelerator stages.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Experiences with SYCL on AMD GPUs with Kokkos

With the recent diversification of the hardware landscape in the high-performance computing (HPC) community, performance-portability solutions are becoming more and more important. One of the most popular choices is Kokkos, which recently became a Linux Foundation project. Most of its development is supported by the US Department of Energy and the French Alternative Energies and Atomic Energy Commission. Kokkos is implemented as a C++ library with multiple backends to support CPUs as well as various GPU architectures. These backends include OpenMP, CUDA, HIP, and also SCYL. This approach enables users to leverage the preferred vendor toolchain for the respective platform (e.g. CUDA, ROCm, OneAPI). The SYCL backend is used to target Intel GPUs, in particular to support the Aurora exascale supercomputer. However, SYCL itself also offers a large degree of portability, and in fact Kokkos’ CI for SYCL has been running on NVIDIA hardware due to a lack of access to Intel GPUs. In this report, we describe our experience with using Kokkos SYCL backend on AMD GPUs targeting the Frontier supercomputer at Oak Ridge National Laboratory. The two major SYCL implementations are DPC++ and AdaptiveCpp. While the Kokkos SYCL backend has been implemented using the former, the latter was the first implementation to target AMD GPUs. We will discuss the experience with both of these SYCL implementations in terms of functionality and performance. Using Kokkos to evaluate SYCL toolchains has a number of benefits. Kokkos’ use of SYCL is fairly complex, exercising features such as graphs, relocatable device functions, atomics – including for non-arithmetic types, as well as pinned and page migratable memory allocations. Kokkos also needs to implement capabilities such as Kokkos’ hierarchical parallelism that are not a straight-forward mapping to SYCL capabilities. Furthermore, a large number of libraries and applications that represent diverse use cases are implemented in Kokkos, providing readily available test cases for a toolchain evaluation. Preliminary results show that support for AMD GPUs in DPC++ is much less mature than for NVIDIA GPUs or Intel GPUs. While the situation has improved significantly over the last year, we still encounter many runtime failures, dispatching problems, and code generation issues. With AdaptiveCpp the challenges arise even earlier in the evaluation process. Since Kokkos’ SYCL implementation is largely focused on supporting Intel GPUs, we opted to leverage SYCL extensions which are available in DPC++ but not in AdaptiveCpp. Furthermore, AdaptiveCpp appears to be less conformant with the SYCL2020 standard which Kokkos relies on. In some cases, we are able to work around the lack of feature support, in other cases we have to disable certain Kokkos capabilities to evaluate the toolchain. Our evaluation will leverage Kokkos’ unit tests to establish basic functionality and feature completeness. We then use simple benchmarks for components of a CG implementation as a measure of usability and performance of the SYCL toolchains.

97 MATHEMATICS AND COMPUTING↗

Exploring performance and energy tradeoffs for irregular applications: A case study on the Tilera many-core architecture

High performance, parallel applications with irregular data accesses are becoming a critical workload class for modern systems. In particular, the execution of such workloads on emerging many-core systems is expected to be a significant component of applications in data mining, machine learning, scientific computing and graph analytics. However, power and energy constraints limit the capabilities of individual cores, memory hierarchy and on-chip interconnect of such systems, thus leading to architectural and software trade-os that must be understood in the context of the intended application’s behavior. Irregular applications are notoriously hard to optimize given their data-dependent access patterns, lack of structured locality and complex data structures and code patterns. We have ported two irregular applications, graph community detection using the Louvain method (Grappolo) and high-performance conjugate gradient (HPCCG), to the Tilera many-core system and have conducted a detailed study of platform-independent and platform-specific optimizations that improve their performance as well as reduce their overall energy consumption. To conduct this study, we employ an auto-tuning based approach that explores the optimization design space along three dimensions - memory layout schemes, GCC compiler flag choices and OpenMP loop scheduling options. We leverage MIT’s OpenTuner auto-tuning framework to explore and recommend energy optimal choices for different combinations of parameters. We then conduct an in-depth architectural characterization to understand the memory behavior of the selected workloads. Finally, we perform a correlation study to demonstrate the interplay between the hardware behavior and application characteristics. Using auto-tuning, we demonstrate whole-node energy savings and performance improvements of up to 49:6% and 60% relative to a baseline instantiation, and up to 31% and 45:4% relative to manually optimized variants.

auto tuning↗

Porting fragmentation methods to GPUs using an OpenMP API: Offloading the resolution-of-the-identity second-order Møller–Plesset perturbation method

Here, using an OpenMP Application Programming Interface, the resolution-of-the-identity second-order Møller–Plesset perturbation (RI-MP2) method has been off-loaded onto graphical processing units (GPUs), both as a standalone method in the GAMESS electronic structure program and as an electron correlation energy component in the effective fragment molecular orbital (EFMO) framework. First, a new scheme has been proposed to maximize data digestion on GPUs that subsequently linearizes data transfer from central processing units (CPUs) to GPUs. Second, the GAMESS Fortran code has been interfaced with GPU numerical libraries (e.g., NVIDIA cuBLAS and cuSOLVER) for efficient matrix operations (e.g., matrix multiplication, matrix decomposition, and matrix inversion). The standalone GPU RI-MP2 code shows an increasing speedup of up to 7.5× using one NVIDIA V100 GPU with one IBM 42-core P9 CPU for calculations on fullerenes of increasing size from 40 to 260 carbon atoms using the 6-31G(d)/cc-pVDZ-RI basis sets. A single Summit node with six V100s can compute the RI-MP2 correlation energy of a cluster of 175 water molecules using the correlation consistent basis sets cc-pVDZ/cc-pVDZ-RI containing 4375 atomic orbitals and 14 700 auxiliary basis functions in ~0.85 h. In the EFMO framework, the GPU RI-MP2 component shows near linear scaling for a large number of V100s when computing the energy of an 1800-atom mesoporous silica nanoparticle in a bath of 4000 water molecules. The parallel efficiencies of the GPU RI-MP2 component with 2304 and 4608 V100s are 98.0% and 96.1%, respectively.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Evaluation of Low Power ARM Processors for Monte Carlo Particle Transport: MCNP6 on the ARM Cortex-A8 - Paper 83

In the push towards exascale computing, the limiting factors are power requirements and heat dissipation. Using current technology, the power required to drive an exascale supercomputer would near 400-MW. Clearly, future computing hardware will have to be more energy efficient in order to achieve exascale performance. Most of the current research in advanced computing architectures is in the use of standard CPUs with accelerators. However, these accelerators are not generally supported by the traditional scientific programming languages (C, C++, and Fortran) and the scientific software must be tailored specifically to take advantage of the accelerators. Large scale scientific software (hundreds-of-thousand to millions of lines of code) can not be easily ported to the computing platform du jour. The widespread use of low power ARM based processors in the mobile computing market offer a low power alternative to accelerators that are also supported by the traditional scientific programming languages and programming paradigms such as MPI and OpenMP. Computing architectures for scientific computing are often assessed using standard benchmarks such as LINPACK benchmark or using suites of mini-applications that are greatly simplified stand-ins for scientific software. Often the real applications that will deployed on the scientific supercomputers are not ported or tested until delivery of the machines. In order to evaluate the use of ARM processors for scientific computing the Monte Carlo particle transport code, MCNP version 6 was compiled on the ARM Cortex-A8 and a set of benchmark problems were used to assess performance in comparison to an Intel Xeon E5-2670 and the low power Intel Atom N270. MCNP is a continuous energy Monte Carlo particle transport code in development at Los Alamos National Laboratory since 1977. MCNP is used to transport neutrons, photon, electrons, and charged particles using continuous energy interaction data. It is commonly used in fields of nuclear power, nuclear medicine, radiation protection, and high-energy physics through-out the world. MCNP version 6 is approximately 450 thousand lines of code, written primary in Fortran90, with some C and C++. MCNP supports distributed multiprocessing using MPI and threading with OpenMP. The ARM Cortex-A8 executes MCNP6 using about twice as much power as the Intel Xeon E5-2670 and about the same amount of power as the Intel Atom N270. However, these results are encouraging for the use of low-power ARM chips in future supercomputers. The ARM Cortex-A8 is a 32-bit CPU and it's NEON general-purpose SIMD (single-instruction multiple data) engine does not support double-precision floating-point arithmetic. More recent ARM processors (ARM Cortex-A53 and Cortex-A57) support 64-bit addressing and have NEON co-processors that support double-precision floating-point arithmetic. ARM claims that the A50 series delivers a performance increase of 15 times the Cortex-A8 performance for similar power usage. (authors)

61 RADIATION PROTECTION AND DOSIMETRY↗