Search NASA⌕ Search

SEARCH · Search NASA

Results for “distributed memory parallel systems”

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 361 records · Page 20

LDRD final report : managing shared memory data distribution in hybrid HPC applications.

MPI is the dominant programming model for distributed memory parallel computers, and is often used as the intra-node programming model on multi-core compute nodes. However, application developers are increasingly turning to hybrid models that use threading within a node and MPI between nodes. In contrast to MPI, most current threaded models do not require application developers to deal explicitly with data locality. With increasing core counts and deeper NUMA hierarchies seen in the upcoming LANL/SNL 'Cielo' capability supercomputer, data distribution poses an upper boundary on intra-node scalability within threaded applications. Data locality therefore has to be identified at runtime using static memory allocation policies such as first-touch or next-touch, or specified by the application user at launch time. We evaluate several existing techniques for managing data distribution using micro-benchmarks on an AMD 'Magny-Cours' system with 24 cores among 4 NUMA domains and argue for the adoption of a dynamic runtime system implemented at the kernel level, employing a novel page table replication scheme to gather per-NUMA domain memory access traces.

99 GENERAL AND MISCELLANEOUS↗

CAST: Contraction Algorithm for Symmetric Tensors

Tensor contractions represent the most compute-intensive core kernels in ab initio computational quantum chemistry and nuclear physics. Symmetries in these tensor contractions makes them difficult to load balance and scale to large distributed systems. In this paper, we develop an efficient and scalable algorithm to contract symmetric tensors. We introduce a novel approach that avoids data redistribution in contracting symmetric tensors while also avoiding redundant storage and maintaining load balance. We present experimental results on two parallel supercomputers for several symmetric contractions that appear in the CCSD quantum chemistry method. We also present a novel approach to tensor redistribution that can take advantage of parallel hyperplanes when the initial distribution has replicated dimensions, and use collective broadcast when the final distribution has replicated dimensions, making the algorithm very efficient.

distributed memory↗

Compile-time estimation of communication costs in multicomputers

An important problem facing numerous research projects on parallelizing compilers for distributed memory machines is that of automatically determining a suitable data partitioning scheme for a program. Any strategy for automatic data partitioning needs a mechanism for estimating the performance of a program under a given partitioning scheme, the most crucial part of which involves determining the communication costs incurred by the program. A methodology is described for estimating the communication costs at compile-time as functions of the numbers of processors over which various arrays are distributed. A strategy is described along with its theoretical basis, for making program transformations that expose opportunities for combining of messages, leading to considerable savings in the communication costs. For certain loops with regular dependences, the compiler can detect the possibility of pipelining, and thus estimate communication costs more accurately than it could otherwise. These results are of great significance to any parallelization system supporting numeric applications on multicomputers. In particular, they lay down a framework for effective synthesis of communication on multicomputers from sequential program references.

Gupta, Manish↗

SOMAFOAM: An OpenFOAM based solver for continuum simulations of low-temperature plasmas

Here, we report the development of SOMAFOAM, a finite volume framework for performing continuum simulations of low-temperature plasmas. The primary goal of this work is to discuss the features of SOMAFOAM along with representative results provided as examples for a range of operating conditions and geometries. This includes plasma and plasma–dielectric systems operating in direct current, radio frequency, and microwave regimes from pressures as low as 100 mTorr to atmospheric pressure. The code has several useful features including the ability to run massively parallel simulations using arbitrary geometries, structured/unstructured meshes, choice of models such as drift–diffusion/full-momentum at runtime, and species-dependent timesteps to name a few. The verification/validation studies presented include comparison with previously published continuum simulations (low-pressure direct current and radio frequency plasma), with experiments (Gaseous Electronics Conference Reference Cell and microwave microplasma ignited in a split ring resonator), and previously published kinetic simulations (low-pressure radio frequency plasma). Other examples provided include a direct current atmospheric pressure microplasma bounded by dielectric sidewalls and a helium–nitrogen plasma ignited using a needle electrode facing a dielectric. The performance of the code is also discussed with serial and distributed memory parallel runs demonstrated up to 512 cores. The design and implementation of the code in a modular object-oriented framework allows for easy extension and seamless coupling with other codes and can be expected to play an important role in both academia and industry.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Production Application Performance Data Streaming for System Monitoring

In this article, we present an approach to streaming collection of application performance data. Practical application performance tuning and troubleshooting in production high-performance computing (HPC) environments requires an understanding of how applications interact with the platform, including (but not limited to) parallel programming libraries such as Message Passing Interface (MPI). Several profiling and tracing tools exist that collect heavy runtime data traces either in memory (released only at application exit) or on a file system (imposing an I/O load that may interfere with the performance being measured). Although these approaches are beneficial in development stages and post-run analysis, a systemwide and low-overhead method is required to monitor deployed applications continuously. This method must be able to collect information at both the application and system levels to yield a complete performance picture. In our approach, an application profiler collects application event counters. A sampler uses an efficient inter-process communication method to periodically extract the application counters and stream them into an infrastructure for performance data collection. We implement a tool-set based on our approach and integrate it with the Lightweight Distributed Metric Service (LDMS) system, a monitoring system used on large-scale computational platforms. LDMS provides the infrastructure to create and gather streams of performance data in a low overhead manner. We demonstrate our approach using applications implemented with MPI, as it is one of the most common standards for the development of large-scale scientific applications. We utilize our tool-set to study the impact of our approach on an open source HPC application, Nalu. Our tool-set enables us to efficiently identify patterns in the behavior of the application without source-level knowledge. We leverage LDMS to collect system-level performance data and explore the correlation between the system and application events. Finally, we demonstrate how our tool-set can help detect anomalies with a low latency. We run tests on two different architectures: a system enabled with Intel Xeon Phi and another system equipped with Intel Xeon processor. Our overhead study shows our method imposes at most 0.5% CPU usage overhead on the application in realistic deployment scenarios.

59 BASIC BIOLOGICAL SCIENCES↗

Application of Local Discretization Methods in the NASA Finite-Volume General Circulation Model

We present the basic ideas of the dynamics system of the finite-volume General Circulation Model developed at NASA Goddard Space Flight Center for climate simulations and other applications in meteorology. The dynamics of this model is designed with emphases on conservative and monotonic transport, where the property of Lagrangian conservation is used to maintain the physical consistency of the computational fluid for long-term simulations. As the model benefits from the noise-free solutions of monotonic finite-volume transport schemes, the property of Lagrangian conservation also partly compensates the accuracy of transport for the diffusion effects due to the treatment of monotonicity. By faithfully maintaining the fundamental laws of physics during the computation, this model is able to achieve sufficient accuracy for the global consistency of climate processes. Because the computing algorithms are based on local memory, this model has the advantage of efficiency in parallel computation with distributed memory. Further research is yet desirable to reduce the diffusion effects of monotonic transport for better accuracy, and to mitigate the limitation due to fast-moving gravity waves for better efficiency.

Yeh, Kao-San↗

Tula: Optimizing Time, Cost, and Generalization in Distributed Large-Batch Training

Distributed training increases the number of batches processed per iteration either by scaling-out (adding more nodes) or scaling-up (increasing the batch-size). However, the largest configuration does not necessarily yield the best performance. Horizontal scaling introduces additional communication overhead, while vertical scaling is constrained by computation cost and device memory limits. Thus, simply increasing the batch-size leads to diminishing returns: training time and cost decrease initially but eventually plateaus, creating a knee-point in the time/cost vs. batch-size pareto curve. The optimal batch-size therefore depends on the underlying model, data and available compute resources. Large batches also suffer from worse model quality due to the well-known “generalization gap”. In this paper, we present Tula, an online service that automatically optimizes time, cost, and convergence quality for large-batch training of convolutional models. It combines parallel-systems modeling with statistical performance prediction to identify the optimal batchsize. Tula predicts training time and cost within 7.5−14% error across multiple models, and achieves up to 20× overall speedup and improves test accuracy by ≈9% on average over standard large-batch training on various vision tasks, thus successfully mitigating the generalization gap and accelerating training at the same time.

Tyagi, Sahil [ORNL] (ORCID:0009000783144745)↗

Global Arrays Parallel Programming Toolkit

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

ARMCI↗

Virtual Machine Language 2.1

VML (Virtual Machine Language) is an advanced computing environment that allows spacecraft to operate using mechanisms ranging from simple, time-oriented sequencing to advanced, multicomponent reactive systems. VML has developed in four evolutionary stages. VML 0 is a core execution capability providing multi-threaded command execution, integer data types, and rudimentary branching. VML 1 added named parameterized procedures, extensive polymorphism, data typing, branching, looping issuance of commands using run-time parameters, and named global variables. VML 2 added for loops, data verification, telemetry reaction, and an open flight adaptation architecture. VML 2.1 contains major advances in control flow capabilities for executable state machines. On the resource requirements front, VML 2.1 features a reduced memory footprint in order to fit more capability into modestly sized flight processors, and endian-neutral data access for compatibility with Intel little-endian processors. Sequence packaging has been improved with object-oriented programming constructs and the use of implicit (rather than explicit) time tags on statements. Sequence event detection has been significantly enhanced with multi-variable waiting, which allows a sequence to detect and react to conditions defined by complex expressions with multiple global variables. This multi-variable waiting serves as the basis for implementing parallel rule checking, which in turn, makes possible executable state machines. The new state machine feature in VML 2.1 allows the creation of sophisticated autonomous reactive systems without the need to develop expensive flight software. Users specify named states and transitions, along with the truth conditions required, before taking transitions. Transitions with the same signal name allow separate state machines to coordinate actions: the conditions distributed across all state machines necessary to arm a particular signal are evaluated, and once found true, that signal is raised. The selected signal then causes all identically named transitions in all present state machines to be taken simultaneously. VML 2.1 has relevance to all potential space missions, both manned and unmanned. It was under consideration for use on Orion.

Riedel, Joseph E.↗

Ad Hoc File Systems for High-Performance Computing

Storage backends of parallel compute clusters are still based mostly on magnetic disks, while newer and faster storage technologies such as flash-based SSDs or non-volatile random access memory (NVRAM) are deployed within compute nodes. Including these new storage technologies into scientific workflows is unfortunately today a mostly manual task, and most scientists therefore do not take advantage of the faster storage media. One approach to systematically include nodelocal SSDs or NVRAMs into scientific workflows is to deploy ad hoc file systems over a set of compute nodes, which serve as temporary storage systems for single applications or longer-running campaigns. This paper presents results from the Dagstuhl Seminar 17202 “Challenges and Opportunities of User-Level File Systems for HPC” and discusses application scenarios as well as design strategies for ad hoc file systems using node-local storage media. The discussion includes open research questions, such as how to couple ad hoc file systems with the batch scheduling environment and how to schedule stage-in and stage-out processes of data between the storage backend and the ad hoc file systems. Also presented are strategies to build ad hoc file systems by using reusable components for networking and how to improve storage device compatibility. Various interfaces and semantics are presented, for example those used by the three ad hoc file systems BeeOND, GekkoFS, and BurstFS. Their presentation covers a range from file systems running in production to cutting-edge research focusing on reaching the performance limits of the underlying devices.

97 MATHEMATICS AND COMPUTING↗

Predeployment validation of fault-tolerant systems through software-implemented fault insertion

Fault injection-based automated testing (FIAT) environment, which can be used to experimentally characterize and evaluate distributed realtime systems under fault-free and faulted conditions is described. A survey is presented of validation methodologies. The need for fault insertion based on validation methodologies is demonstrated. The origins and models of faults, and motivation for the FIAT concept are reviewed. FIAT employs a validation methodology which builds confidence in the system through first providing a baseline of fault-free performance data and then characterizing the behavior of the system with faults present. Fault insertion is accomplished through software and allows faults or the manifestation of faults to be inserted by either seeding faults into memory or triggering error detection mechanisms. FIAT is capable of emulating a variety of fault-tolerant strategies and architectures, can monitor system activity, and can automatically orchestrate experiments involving insertion of faults. There is a common system interface which allows ease of use to decrease experiment development and run time. Fault models chosen for experiments on FIAT have generated system responses which parallel those observed in real systems under faulty conditions. These capabilities are shown by two example experiments each using a different fault-tolerance strategy.

Czeck, Edward W.↗

Optimizing checkpoint data placement with guaranteed burst buffer endurance in large-scale hierarchical storage systems

Non-volatile devices, such as SSDs, will be an integral part of the deepening storage hierarchy on large-scale HPC systems. These devices can be on the compute nodes as part of a distributed burst buffer service or they can be external. Wherever they are located in the hierarchy, one critical design issue is the SSD endurance under the write-heavy workloads, such as the checkpoint I/O for scientific applications. For these environments, it is widely assumed that checkpoint operations can occur once every 60 min and for each checkpoint step as much as half of the system memory can be written out. Unfortunately, for large-scale HPC applications, the burst buffer SSDs can be worn out much more quickly given the extensive amount of data written at every checkpoint step. One possible solution is to control the amount of data written by reducing the checkpoint frequency. However, a direct effect caused by reduced checkpoint frequency is the increased vulnerability window of system failures and therefore potentially wasted computation time, especially for large-scale compute jobs.

Computer Science↗

A parallel strategy for density functional theory computations on accelerated nodes

Using the Löwdin orthonormalization of tall-skinny matrices as a proxy-app for wavefunction-based Density Functional Theory solvers, we investigate a distributed memory parallel strategy focusing on Graphics Processing Unit (GPU)-accelerated nodes as available on some of the top ranked supercomputers at the present time. Here we present numerical results in the strong limit regime, as it is particularly relevant for First-Principles Molecular Dynamics. We also examine how matrix product-based iterative solvers provide a competitive alternative to dense eigensolvers on GPUs, allowing to push the strong scaling limit of these computations to a larger number of distributed tasks. Our strategy, which relies on replicated Gram matrices and efficient collective communications using the NCCL library, leads to a time-to-solution under 0.5 s for the Löwdin orthonormalization of a tall-skinny matrix of 3000 columns on Summit at Oak Ridge Leadership Facility (OLCF). Given the similarity in computational operations between one iteration of a DFT solver and this proxy-app, this shows the possibility of solving accurately the DFT equations well under a minute for 3000 electronic wave functions, and thus perform First-Principles molecular dynamics of physical systems much larger than traditionally solved on CPU systems.

97 MATHEMATICS AND COMPUTING↗

An Asynchronous Task-based Fan-Both Sparse Cholesky Solver

Systems of linear equations arise at the heart of many scientific and engineering applications. Many of these linear systems are sparse; i.e., most of the elements in the coefficient matrix are zero. Direct methods based on matrix factorizations are sometimes needed to ensure accurate solutions. For example, accurate solution of sparse linear systems is needed in shift-invert Lanczos to compute interior eigenvalues. The performance and resource usage of sparse matrix factorizations are critical to time-to-solution and maximum problem size solvable on a given platform. In many applications, the coefficient matrices are symmetric, and exploiting symmetry will reduce both the amount of work and storage cost required for factorization. When the factorization is performed on large-scale distributed memory platforms, communication cost is critical to the performance of the algorithm. At the same time, network topologies have become increasingly complex, so that modern platforms exhibit a high level of performance variability. This makes scheduling of computations an intricate and performance-critical task. In this paper, we investigate the use of an asynchronous task paradigm, one-sided communication and dynamic scheduling in implementing sparse Cholesky factorization (symPACK) on large-scale distributed memory platforms. Our solver symPACK relies on efficient and flexible communication primitives provided by the UPC++ library. Performance evaluation shows good scalability and that symPACK outperforms state-of-the-art parallel distributed memory factorization packages, validating our approach on practical cases.

Jacquelin, Mathias↗

Scalable, In-situ Data Clustering Data Analysis for Extreme Scale Scientific Computing (Final Report)

The objective of this project is to address challenges in the design and development of scalable in-situ data clustering and analytics algorithms and software. Our goal is to develop parallel software consisting of a set of spatio-temporal data clustering and anomaly detection functions, both of which are very important for large-scale analysis and have wide applicability for in-situ runs as well as post-processing analysis. Our design principles for in-situ analysis consider the following: (1) identify parts of the computation can be done close to the data within the nodes, while it is still in memory; (2) extract analysis components can (and should) be performed in remote staging and analysis nodes; (3) develop error-bound approximation methods for applications tolerable for small errors; (4) identify the type of derived distributions and statistics, for spatio-temporal data, that can be kept locally in order to both accelerate computations and meet energy constraints in subsequent iterations and phases; (5) use a self-describing data format so that data can be consistent and understood among local storage (memory and SSDs) and at staging and analysis nodes, thereby providing portability and flexibility; (6) develop service-oriented functions that can schedule in-situ and post-hoc analysis tasks based on the dynamic requirements of applications. Our development focus is to produce the parallel data analysis software/library that will be scalable, reusable, extensible, and generic for applications in different disciplines. The software will be able to run in-situ with the simulations as well as post-hoc analysis. This approach will satisfy many synergistic requirements for data intensive applications executed on data coming from instruments and experiments. In particular, the proposed multilevel approach is directly applicable to perform design tradeoffs for running part of the algorithms near the instruments and the rest on remote (analysis) systems.

97 MATHEMATICS AND COMPUTING↗

NVMalloc: Exposing an Aggregate SSD Store as a Memory Partition in Extreme-Scale Machines

DRAM is a precious resource in extreme-scale machines and is increasingly becoming scarce, mainly due to the growing number of cores per node. On future multi-petaflop and exaflop machines, the memory pressure is likely to be so severe that we need to rethink our memory usage models. Fortunately, the advent of non-volatile memory (NVM) offers a unique opportunity in this space. Current NVM offerings possess several desirable properties, such as low cost and power efficiency, but suffer from high latency and lifetime issues. We need rich techniques to be able to use them alongside DRAM. In this talk, we present a novel approach for exploiting NVM as a secondary memory partition so that applications can explicitly allocate and manipulate memory regions therein. More specifically, we propose an NVMalloc library with a suite of services that enables applications to access a distributed NVM storage system. We have devised ways within NVMalloc so that the storage system, built from compute node-local NVM devices, can be accessed in a byte-addressable fashion using the memory mapped I/O interface. Our approach has the potential to re-energize out-of-core computations on largescale machines by having applications allocate certain variables through NVMalloc, thereby increasing the overall memory capacity available. Our evaluation on a 128-core cluster shows that NVMalloc enables applications to compute problem sizes larger than the physical memory in a cost-effective manner. It can bring more performance/efficiency gain with increased computation time between NVM memory accesses or increased data access locality. In addition, our results suggest that while NVMalloc enables transparent access to NVM-resident variables, the explicit control it provides is crucial to optimize application performance.

99 GENERAL AND MISCELLANEOUS↗

SPARC: Demonstrate burst-buffer-based checkpoint/restart on ATS-1.

Recent high-performance computing (HPC) platforms such as the Trinity Advanced Technology System (ATS-1) feature burst buffer resources that can have a dramatic impact on an application’s I/O performance. While these non-volatile memory (NVM) resources provide a new tier in the storage hierarchy, developers must find the right way to incorporate the technology into their applications in order to reap the benefits. Similar to other laboratories, Sandia is actively investigating ways in which these resources can be incorporated into our existing libraries and workflows without burdening our application developers with excessive, platform-specific details. This FY18Q1 milestone summaries our progress in adapting the Sandia Parallel Aerodynamics and Reentry Code (SPARC) in Sandia’s ATDM program to leverage Trinity’s burst buffers for checkpoint/restart operations. We investigated four different approaches with varying tradeoffs in this work: (1) simply updating job script to use stage-in/stage out burst buffer directives, (2) modifying SPARC to use LANL’s hierarchical I/O (HIO) library to store/retrieve checkpoints, (3) updating Sandia’s IOSS library to incorporate the burst buffer in all meshing I/O operations, and (4) modifying SPARC to use our Kelpie distributed memory library to store/retrieve checkpoints. Team members were successful in generating initial implementation for all four approaches, but were unable to obtain performance numbers in time for this report (reasons: initial problem sizes were not large enough to stress I/O, and SPARC refactor will require changes to our code). When we presented our work to the SPARC team, they expressed the most interest in the second and third approaches. The HIO work was favored because it is lightweight, unobtrusive, and should be portable to ATS-2. The IOSS work is seen as a long-term solution, and is favored because all I/O work (including checkpoints) can be deferred to a single library.

97 MATHEMATICS AND COMPUTING↗

Communication-Avoiding and Memory-Constrained Sparse Matrix-Matrix Multiplication at Extreme Scale

Sparse matrix-matrix multiplication (SpGEMM) is a widely used kernel in various graph, scientific computing and machine learning algorithms. In this paper, we consider SpGEMMs performed on hundreds of thousands of processors generating trillions of nonzeros in the output matrix. Distributed SpGEMM at this extreme scale faces two key challenges: (1) high communication cost and (2) inadequate memory to generate the output. Furthermore, we address these challenges with an integrated communication-avoiding and memory-constrained SpGEMM algorithm that scales to 262,144 cores (more than 1 million hardware threads) and can multiply sparse matrices of any size as long as inputs and a fraction of output fit in the aggregated memory. As we go from 16,384 cores to 262,144 cores on a Cray XC40 supercomputer, the new SpGEMM algorithm runs 10x faster when multiplying large-scale protein-similarity matrices.

97 MATHEMATICS AND COMPUTING↗