Search NASA⌕ Search

SEARCH · Search NASA

Results for “Memory Hierarchy”

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 127 records · Page 7

Accelerating Random Forest Classification on GPU and FPGA

Random Forests (RFs) are a commonly used machine learning method for classification and regression tasks spanning a variety of application domains, including bioinformatics, business analytics, and software optimization. While prior work has focused primarily on improving performance of the training of RFs, many applications, such as malware identification, cancer prediction, and banking fraud detection, require fast RF classification. In this work, we accelerate RF classification on GPU and FPGA. In order to provide efficient support for large datasets, we propose a hierarchical memory layout suitable to the GPU/FPGA memory hierarchy. We design three RF classification code variants based on that layout, and we investigate GPU- and FPGA-specific considerations for these kernels. Our experimental evaluation, performed on an Nvidia Xp GPU and on a Xilinx Alveo U250 FPGA accelerator card using publicly available datasets on the scale of millions of samples and tens of features, covers various aspects. First, we evaluate the performance benefits of our hierarchical data structure over the standard compressed sparse row (CSR) format. Second, we compare our GPU implementation with cuML, a machine learning library targeting Nvidia GPUs. Third, we explore the performance/accuracy tradeoff resulting from the use of different tree depths in the RF. Finally, we perform a comparative performance analysis of our GPU and FPGA implementations. Our evaluation shows that for high accuracy targets, our GPU implementation yields 5-9x speedup over CSR, and up to a 2x speedup over cuML.

FPGA, Xilinx FPGA, GPU, Random Forest classificati↗

Extensions to the SENSEI In situ Framework for Heterogeneous Architectures

The proliferation of GPUs and accelerators in recent supercomputing systems, so called heterogeneous architectures, has led to increased complexity in execution environments and programming models as well as to deeper memory hierarchies on these systems. In this work, we discuss challenges that arise in in situ code coupling on these heterogeneous architectures. In particular, we present data and execution model extensions to the SENSEI in situ framework that are targeted at the effective use of systems with heterogeneous architectures. We then use these new data and execution model extensions to investigate several in situ placement and execution configurations and to analyze the impact these choices have on overall performance.

Loring, Burlen↗

Performance Models for the Spike Banded Linear System Solver

With availability of large-scale parallel platforms comprised of tens-of-thousands of processors and beyond, there is significant impetus for the development of scalable parallel sparse linear system solvers and preconditioners. An integral part of this design process is the development of performance models capable of predicting performance and providing accurate cost models for the solvers and preconditioners. There has been some work in the past on characterizing performance of the iterative solvers themselves. In this paper, we investigate the problem of characterizing performance and scalability of banded preconditioners. Recent work has demonstrated the superior convergence properties and robustness of banded preconditioners, compared to state-of-the-art ILU family of preconditioners as well as algebraic multigrid preconditioners. Furthermore, when used in conjunction with efficient banded solvers, banded preconditioners are capable of significantly faster time-to-solution. Our banded solver, the Truncated Spike algorithm is specifically designed for parallel performance and tolerance to deep memory hierarchies. Its regular structure is also highly amenable to accurate performance characterization. Using these characteristics, we derive the following results in this paper: (i) we develop parallel formulations of the Truncated Spike solver, (ii) we develop a highly accurate pseudo-analytical parallel performance model for our solver, (iii) we show excellent predication capabilities of our model – based on which we argue the high scalability of our solver. Our pseudo-analytical performance model is based on analytical performance characterization of each phase of our solver. These analytical models are then parameterized using actual runtime information on target platforms. An important consequence of our performance models is that they reveal underlying performance bottlenecks in both serial and parallel formulations. All of our results are validated on diverse heterogeneous multiclusters – platforms for which performance prediction is particularly challenging. Finally, we provide predict the scalability of the Spike algorithm using up to 65,536 cores with our model. In this paper we extend the results presented in the Ninth International Symposium on Parallel and Distributed Computing.

Manguoglu, Murat↗

Multi-core and many-core shared-memory parallel raycasting volume rendering optimization and tuning

Given the computing industry trend of increasing processing capacity by adding more cores to a chip, the focus of this work is tuning the performance of a staple visualization algorithm, raycasting volume rendering, for shared-memory parallelism on multi-core CPUs and many-core GPUs. Our approach is to vary tunable algorithmic settings, along with known algorithmic optimizations and two different memory layouts, and measure performance in terms of absolute runtime and L2 memory cache misses. Our results indicate there is a wide variation in runtime performance on all platforms, as much as 254% for the tunable parameters we test on multi-core CPUs and 265% on many-core GPUs, and the optimal configurations vary across platforms, often in a non-obvious way. For example, our results indicate the optimal configurations on the GPU occur at a crossover point between those that maintain good cache utilization and those that saturate computational throughput. This result is likely to be extremely difficult to predict with an empirical performance model for this particular algorithm because it has an unstructured memory access pattern that varies locally for individual rays and globally for the selected viewpoint. Our results also show that optimal parameters on modern architectures are markedly different from those in previous studies run on older architectures. In addition, given the dramatic performance variation across platforms for both optimal algorithm settings and performance results, there is a clear benefit for production visualization and analysis codes to adopt a strategy for performance optimization through auto-tuning. These benefits will likely become more pronounced in the future as the number of cores per chip and the cost of moving data through the memory hierarchy both increase.

97 MATHEMATICS AND COMPUTING↗

Scaling soft matter physics to thousands of graphics processing units in parallel

We describe a multi-graphics processing unit (GPU) implementation of the Ludwig application, which specialises in simulating a variety of complex fluids via lattice Boltzmann fluid dynamics coupled to additional physics describing complex fluid constituents. We describe our methodology in augmenting the original central processing unit (CPU) version with GPU functionality in a maintainable fashion. We present several optimisations that maximise performance on the GPU architecture through tuning for the GPU memory hierarchy. We describe how we implement particles within the fluid in such a way to avoid a major diversion of the CPU and GPU codebases, whilst minimising data transfer at each time step. We detail our halo-exchange communication phase for the code, which exploits overlapping to allow efficient parallel scaling to many GPUs. We present results showing that the application demonstrates excellent scaling to at least 8192 GPUs in parallel, the largest system tested at the time of writing. The GPU version (on NVIDIA K20X GPUs) is around 3.5–5 times faster that the CPU version (on fully utilised AMD Opteron 6274 16-core CPUs), comparing equal numbers of CPUs and GPUs.

Computer Science↗

Exploring the feasibility of lossy compression for PDE simulations

Checkpoint restart plays an important role in high-performance computing (HPC) applications, allowing simulation runtime to extend beyond a single job allocation and facilitating recovery from hardware failure. Yet, as machines grow in size and in complexity, traditional approaches to checkpoint restart are becoming prohibitive. Current methods store a subset of the application’s state and exploit the memory hierarchy in the machine. However, as the energy cost of data movement continues to dominate, further reductions in checkpoint size are needed. Lossy compression, which can significantly reduce checkpoint sizes, offers a potential to reduce computational cost in checkpoint restart. This article investigates the use of numerical properties of partial differential equation (PDE) simulations, such as bounds on the truncation error, to evaluate the feasibility of using lossy compression in checkpointing PDE simulations. Restart from a checkpoint with lossy compression is considered for a fail-stop error in two time-dependent HPC application codes: PlasComCM and Nek5000. Results show that error in application variables due to a restart from a lossy compressed checkpoint can be masked by the numerical error in the discretization, leading to increased efficiency in checkpoint restart without influencing overall accuracy in the simulation.

97 MATHEMATICS AND COMPUTING↗

EXAGRAPH: Graph and combinatorial methods for enabling exascale applications

Combinatorial algorithms in general and graph algorithms in particular play a critical enabling role in numerous scientific applications. However, the irregular memory access nature of these algorithms makes them one of the hardest algorithmic kernels to implement on parallel systems. With tens of billions of hardware threads and deep memory hierarchies, the exascale computing systems in particular pose extreme challenges in scaling graph algorithms. The codesign center on combinatorial algorithms, ExaGraph, was established to design and develop methods and techniques for efficient implementation of key combinatorial (graph) algorithms chosen from a diverse set of exascale applications. Algebraic and combinatorial methods have a complementary role in the advancement of computational science and engineering, including playing an enabling role on each other. In this paper, we survey the algorithmic and software development activities performed under the auspices of ExaGraph from both a combinatorial and an algebraic perspective. In particular, we detail our recent efforts in porting the algorithms to manycore accelerator (GPU) architectures. We also provide a brief survey of the applications that have benefited from the scalable implementations of different combinatorial algorithms to enable scientific discovery at scale. We believe that several applications will benefit from the algorithmic and software tools developed by the ExaGraph team.

97 MATHEMATICS AND COMPUTING↗

Quantum Monte Carlo Endstation for Petascale Computing

The major achievements enabled by QMC Endstation grant include * Performance improvement on clusters of x86 multi-core systems, especially on Cray XT systems * New and improved methods for the wavefunction optimizations * New forms of trial wavefunctions * Implementation of the full application on NVIDIA GPUs using CUDA The scaling studies of QMCPACK on large-scale systems show excellent parallel efficiency up to 216K cores on Jaguarpf (Cray XT5). The GPU implementation shows speedups of 10-15x over the CPU implementation on older generation of x86. We have implemented hybrid OpenMP/MPI scheme in QMC to take advantage of multi-core shared memory processors of petascale systems. Our hybrid scheme has several advantages over the standard MPI-only scheme. * Memory optimized: large read-only data to store one-body orbitals and other shared properties to represent the trial wave function and many-body Hamiltonian can be shared among threads, which reduces the memory footprint of a large-scale problem. * Cache optimized: the data associated with an active Walker are in cache during the compute-intensive drift-diffusion process and the operations on an Walker are optimized for cache reuse. Thread-local objects are used to ensure the data affinity to a thread. * Load balanced: Walkers in an ensemble are evenly distributed among threads and MPI tasks. The two-level parallelism reduces the population imbalance among MPI tasks and reduces the number of point-to-point communications of large messages (serialized objects) for the Walker exchange. * Communication optimized: the communication overhead, especially for the collective operations necessary to determine ET and measure the properties of an ensemble, is significantly lowered by using less MPI tasks. The multiple forms of parallelism afforded by QMC algorithms make them ideal candidates for acceleration in the many-core paradigm. We presented the results of our effort to port the QMCPACK simulation code to the NVIDIA CUDA GPU platform. We restructured the CPU algorithms to express additional parallelism, minimize GPU-CPU communication, and efficiently utilize the GPU memory hierarchy. Using mixed precision on GT200 GPUs and MPI for intercommunication and load balancing, we observe typical full-application speedups of approximately 10x to 15x relative to quad-core Xeon CPUs alone, while reproducing the double-precision CPU results within statistical error. We developed an all-electron quantum Monte Carlo (QMC) method for solids that does not rely on pseudopotentials, and used it to construct a primary ultra-high-pressure calibration based on the equation of state of cubic boron nitride. We computed the static contribution to the free energy with the QMC method and obtained the phonon contribution from density functional theory, yielding a high-accuracy calibration up to 900 GPa usable directly in experiment. We computed the anharmonic Raman frequency shift with QMC simulations as a function of pressure and temperature, allowing optical pressure calibration. In contrast to present experimental approaches, small systematic errors in the theoretical EOS do not increase with pressure, and no extrapolation is needed. This all-electron method is applicable to first-row solids, providing a new reference for ab initio calculations of solids and benchmarks for pseudopotential accuracy. We compared experimental and theoretical results on the momentum distribution and the quasiparticle renormalization factor in sodium. From an x-ray Compton-profile measurement of the valence-electron momentum density, we derived its discontinuity at the Fermi wavevector finding an accurate measure of the renormalization factor that we compared with quantum-Monte-Carlo and G0W0 calculations performed both on crystalline sodium and on the homogeneous electron gas. Our calculated results are in good agreement with the experiment. We have been studying the heat of formation for various Kubas complexes of molecular hydrogen on Ti(1,2)ethylene-nH2 using Diffusion Monte Carlo. This work has been started and is ongoing. We are studying systems involving 1 and 2 Ti bonding sites with up to 10 hydrogen molecules in numerous configurations. This work will establish a benchmark that will test the accuracy of density functional calculations and establish the feasibility of our methods for similar systems.

97 MATHEMATICS AND COMPUTING↗

Exascale Hardware Architectures Working Group

The ASC Exascale Hardware Architecture working group is challenged to provide input on the following areas impacting the future use and usability of potential exascale computer systems: processor, memory, and interconnect architectures, as well as the power and resilience of these systems. Going forward, there are many challenging issues that will need to be addressed. First, power constraints in processor technologies will lead to steady increases in parallelism within a socket. Additionally, all cores may not be fully independent nor fully general purpose. Second, there is a clear trend toward less balanced machines, in terms of compute capability compared to memory and interconnect performance. In order to mitigate the memory issues, memory technologies will introduce 3D stacking, eventually moving on-socket and likely on-die, providing greatly increased bandwidth but unfortunately also likely providing smaller memory capacity per core. Off-socket memory, possibly in the form of non-volatile memory, will create a complex memory hierarchy. Third, communication energy will dominate the energy required to compute, such that interconnect power and bandwidth will have a significant impact. All of the above changes are driven by the need for greatly increased energy efficiency, as current technology will prove unsuitable for exascale, due to unsustainable power requirements of such a system. These changes will have the most significant impact on programming models and algorithms, but they will be felt across all layers of the machine. There is clear need to engage all ASC working groups in planning for how to deal with technological changes of this magnitude. The primary function of the Hardware Architecture Working Group is to facilitate codesign with hardware vendors to ensure future exascale platforms are capable of efficiently supporting the ASC applications, which in turn need to meet the mission needs of the NNSA Stockpile Stewardship Program. This issue is relatively immediate, as there is only a small window of opportunity to influence hardware design for 2018 machines. Given the short timeline a firm co-design methodology with vendors is of prime importance.

97 MATHEMATICS AND COMPUTING↗

From Petascale to Exascale: Eight Focus Areas of R&D Challenges for HPC Simulation Environments

Programming models bridge the gap between the underlying hardware architecture and the supporting layers of software available to applications. Programming models are different from both programming languages and application programming interfaces (APIs). Specifically, a programming model is an abstraction of the underlying computer system that allows for the expression of both algorithms and data structures. In comparison, languages and APIs provide implementations of these abstractions and allow the algorithms and data structures to be put into practice - a programming model exists independently of the choice of both the programming language and the supporting APIs. Programming models are typically focused on achieving increased developer productivity, performance, and portability to other system designs. The rapidly changing nature of processor architectures and the complexity of designing an exascale platform provide significant challenges for these goals. Several other factors are likely to impact the design of future programming models. In particular, the representation and management of increasing levels of parallelism, concurrency and memory hierarchies, combined with the ability to maintain a progressive level of interoperability with today's applications are of significant concern. Overall the design of a programming model is inherently tied not only to the underlying hardware architecture, but also to the requirements of applications and libraries including data analysis, visualization, and uncertainty quantification. Furthermore, the successful implementation of a programming model is dependent on exposed features of the runtime software layers and features of the operating system. Successful use of a programming model also requires effective presentation to the software developer within the context of traditional and new software development tools. Consideration must also be given to the impact of programming models on both languages and the associated compiler infrastructure. Exascale programming models must reflect several, often competing, design goals. These design goals include desirable features such as abstraction and separation of concerns. However, some aspects are unique to large-scale computing. For example, interoperability and composability with existing implementations will prove critical. In particular, performance is the essential underlying goal for large-scale systems. A key evaluation metric for exascale models will be the extent to which they support these goals rather than merely enable them.

97 MATHEMATICS AND COMPUTING↗

Scientific Application Requirements for Leadership Computing at the Exascale

The Department of Energy’s Leadership Computing Facility, located at Oak Ridge National Laboratory’s National Center for Computational Sciences, recently polled scientific teams that had large allocations at the center in 2007, asking them to identify computational science requirements for future exascale systems (capable of an exaflop, or 10 18 floating point operations per second). These requirements are necessarily speculative, since an exascale system will not be realized until the 2015–2020 timeframe, and are expressed where possible relative to a recent petascale requirements analysis of similar science applications. Our initial findings, which beg further data collection, validation, and analysis, did in fact align with many of our expectations and existing petascale requirements, yet they also contained some surprises, complete with new challenges and opportunities. First and foremost, the breadth and depth of science prospects and benefits on an exascale computing system are striking. Without a doubt, they justify a large investment, even with its inherent risks. The possibilities for return on investment (by any measure) are too large to let us ignore this opportunity. The software opportunities and challenges are enormous. In fact, as one notable computational scientist put it, the scale of questions being asked at the exascale is tremendous and the hardware has gotten way ahead of the software. We are in grave danger of failing because of a software crisis unless concerted investments and coordinating activities are undertaken to reduce and close this hardware-software gap over the next decade. Key to success will be a rigorous requirement for natural mapping of algorithms to hardware in a way that complements (rather than competes with) compilers and runtime systems. The level of abstraction must be raised, and more attention must be paid to functionalities and capabilities that incorporate intent into data structures, are aware of memory hierarchy, possess fault tolerance, exploit asynchronism, and are power-consumption aware. On the other hand, we must also provide application scientists with the ability to develop software without having to become experts in the computer science components.

Ahern, Sean↗

High performance graphics processor based computed tomography reconstruction algorithms for nuclear and other large scale applications.

The goal of this work is to develop a fast computed tomography (CT) reconstruction algorithm based on graphics processing units (GPU) that achieves significant improvement over traditional central processing unit (CPU) based implementations. The main challenge in developing a CT algorithm that is capable of handling very large datasets is parallelizing the algorithm in such a way that data transfer does not hinder performance of the reconstruction algorithm. General Purpose Graphics Processing (GPGPU) is a new technology that the Science and Technology (S&T) community is starting to adopt in many fields where CPU-based computing is the norm. GPGPU programming requires a new approach to algorithm development that utilizes massively multi-threaded environments. Multi-threaded algorithms in general are difficult to optimize since performance bottlenecks occur that are non-existent in single-threaded algorithms such as memory latencies. If an efficient GPU-based CT reconstruction algorithm can be developed; computational times could be improved by a factor of 20. Additionally, cost benefits will be realized as commodity graphics hardware could potentially replace expensive supercomputers and high-end workstations. This project will take advantage of the CUDA programming environment and attempt to parallelize the task in such a way that multiple slices of the reconstruction volume are computed simultaneously. This work will also take advantage of the GPU memory by utilizing asynchronous memory transfers, GPU texture memory, and (when possible) pinned host memory so that the memory transfer bottleneck inherent to GPGPU is amortized. Additionally, this work will take advantage of GPU-specific hardware (i.e. fast texture memory, pixel-pipelines, hardware interpolators, and varying memory hierarchy) that will allow for additional performance improvements.

97 MATHEMATICS AND COMPUTING↗

MPQC: Performance Analysis and Optimization

MPQC (Massively Parallel Quantum Chemistry) is a widely used computational quantum chemistry code. It is capable of performing a number of computations commonly occurring in quantum chemistry. In order to achieve better performance of MPQC, in this report we present a detailed performance analysis of this code. We then perform loop and memory access optimizations, and measure performance improvements by comparing the performance of the optimized code with that of the original MPQC code. We observe that the optimized MPQC code achieves a significant improvement in the performance through a better utilization of vector processing and memory hierarchies.

97 MATHEMATICS AND COMPUTING↗

Enabling and Supporting HPC [Poster]

We are helping our user community deal with HPC Architectural changes such as many-core nodes, data management, and multi-level memory hierarchies.

97 MATHEMATICS AND COMPUTING↗

Software Synthesis for High Productivity Exascale Computing

Over the three years of our project, we accomplished three key milestones: We demonstrated how ideas from generative programming and software synthesis can help support the development of bulk-synchronous distributed memory kernels. These ideas are realized in a new language called MSL, a C-like language that combines synthesis features with high level notations for array manipulation and bulk-synchronous parallelism to simplify the semantic analysis required for synthesis. We also demonstrated that these high level notations map easily to low level C code and show that the performance of this generated code matches that of handwritten Fortran. Second, we introduced the idea of solver-aided domain-specific languages (SDSLs), which are an emerging class of computer-aided programming systems. SDSLs ease the construction of programs by automating tasks such as verification, debugging, synthesis, and non-deterministic execution. SDSLs are implemented by translating the DSL program into logical constraints. Next, we developed a symbolic virtual machine called Rosette, which simplifies the construction of such SDSLs and their compilers. We have used Rosette to build SynthCL, a subset of OpenCL that supports synthesis. Third, we developed novel numeric algorithms that move as little data as possible, either between levels of a memory hierarchy or between parallel processors over a network. We achieved progress in three aspects of this problem. First we determined lower bounds on communication. Second, we compared these lower bounds to widely used versions of these algorithms, and noted that these widely used algorithms usually communicate asymptotically more than is necessary. Third, we identified or invented new algorithms for most linear algebra problems that do attain these lower bounds, and demonstrated large speed-ups in theory and practice.

97 MATHEMATICS AND COMPUTING↗

Extreme-scale Algorithms and Solver Resilience

A widening gap exists between the peak performance of high-performance computers and the performance achieved by complex applications running on these platforms. Over the next decade, extreme-scale systems will present major new challenges to algorithm development that could amplify this mismatch in such a way that it prevents the productive use of future DOE Leadership computers due to the following; Extreme levels of parallelism due to multicore processors; An increase in system fault rates requiring algorithms to be resilient beyond just checkpoint/restart; Complex memory hierarchies and costly data movement in both energy and performance; Heterogeneous system architectures (mixing CPUs, GPUs, etc.); and Conflicting goals of performance, resilience, and power requirements.

97 MATHEMATICS AND COMPUTING↗

Characterization of bandwidth requirements of algorithms for extreme scale science

Technology trends have resulted in widely different rates of improvement in computational throughput as compared to data movement rates in computer systems. With future systems, the cost of data movement through the memory hierarchy is expected to become even more dominant relative to the cost of performing arithmetic operations, both in terms of time and energy. Therefore, optimizing data access costs will become ever more critical in the coming years. However, unlike the well understood model of computational complexity in terms of the number of elementary operations to execute an algorithm, the inherent data-movement complexity of computations is unknown for most algorithms. Given the crucial importance of optimizing data access costs in scalable parallel systems, it is of great interest to develop techniques for understanding the inherent data movement complexity of an algorithm. In this project, we focus on this fundamental and significant problem.

97 MATHEMATICS AND COMPUTING↗

Exascale Operating Systems and Runtime Software Report

Here U.S. Department of Energy (DOE) workshops and reports have identified four key exascale challenges: dramatically improving power efficiency; improving resilience in the presence of increasing faults; enabling efficient data movement across deepening memory hierarchies and new storage technologies; and managing dramatically increased parallelism, especially at the node level. Software solutions that address these challenges must also improve programmability, expanding the community of computational scientists who can use leadership-class platforms. To address these challenges, DOE must develop new techniques, novel designs, and advanced software architectures for next-generation exascale software infrastructure. In this report, we discuss challenges and approaches to exascale operating system and runtime (OS/R) software. The Exascale Operating Systems and Runtime (OS/R) Software Technical Council canvassed hardware architects, application programmers, parallel tool developers, DOE high-performance computing (HPC) facilities, and the vendors that sell and support integrated platforms. After considering the collective requirements of these constituents and examining the future research challenges and current software gaps, the council recommends that DOE invest in targeted advanced computer science research and in specific coordinating activities to enable the effective and successful development of anticipated exascale systems.

97 MATHEMATICS AND COMPUTING↗