Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Intel”

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 829 records · Page 46

Evaluating the Suitability of Commercial Clouds for NASA's High Performance Computing Applications: A Trade Study

NASA’s High-End Computing Capability (HECC) Project is periodically asked if it could be more cost effective through the use of commercial cloud resources. To answer the question, HECC’s Application Performance and Productivity (APP) team undertook a performance and cost evaluation comparing three domains: two commercial cloud providers, Amazon and Penguin, and HECC’s in-house resources—the Pleiades and Electra systems. In the study, the APP team used a combination of the NAS Parallel Benchmarks (NPB) and six full applications from NASA’s workload on Pleiades and Electra to compare performance of nodes based on three different generations of Intel Xeon processors—Haswell, Broadwell, and Skylake. Because of export control limitations, the most heavily used applications on Pleiades and Electra could not be used in the cloud; therefore, only one of the applications, OpenFOAM, represents work from the Aeronautics Research Mission Directorate and the Human and Exploration Mission Directorate. The other five applications are from the Science Mission Directorate.

High Performance↗

High-Performance Computing Optimization for Aladyn – Adaptive Neural Network Molecular Dynamics Mini-Application

This report provides a description and performance evaluation of the optimization techniques for high performance computing (HPC) implementation of the open source Computational Materials mini-application Aladyn (https://github.com/nasa/aladyn). Aladyn is a basic molecular dynamics code written in FORTRAN 2003, which is designed to demonstrate the use of adaptive neural networks (ANNs) in atomistic simulations. The role of ANNs is to efficiently reproduce the very complex energy landscape resulting from the atomic interactions in materials with the accuracy of the more expensive quantum mechanics-based calculations. The ANN is trained on a large set of atomic structures calculated using the density functional theory (DFT) method. While achieving orders of magnitude faster computational performance than DFT, the ANN-based approach was still very computationally demanding compared to the conventional approach of using empirically fitted energy functions. After its initial development, Aladyn was evaluated and optimized by experts at the NASA Advanced Supercomputing (NAS) division to exploit modern supercomputer architectures. The code has been optimized for execution on multicore central processing units (CPUs), including Intel® Skylake microarchitecture, and on graphic accelerators, such as Nvidia® V100 graphic processing units (GPUs), using Open Multi-Processing (OpenMP) and Open Accelerators (OpenACC) programming interfaces. The optimization achieved a speedup of 4.7 times the baseline version on CPU performance and an additional 2.4 times on CPU+GPU performance. Atomistic computer simulations are a fundamental tool in materials research to model material properties form physics-based first principles. Atomic interaction, governed by Quantum Mechanics (QM) require sophisticated and highly computationally demanding mathematical models to calculate [1]. Classical methods use approximate functional forms, empirically fitted through a set of variable parameters to emulate atomic energies as direct functions of atomic coordinates [2]. While empirical potentials are computationally much simpler, allowing simulations of large-scale systems of up to a trillion (1012) atoms [3], they are substantially less accurate compared to quantum calculations and applicable only to very specific atomic configurations or predefined crystallographic phases. A recently suggested approach is to use heuristic machine learning methods [4], such as those based on Adaptive Neural Networks (ANNs) to predict atomic energies, after being trained on a sufficiently large database of QM-calculated structures [5,6]. This approach reduces significantly the computational complexity, allowing for simulations of orders of magnitude larger systems compared to QM-based methods without compromising accuracy. Still, compared to classical methods using empirical energy functions, ANN methods remain two- to three orders of magnitude more computationally demanding. Hence, the computational cost of simulations, together with the need for extensive training of ANNs, still makes the practical implementation of ANN-based methods quite challenging. The purpose of the Aladyn mini-application software [7], available as open source at https://github.com/nasa/aladyn, is to be a testbed for exploring possible optimization strategies to develop highly scalable parallel algorithms for ANN-based atomistic simulations. Aladyn is aimed at utilizing the architecture of the high-end modern highperformance computing (HPC) hardware based on multicore central processing units (CPUs) equipped with graphic processing unit (GPU) accelerators. Specifically, the goal is to optimize the performance on a single HPC compute node, before implementing scaling to multi-node parallelization using message passing interface (MPI). At the same time, the open source code of Aladyn can serve as a training model for students and professors in academia.

Yamakov, Vesselin I.↗

Accelerated Simulation of Air Pollution Using NVIDIA RAPIDS

Atmospheric chemistry models are a central tool to study and forecast the impact of air pollution on the environment, vegetation, and human health. However, the numerical simulation of chemical kinetics is computationally expensive due to the stiffness of the system of ordinary differential equations that describes atmospheric chemistry. Here we present an alternative approach to the computation of atmospheric chemistry based on machine learning. Our training data set is produced using the NASA Goddard Earth Observing System (GEOS) model with GEOS-Chem chemistry, run on the NASA Center for Climate Simulation (NCCS) Discover supercomputing cluster on 384 Intel Xeon Haswell cores. This model spends more than 50% of total run time on solving atmospheric chemistry. The data set contains as input features the air pollution concentrations before solving the differential equations, together with some key physical parameters such as temperature and sun intensity. As target variables we define the air pollution concentrations after solving the differential equations. Using Dask-cuDF and Dask-XGBoost on the NVIDIA RAPIDS platform on 8 Tesla V100 GPUs, we generate from this training set gradient boosted decision tree models that can reproduce the simulation of chemical kinetics. We do this on the NCCS Advanced Data Analytics Platform (ADAPT) science cloud environment. Our application takes full advantage of recent advances in Dask-XGBoost, such as multi-node and multi-GPU scaling for distributed training with large data sets. The increase in training data size enabled by this is critical to capture the full range of chemical environments encountered across the globe and all annual seasons.The boosted tree models offer good predictability and show many of the features of the full chemistry reference simulation. Further improvements can be achieved through mass balance considerations and by accounting for error correlations. We incorporate the boosted tree models into the GEOS reference model using XGBoost's C API. This enables a seamless integration of the GPU trained models into GEOS-Chem, which is written in Fortran and optimized for use in a massively parallel CPU environment. We show the benefits of this approach and discuss the potential speedup of this machine learning accelerated atmospheric chemistry model.

Keller, Christoph A.↗

Optimization of a Solver for Computational Materials and Structures Problems on NVIDIA Volta and AMD Instinct GPUs

The Scalable Implementation of Finite Elements by NASA (ScIFEN) is a software package developed to solve complex computational materials and structures problems using the finite element method (FEM). In this paper, we describe optimization techniques to speed up the linear solver computation that occurs within the ScIFEN application. We consider GPUs from two different vendors, NVIDIA and AMD as our target platforms for optimization and highlight differences in performance and optimization techniques. The NVIDIA GPU Volta V100 is used in the Summit system deployed at Oak Ridge National Laboratory, and the new exascale system, Frontier, will be using AMD Radeon Instinct GPU. We evaluated the performance of various optimization techniques on test matrices, ranging in size from100K to 4M, that are representative of ScIFEN applications. The linear solver computation is memory-bound on both GPUs. Our experiments show that on the NVIDIA GPU we obtained up to79%of the theoretical peak bandwidth, while the AMD GPU achieved 59%. Overall, the NVIDIA V100 GPU outperforms the AMD MI 25 GPU1. We observed an overall speedup of up to37X on an NVIDIA V100 compared to an Intel Skylake 12-coremachine. The solver for a 4M degree of freedom system took under 2.5 seconds.

Mohammad Zubair↗

TPSAS-NF1676L-30692-DND

Objectives: transform a branch of the FUN3D mini-app into OpenCL, benchmark OpenCL routines against sequential and CUDA versions of the same routines, use Intel OpenCL SDK to run on FPGA, and compare performance and suggest opportunities for optimization.

Jason Orender↗

TechEdSat-13: The First Flight of an AI/ML Neuromorphic Processor

Neuromorphic processors, inspired by the wiring of the brain, permit certain classes of Artificial Intelligence/Machine Learning (AI/ML) algorithms to run far more efficiently. Ultimately, such systems will make the small- and nano-satellite platforms even more useful in terms of greatly improved power, communication and internal data management. In addition, on-board processing of images and data will help to not only rapidly interpret the information, but also reduce the amount of data that needs to be transmitted to ground stations. This initial flight experiment uses the Intel/ Loihi processor combined with a custom interface board and three communication channels to run the AI/ML scripts. These will vary with increasing length and complexity during the course of the mission. The algorithms will use the (at first) limited sensor data to ‘learn’ - with comparisons to similar architecture in comparable ground experiments. Some of the applications for successor flights in the TES-n flight series include Cognitive Communications, whereby the overall communication system is optimized per overflight – by optimizing timing and data transmission functions. Lastly, the performance of the Loihi 14nm process technology will be monitored for performance in the LEO radiation environment, thus looking for induced hardware and software errors. This information will help guide future radiation protection techniques to extend the lifetime and overall utility. The TES-13 is a 3U nanosat successfully launched by the Virgin Orbit Launcher-1 on January 13, 2022, and will presage more flights and AI/ML applications to come.

TechEdSat↗

Using Federated Learning to Overcome Data Gravity in Space

Humans intend to take longer missions to outer space. Understanding the impact that space has on human health is paramount to the success of these missions. Controlled experiments with model organisms are run to infer the impact of space conditions on human health, but the data these experiments generate are too large to transfer to Earth for building models. The same is true for space-relevant data generated on Earth. Ideally, these datasets should be combined to improve statistical power and model accuracy without having to transfer data. Federated learning is such a method which trains an algorithm across decentralized computing systems, each of which has their own local copy of training and testing data. In this research, made possible by NASA@Work, the AI for Life in Space group at NASA demonstrates the use of federated learning to train an ensemble of causality inference models on a combination of data residing on the International Space Station (ISS) and in the cloud. Our work leverages CRISP, a causal inference platform developed during the 2020 Frontier Development Lab’s “Astronaut Health Challenge.” We also leverage the OpenFL federated learning library which was collaboratively developed at Intel and UPenn. We used publicly available data from the NASA Ames Life Sciences Data Archive to identify features in ionizing radiation experiments as causal of changes in cardiac blood velocity. This research demonstrates, for the first time, the possibility of running machine learning algorithms on datasets separated by astronomical distances. In this experiment, all the data were generated in terra, half of which were transferred to the ISS and analyzed on the Spaceborne Computer. In the future, our research will leverage federated learning on data generated in situ on the ISS with data generated terrestrially to predict the impact of spaceflight on mammalian female reproductive capacity.

James Casaletto↗

Porting OVERFLOW CFD Code to GPUs: To Hackathons and Beyond!

OVERFLOW is an overset, structured computational fluid dynamics (CFD) code written in Fortran which is widely used in the government, industry, and academia. Over the last several years the OVERFLOW developers have been working to port miniapps based on computationally expensive parts of OVERFLOW to run on GPUs, primarily using OpenACC. This effort started at our first hackathon in 2019 and since then the OVERFLOW team has attended two additional hackathons (virtually). These hackathon environments have provided a great place to collaborate with others and learn from experts. These learning experiences enabled porting two miniapps to run effectively on NVIDIA GPUs using OpenACC. The first miniapp focused on motifs found in the solver itself and the final ported version runs three times fast ona single V100 compared to a 40 core, dual-socket Intel Skylake node. The speed up in this solverminiapp required multiple design changes including increasing the amount of parallelism available and the amount of work performed in each kernel. The second miniapp focused on overset MPI communication, also saw significant speedups over the CPU implementation using a CUDA-aware MPI implementation through OpenACC. This presentation will discuss our experience at the hackathons, our process of porting the miniapps to run on the GPUs, and several lessons learned throughout.

OpenACC↗

Event Report for The Ethical Artificial Intelligence Quantification Workshop

Artificial Intelligence (AI) is a powerful emerging technology area which requires special attention to using it ethically. AI ethics is still an emerging field, and the partners for this workshop and report seek to move AI ethics discussion ahead by experimenting with ways to measure AI ethics criteria. The following document describes the outcomes and learnings from The Ethical Artificial Intelligence Quantification Workshop held at the National Institute for Aerospace (NIA), Hampton, Virginia on May 12th, 2022. The purpose of the workshop was for participants to evaluate and experiment-with the methodology and process presented by AIEthics.World in cooperation with Intel Corporation. The meeting participants learned about the Ethical AI Certification and Maturity Model™ and applied the methodology to selected notional AI systems. The workshop facilitated the evaluation of the maturity of the AI system according to ethical considerations relevant to NASA, NIA and other participants. The workshop consisted of three main phases. The first phase focused on understanding and summarizing NASA’s ethical approaches, mission and values based on published documentation, discussions and individual insights & opinions of participants. This information was prioritized, weighted, ordered, and quantified in phase two, to formulate an alignment between human values (ethics) and their applicability to AI systems during all lifecycle phases. The first two phases were summarized as a form of ethical genealogy for artificial intelligence, specific to NASA’s ethical approaches. In the third and last phase of the workshop the participants evaluated notional examples of artificial intelligence to qualify and quantify its ability to adhere to the organizational ethics approaches, using the Ethical AI Certification and Maturity Model™. The workshop uses the concept of genealogy, in the traditional sense: the study and traceability of lines of ancestors in the process of evolutionary development from earlier forms. However, as it is applied to an Ethical AI definition, it is providing the insights to the necessary and mandatory traceability of content, data, metrics, telemetry, elements, and structures which are used in the AI’s lifecycle to foster and measure AI ethics in all steps of its lifecycle. The Ethical Artificial Intelligence Quantification Workshop provided NASA with the opportunity to apply the Ethical AI Certification and Maturity Model™, in combination with existing and well-known decision-making and quality control methods to identify the metrics and measurements for an Ethical AI and assess its ethical condition and quality aligned with NASA ethics approaches. The result of the workshop is the capacity for NASA to apply the maturity model assessment to its AI Systems as desired and if necessary, publish the ability of these AI Systems to adhere to the organizational ethical goals. AI ethics frameworks need to be customized for each application domain, for example, individual NASA Mission Directorates. General principles that work in one area such as AI/Machine Learning-based text analysis (the ethics of information-extraction) may need to be adapted for another such as sense-and-avoid decision-making in a flight environment. The workshop was conducted among approximately twenty NASA subject matter experts, so the elements noted above should be considered examples, not definitive NASA ethical AI principles, genealogy, etc. Generating a definitive AI ethics framework for an organization as diverse as NASA would require far more discussion, debate, review, etc. However, the workshop provided valuable insight into mechanisms and processes for quantifying AI ethical qualities.

Artificial Intelligence↗

A Multi-Architecture Approach for Implicit Computational Fluid Dynamics on Unstructured Grids

High-performance computing (HPC) architectures are trending toward manycore paradigms such as graphics processing units (GPUs). Approximately half of the top 100 publicly disclosed supercomputers in the world utilize GPU accelerators for performance. This is in contrast to a decade ago, where there were only a few such machines in the top 100. It is not currently possible to compile and run legacy central processing unit (CPU) software efficiently on GPUs without significant refactoring. Though a number of frameworks offering performance portability exist, none offer a standardized specification that is supported by all major hardware vendors. Additionally, experiences show that obtaining a high percentage of peak performance often requires architecture-specific code. This work details a pragmatic multi-architecture computational fluid dynamics library focused on aerospace problems across the speed range from low subsonic to hypersonic flows involving thermochemical nonequilibrium. A thin abstraction layer above NVIDIA CUDA C++ is utilized, which enables primarily single-source software currently capable of running efficiently on multicore CPUs, NVIDIA GPUs, AMD GPUs, and Intel GPUs. Results on various problems of interest across the speed range are presented and performance is compared between various architectures.

GPU↗

A Multi-Architecture Approach for Implicit Computational Fluid Dynamics on Unstructured Grids

High-performance computing (HPC) architectures are trending toward manycore paradigms such as graphics processing units (GPUs). Approximately half of the top 100 publicly disclosed supercomputers in the world utilize GPU accelerators for performance. This is in contrast to a decade ago, where there were only a few such machines in the top 100. It is not currently possible to compile and run legacy central processing unit (CPU) software efficiently on GPUs without significant refactoring. Though a number of frameworks offering performance portability exist, none offer a standardized specification that is supported by all major hardware vendors. Additionally, experiences show that obtaining a high percentage of peak performance often requires architecture-specific code. This work details a pragmatic multi-architecture computational fluid dynamics library focused on aerospace problems across the speed range from low subsonic to hypersonic flows involving thermochemical nonequilibrium. A thin abstraction layer above NVIDIA CUDA C++ is utilized, which enables primarily single-source software currently capable of running efficiently on multicore CPUs, NVIDIA GPUs, AMD GPUs, and Intel GPUs. Results on various problems of interest across the speed range are presented and performance is compared between various architectures.

GPU↗

New Rover Conops with High-Performance Onboard Computing: Give Up Raw Data to Reduce Ops Cost and Do More Science

A major portion of time during the tactical operation of Mars rovers is spent for selecting, prioritizing, and coordinating sciences and engineering activities such that they fit within resource constraints, including the downlink data volume, energy, and time. In particular, the downlink data volume constraint is getting particularly tighter in recent missions because modern instruments produce increasingly high data volume while the communication bandwidth is essentially bounded by the law of physics. Tactical operation would be substantially simplified, hence the operation cost could be reduced, if the data volume constraint is relaxed or even removed. In this abstract, we propose a new operation paradigm for achieving this goal. The key observation is that, both in science and engineering applications, the bit size of raw data is typically much greater than the volume of processed information that is needed for scientific or engineering analysis. For example, a full-resolution image from Mastcam-Z, the main science camera on Perseverance, is about 700 kB in volume and we downlinked 29,685 images up to Sol 243, totaling ~20 GB of data. But of course, scientists do not use every pixel of these images; what they really look for in the images are geological features, typically represented by specific geometric configurations or textures. An end product after processing hundreds of Mascam-Z images could be a single geological map summarizing the spatial distribution of the features. For another example, a 100-meter drive of Perseverance produces 7-12 MB of drive telemetry, which records every detail of the rover's motion at 8 Hz, including position, attitude, steering angles, encoder readings, motor currents and many other information. But what the ground engineers eventually pay attention to is the signs of anomaly, such as excessive motor currents or high slip; if a drive is nominal, the vast majority of this data is unused. What if, then, we process the raw data onboard and only downlink the processed data that is relevant to scientific or engineering analyses, such as a list of detected science features (with cropped images) or a list of potential signs of anomaly while driving? A major roadblock for such onboard, high-level information processing has been the onboard computational resource. RAD750, the main onboard computer of Perseverance, is obviously not sufficient for performing complex image or signal processing such as object detection, semantic segmentation, or anomaly detection. Interestingly, RAD750 is not the best processor that Perseverance has; Qualcomm's Snapdragon 801, a modern mobile processor, is on her Heli Base Station, a device for communicating with Mars Helicopter Ingenuity; also, Intel's Atom E3845 processors are on engineering cameras. In the reminder of this paper, we will introduce two particular uses cases of these high-performance co-processors (meaning auxiliary CPU, GPU, or other types of processors that are separate from the main processor that runs the main flight software) for lowering operation cost and accommodating more science activities for a given communication constraint.

Didier, A.↗

3D Scanning System to Assess Gravity-Dependent Body Shape Changes

The human body shows unique morphological changes when exposed to different gravity conditions, including muscle atrophy, fluid shift, and spinal elongation. Such changes need to be incorporated for human-system integration in the vehicle habitat, garment, and spacesuit designs, as inaccurate body measurements can result in suboptimal crew protection that can potentially decrease injury tolerance. However, measurement tools have not been available for accurate assessments of body shape changes. This work aimed to develop a prototype 3-D body scanning system with the configuration and performance optimized for in-flight crewmember body scanning. A scan hardware system was developed using Intel RealSense commercial off-the-shelf 3D sensors. The sensor parameters were iteratively optimized and tested to obtain the performance level needed for body scanning. A scan booth structure was fabricated, with the overall size 4 x 4 x 8 feet. The specific number of sensors and mounting positions were determined by iterative simulations, which indicated that 16 cameras can capture 94% and 96% of body surface area from the 1st percentile female and 99th percentile male crew population subject. The mounted sensors were linked through a mix of USB-C and USB-3 cables and operated for data acquisition from a Linux laptop computer. A software prototype was developed using Python and Tkinter graphical user interface toolkit. A calibration procedure was also built using a panel of QR codes. A computer vision tool detected and decoded the unique ID and pattern locations of the QR codes. The calibration information determined the position and orientation of the 3D sensors with respect to each other. The scanner performance was assessed using a set of 3D printed custom manikins. The manikin size and shape were derived from the previous ISS study, which measured the crewmembers’ anthropometry changes across the different flight phases. The average anthropometric measurements at the pre-flight and flight day 15 were sampled and projected onto the 1st percentile female and 99th percentile male body shapes. Another pair of manikins were also 3D printed to simulate the neutral body posture, estimated from ISS microgravity environments. A preliminary analysis assessed the performance of the newly developed scanner against the reference scanner, which has been used at the NASA JSC for crew and test subject anthropometry. Although the new scanner data showed several artifacts and missing geometries in the occluded body areas such as armpits and crotch, overall shape matched with the reference scan. When the manikin surface coordinates were compared, a root mean square error of 1.3 cm was observed from the manikin torso segment. Linear measurements including the stature, knee height and circumference measurements at the chest and calf showed differences from the reference scan measurements, ranging between 0.3 and 0.9 cm. Overall, this work demonstrated a development framework for an in-flight scanner with design and operation optimized for crewmember body scanning. Further improvement can potentially provide previously unavailable anthropometric data from different gravitational environments, including 0-g, 1/6-g, and 1-g. Such data can improve suit fit, habitat design, exercise efficacy quantification and sizing of orthostatic intolerance garments.

K H Kim↗

Empirical Memory-Access Cost Models in Multicore NUMA Architectures

Data location is of prime importance when scheduling tasks in a non-uniform memory access (NUMA) architecture. The characteristics of the NUMA architecture must be understood so tasks can be scheduled onto processors that are close to the task's data. However, in modern NUMA architectures, such as AMD Magny-Cours and Intel Nehalem, there may be a relatively large number of memory controllers with sockets that are connected in a non-intuitive manner, leading to performance degradation due to uninformed task-scheduling decisions. In this paper, we provide a method for experimentally characterizing memory-access costs for modern NUMA architectures via memory latency and bandwidth microbenchmarks. Using the results of these benchmarks, we propose a memory-access cost model to improve task-scheduling decisions by scheduling tasks near the data they need. Simple task-scheduling experiments using the memory-access cost models validate the use of empirical memory-access cost models to significantly improve program performance.

BENCHMARKS↗

First record of single event upset on the ground, Cray-1 computer memory at Los Alamos in 1976

Records of bit flips in the Cray-1 computer installed at Los Alamos in 1976 lead to an upset rate in the Cray-1 's bipolar SRAMs that correlates with the SEUs being induced by the atmospheric neutrons. In 1976 the Cray Research Company delivered its first supercomputer, the Cray-1, installing it at Los Alamos National Laboratory. Los Alamos had competed with the Lawrence Livermore National Laboratory for the Cray-1 and won, reaching an agreement with Seymour Cray to install the machine for a period of six months for free, after which they could decide whether to buy, lease or return it. As a result, Los Alamos personnel kept track of the computer reliability and performance and so we know that during those six months of operation, 152 memory parity errors were recorded. The computer memory consisted of approximately 70,000 1Kx1 bipolar ECL static RAMs, the Fairchild 10415. What the Los Alamos engineers didn't know is that those bit flips were the result of single event upsets (SEUs) caused by the atmospheric neutrons. Thus, these 152 bit flips were the first recorded SEUs on the earth, and were observed 2 years before the SEUs in the Intel DRAMs that had been found by May and Woods in 1978. The upsets in the DRAMs were shown to have been caused by alpha particles from the chip packaging material. In this paper we will demonstrate that the Cray-1 bit flips, which were found through the use of parity bits in the Cray-1, were likely due to atmospheric neutrons. This paper will follow the same approach as that of the very first paper to demonstrate single event effects, which occurred in satellite flip-flop circuits in 1975. The main difference is that in the four events that occurred over the course of 17 satellite years of operation were shown to be due to single event effects just a few years after those satellite anomalies were recorded. In the case of the Cray-1 bit flips, there has been a delay of more than 30 years between the occurrence of the bit flips and the identification of their cause as SEUs induced by the atmospheric neutrons.

42 ENGINEERING↗

Computation of Large Covariance Matrices by SAMMY on Graphical Processing Units and Multicore CPUs

Computational power of Graphical Processing Units and multicore CPUs was harnessed by the nuclear data evaluation code SAMMY to speed up computations of large Resonance Parameter Covariance Matrices (RPCMs). This was accomplished by linking SAMMY to vendor-optimized implementations of the matrix-matrix multiplication subroutine of the Basic Linear Algebra Library to compute the most time-consuming step. The U-235 RPCM computed previously using a triple-nested loop was re-computed using the NVIDIA implementation of the subroutine on a single Tesla Fermi Graphical Processing Unit, and also using the Intel's Math Kernel Library implementation on two different multicore CPU systems. A multiplication of two matrices of dimensions 16,000 x 20,000 that had previously taken days, took approximately one minute on the GPU. Similar performance was achieved on a dual six-core CPU system. The magnitude of the speed-up suggests that these, or similar, combinations of hardware and libraries may be useful for large matrix operations in SAMMY. Uniform interfaces of standard linear algebra libraries make them a promising candidate for a programming framework of a new generation of SAMMY for the emerging heterogeneous computing platforms.

99 GENERAL AND MISCELLANEOUS↗

Concurrent Collections (CnC): A new approach to parallel programming

A common approach in designing parallel languages is to provide some high level handles to manipulate the use of the parallel platform. This exposes some aspects of the target platform, for example, shared vs. distributed memory. It may expose some but not all types of parallelism, for example, data parallelism but not task parallelism. This approach must find a balance between the desire to provide a simple view for the domain expert and provide sufficient power for tuning. This is hard for any given architecture and harder if the language is to apply to a range of architectures. Either simplicity or power is lost. Instead of viewing the language design problem as one of providing the programmer with high level handles, we view the problem as one of designing an interface. On one side of this interface is the programmer (domain expert) who knows the application but needs no knowledge of any aspects of the platform. On the other side of the interface is the performance expert (programmer or program) who demands maximal flexibility for optimizing the mapping to a wide range of target platforms (parallel / serial, shared / distributed, homogeneous / heterogeneous, etc.) but needs no knowledge of the domain. Concurrent Collections (CnC) is based on this separation of concerns. The talk will present CnC and its benefits. About the speaker. Kathleen Knobe has focused throughout her career on parallelism especially compiler technology, runtime system design and language design. She worked at Compass (aka Massachusetts Computer Associates) from 1980 to 1991 designing compilers for a wide range of parallel platforms for Thinking Machines, MasPar, Alliant, Numerix, and several government projects. In 1991 she decided to finish her education. After graduating from MIT in 1997, she joined Digital Equipment’s Cambridge Research Lab (CRL). She stayed through the DEC/Compaq/HP mergers and when CRL was acquired and absorbed by Intel. She currently works in the Software and Services Group / Technology Pathfinding and Innovation.

97 MATHEMATICS AND COMPUTING↗

A communication-avoiding, hybrid-parallel, rank-revealing orthogonalization method.

Orthogonalization consumes much of the run time of many iterative methods for solving sparse linear systems and eigenvalue problems. Commonly used algorithms, such as variants of Gram-Schmidt or Householder QR, have performance dominated by communication. Here, 'communication' includes both data movement between the CPU and memory, and messages between processors in parallel. Our Tall Skinny QR (TSQR) family of algorithms requires asymptotically fewer messages between processors and data movement between CPU and memory than typical orthogonalization methods, yet achieves the same accuracy as Householder QR factorization. Furthermore, in block orthogonalizations, TSQR is faster and more accurate than existing approaches for orthogonalizing the vectors within each block ('normalization'). TSQR's rank-revealing capability also makes it useful for detecting deflation in block iterative methods, for which existing approaches sacrifice performance, accuracy, or both. We have implemented a version of TSQR that exploits both distributed-memory and shared-memory parallelism, and supports real and complex arithmetic. Our implementation is optimized for the case of orthogonalizing a small number (5-20) of very long vectors. The shared-memory parallel component uses Intel's Threading Building Blocks, though its modular design supports other shared-memory programming models as well, including computation on the GPU. Our implementation achieves speedups of 2 times or more over competing orthogonalizations. It is available now in the development branch of the Trilinos software package, and will be included in the 10.8 release.

99 GENERAL AND MISCELLANEOUS↗