Search NASA⌕ Search

SEARCH · Search NASA

Results for “Compiler techniques”

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 145 records · Page 8

Contributed Review: Culet diameter and the achievable pressure of a diamond anvil cell: Implications for the upper pressure limit of a diamond anvil cell

Recently, static pressures of more than 1.0 TPa have been reported, which raises the question: what is the maximum static pressure that can be achieved using diamond anvil cell techniques? Here in this paper we compile culet diameters, bevel diameters, bevel angles, and reported pressures from the literature. We fit these data and find an expression that describes the maximum pressure as a function of the culet diameter. An extrapolation of our fit reveals that a culet diameter of 1 μm should achieve a pressure of ~1.8 TPa. Additionally, for pressure generation of ~400 GPa with a single beveled diamond anvil, the most commonly reported parameters are a culet diameter of ~20 μm, a bevel angle of 8.5°, and a bevel diameter to culet diameter ratio between 14 and 18. Our analysis shows that routinely generating pressures more than ~300 GPa likely requires diamond anvil geometries that are fundamentally different from a beveled or double beveled anvil (e.g., toroidal or double stage anvils) and culet diameters that are ≤20 μm.

36 MATERIALS SCIENCE↗

Modeling Cooperative Threads to Project GPU Performance for Adaptive Parallelism

Most accelerators, such as graphics processing units (GPUs) and vector processors, are particularly suitable for accelerating massively parallel workloads. On the other hand, conventional workloads are developed for multi-core parallelism, which often scale to only a few dozen OpenMP threads. When hardware threads significantly outnumber the degree of parallelism in the outer loop, programmers are challenged with efficient hardware utilization. A common solution is to further exploit the parallelism hidden deep in the code structure. Such parallelism is less structured: parallel and sequential loops may be imperfectly nested within each other, neigh boring inner loops may exhibit different concurrency patterns (e.g. Reduction vs. Forall), yet have to be parallelized in the same parallel section. Many input-dependent transformations have to be explored. A programmer often employs a larger group of hardware threads to cooperatively walk through a smaller outer loop partition and adaptively exploit any encountered parallelism. This process is time-consuming and error-prone, yet the risk of gaining little or no performance remains high for such workloads. To reduce risk and guide implementation, we propose a technique to model workloads with limited parallelism that can automatically explore and evaluate transformations involving cooperative threads. Eventually, our framework projects the best achievable performance and the most promising transformations without implementing GPU code or using physical hardware. We envision our technique to be integrated into future compilers or optimization frameworks for autotuning.

Meng, Jiayuan↗

Computer enhancement through interpretive techniques

The improvement in the usage of the digital computer through the use of the technique of interpretation rather than the compilation of higher ordered languages was investigated by studying the efficiency of coding and execution of programs written in FORTRAN, ALGOL, PL/I and COBOL. FORTRAN was selected as the high level language for examining programs which were compiled, and A Programming Language (APL) was chosen for the interpretive language. It is concluded that APL is competitive, not because it and the algorithms being executed are well written, but rather because the batch processing is less efficient than has been admitted. There is not a broad base of experience founded on trying different implementation strategies which have been targeted at open competition with traditional processing methods.

Foster, G.↗

Knowledge-based design of generate-and-patch problem solvers that solve global resource assignment problems

We present MENDER, a knowledge based system that implements software design techniques that are specialized to automatically compile generate-and-patch problem solvers that satisfy global resource assignments problems. We provide empirical evidence of the superior performance of generate-and-patch over generate-and-test: even with constrained generation, for a global constraint in the domain of '2D-floorplanning'. For a second constraint in '2D-floorplanning' we show that even when it is possible to incorporate the constraint into a constrained generator, a generate-and-patch problem solver may satisfy the constraint more rapidly. We also briefly summarize how an extended version of our system applies to a constraint in the domain of 'multiprocessor scheduling'.

Voigt, Kerstin↗

Constructing the 'Best' Reliability Data for the Job - Developing Generic Reliability Data from Alternative Sources Early in a Product's Development Phase

Reliability practitioners advocate getting reliability involved early in a product development process. However, when assigned to estimate or assess the (potential) reliability of a product or system early in the design and development phase, they are faced with lack of reasonable models or methods for useful reliability estimation. Developing specific data is costly and time consuming. Instead, analysts rely on available data to assess reliability. Finding data relevant to the specific use and environment for any project is difficult, if not impossible. Instead, analysts attempt to develop the "best" or composite analog data to support the assessments. Industries, consortia and vendors across many areas have spent decades collecting, analyzing and tabulating fielded item and component reliability performance in terms of observed failures and operational use. This data resource provides a huge compendium of information for potential use, but can also be compartmented by industry, difficult to find out about, access, or manipulate. One method used incorporates processes for reviewing these existing data sources and identifying the available information based on similar equipment, then using that generic data to derive an analog composite. Dissimilarities in equipment descriptions, environment of intended use, quality and even failure modes impact the "best" data incorporated in an analog composite. Once developed, this composite analog data provides a "better" representation of the reliability of the equipment or component. It can be used to support early risk or reliability trade studies, or analytical models to establish the predicted reliability data points. It also establishes a baseline prior that may updated based on test data or observed operational constraints and failures, i.e., using Bayesian techniques. This tutorial presents a descriptive compilation of historical data sources across numerous industries and disciplines, along with examples of contents and data characteristics. It then presents methods for combining failure information from different sources and mathematical use of this data in early reliability estimation and analyses.

Kleinhammer, Roger K.↗

Detection of counterfeit and compromised devices using system and function call tracing techniques

Frameworks, methods, and systems for securing a smart grid are provided. A framework can include data collection, call tracing techniques, and preparing call lists to detect counterfeit or compromised devices. The call tracing techniques can include call tracing and compiling all system and function calls over a time interval. The framework can further include data processing, in which a genuine device is identified and compared to unknown devices. A first statistical correlation can be used for resource-rich systems, and a second statistical correlation can be used for resource-limited systems. Threats of information leakage, measurement poisoning and store-and-send-later can be considered.

Babun, Leonardo↗

Advanced smoke meter development survey and analysis

Ideal smoke meter characteristics are determined to provide a basis for evaluation of candidate systems. Five promising techniques are analyzed in detail to evaluate compilance with the practical smoke meter requirements. Four of the smoke measurement concepts are optical methods: Modulated Transmission (MODTRAN), Cross Beam Absorption Counter (CBAC), Laser Induced Incandescence (LIN), and Photoacoustic Spectroscopy (PAS). A rapid response filter instrument called a Taper Element Oscillating Microbalance (TEOM) is also evaluated. For each technique, the theoretical principles are described, the expected performance is determined, and the advantages and disadvantages are discussed The expected performance is evaluated against each of the smoke meter specifications, and the key questions for further study are given. The most promising smoke meter technique analyzed was MODTRAN, which is a variation on a direct transmission measurement. The soot-laden gas is passed through a transmission cell, and the gas pressure is modulated by a speaker.

Pitz, R. W.↗

Optical Microscopy Techniques to Inspect for Metallic Whiskers

Metal surface finishes of tin, zinc and cadmium are often applied to electronic components, mechanical hardware and other structures. These finishes sometimes unpredictably may form metal whiskers over periods that can take from hours to months or even many years. The metal whiskers are crystalline structures commonly having uniform cross sectional area along their entire length. Typical whisker dimensions are nominally on the order of only a few microns (um) across while their lengths can extend from a few microns to several millimeters. Metal whiskers pose a reliability hazard to electronic systems primarily as an electrical shorting hazard. The extremely narrow dimensions of metal whiskers can make observation with optical techniques very challenging. The videos herein were compiled to demonstrate the complexities associated with optical microscope inspection of electronic and mechanical components and assemblies for the presence or absence of metal whiskers. The importance of magnification, light source and angle of illumination play critical roles in being able to detect metal whiskers when present. Furthermore, it is demonstrated how improper techniques can easily obscure detection. It is hoped that these videos will improve the probability of detecting metal whiskers with optical inspection techniques.

Brusse, Jay A.↗

HERCULES: Strong Patterns towards More Intelligent Predictive Modeling, In: 2014 43rd International Conference on Parallel Processing

Recent work has shown that program analysis techniques to select meaningful code features of programs are important in the task of deciding the best compiler optimizations. Although, there are many successful state-of-the-art program analysis techniques, they often do not provide a simple method to extract the most expressive information about loops, especially when a target program is computationally intensive with complex loops and data dependencies. In this paper, we introduce a static technique to characterize a program using a pattern-driven system named HERCULES. This characterization technique not only helps a user to understand programs by searching pattern-of-interests, but also can be used for a predictive model that effectively selects the proper compiler optimizations. We formulated 35 loop patterns, then evaluated our characterization technique by comparing the predictive models constructed using HERCULES to three other state-of-the-art characterization methods. We show that our models outperform three state-of-the-art program characterization techniques on two multicore systems in selecting the best optimization combination from a given loop transformation space. We achieved up to 67% of the best possible speedup achievable with the optimization search space we evaluated.

Computer Science↗

Risk management technique for liquefied natural gas facilities

Checklists have been compiled for planning, design, construction, startup and debugging, and operation of liquefied natural gas facilities. Lists include references to pertinent safety regulations. Methods described are applicable to handling of other hazardous materials.

Fedor, O. H.↗

Technique for grouping instructions into independent strands

A device compiler and linker is configured to group instructions into different strands for execution by different threads based on the dependence of those instructions on other, long-latency instructions. A thread may execute a strand that includes long-latency instructions, and then hardware resources previously allocated for the execution of that thread may be de-allocated from the thread and re-allocated to another thread. The other thread may then execute another strand while the long-latency instructions are in flight. With this approach, the other thread is not required to wait for the long-latency instructions to complete before acquiring hardware resources and initiating execution of the other strand, thereby eliminating at least a portion of the time that the other thread would otherwise spend waiting.

Mehrara, Mojtaba↗

OpenARC

OpenARC is an open-sourced, very High-Level Intermediate Representation (HLIR)-based, extensible compiler framework, where various performance optimizations, traceability mechanisms, fault tolerance techniques, etc., can be built for better debuggability/performance/resilience on the complex accelerator computing. OpenARC is the first OpenACC compiler supporting Altera FPGAs, in addition to NVIDIA GPUs, AMD GPUs, and Intel Xeon Phis.

Lee, Seyong [Oak Ridge National Laboratory (ORNL),↗

Memory management and compiler support for rapid recovery from failures in computer systems

This paper describes recent developments in the use of memory management and compiler technology to support rapid recovery from failures in computer systems. The techniques described include cache coherence protocols for user transparent checkpointing in multiprocessor systems, compiler-based checkpoint placement, compiler-based code modification for multiple instruction retry, and forward recovery in distributed systems utilizing optimistic execution.

Fuchs, W. K.↗

Chapter 13. Exploring Use of the Reserved Core

In this chapter, we illustrate benefits of thinking in terms of thread management techniques when using a centralized scheduler model along with interoperability of MPI and PThread. This is facilitated through an exploration of thread placement strategies for an algorithm modeling radiative heat transfer with special attention to the 61st core. This algorithm plays a key role within the Uintah Computational Framework (UCF) and current efforts taking place at the University of Utah to model next-generation, large-scale clean coal boilers. In such simulations, this algorithm models the dominant form of heat transfer and consumes a large portion of compute time. Exemplified by a real-world example, this chapter presents our early efforts in porting a key portion of a scalability-centric codebase to the Intel Xeon Phi coprocessor. Specifically, this chapter presents results from our experiments profiling the native execution of a reverse Monte-Carlo ray tracing-based radiation model on a single coprocessor. These results demonstrate that our fastest run configurations utilized the 61st core and that performance was not profoundly impacted when explicitly oversubscribing the coprocessor operating system thread. Additionally, this chapter presents a portion of radiation model source code, a MIC-centric UCF cross-compilation example, and less conventional thread management technique for developers utilizing the PThreads threading model.

97 MATHEMATICS AND COMPUTING↗

The Surface-Topography Challenge: A Multi-Laboratory Benchmark Study to Advance the Characterization of Topography

Surface performance is critically influenced by topography in virtually all real-world applications. The current standard practice is to describe topography using one of a few industry-standard parameters. The most commonly reported number is Ra, the average absolute deviation of the height from the mean line (at some, not necessarily known or specified, lateral length scale). However, other parameters, particularly those that are scale-dependent, influence surface and interfacial properties; for example the local surface slope is critical for visual appearance, friction, and wear. The present Surface-Topography Challenge was launched to raise awareness for the need of a multi-scale description, but also to assess the reliability of different metrology techniques. In the resulting international collaborative effort, 153 scientists and engineers from 64 research groups and companies across 20 countries characterized statistically equivalent samples from two different surfaces: a “rough” and a “smooth” surface. The results of the 2088 measurements constitute the most comprehensive surface description ever compiled. We find wide disagreement across measurements and techniques when the lateral scale of the measurement is ignored. Consensus is established through scale-dependent parameters while removing data that violates an established resolution criterion and deviates from the majority measurements at each length scale. Our findings suggest best practices for characterizing and specifying topography. The public release of the accumulated data and presented analyses enables global reuse for further scientific investigation and benchmarking.

42 ENGINEERING↗

Automatic Offloading C++ Expression Templates to CUDA Enabled GPUs

In the last few years, many scientific applications have been developed for powerful graphics processing units (GPUs) and have achieved remarkable speedups. This success can be partially attributed to high performance host callable GPU library routines that are offloaded to GPUs at runtime. These library routines are based on C/C++-like programming toolkits such as CUDA from NVIDIA and have the same calling signatures as their CPU counterparts. Recently, with the sufficient support of C++ templates from CUDA, the emergence of template libraries have enabled further advancement in code reusability and rapid software development for GPUs. However, Expression Templates (ET), which have been very popular for implementing data parallel scientific software for host CPUs because of their intuitive and mathematics-like syntax, have been underutilized by GPU development libraries. The lack of ET usage is caused by the difficulty of offloading expression templates from hosts to GPUs due to the inability to pass instantiated expressions to GPU kernels as well as the absence of the exact form of the expressions for the templates at the time of coding. This paper presents a general approach that enables automatic offloading of C++ expression templates to CUDA enabled GPUs by using the C++ metaprogramming technique and Just-In-Time (JIT) compilation methodology to generate and compile CUDA kernels for corresponding expression templates followed by executing the kernels with appropriate arguments. This approach allows developers to port applications to run on GPUs with virtually no code modifications. More specifically, this paper uses a large ET based data parallel physics library called QDP++ as an example to illustrate many aspects of the approach to offload expression templates automatically and to demonstrate very good speedups for typical QDP++ applications running on GPUs against running on CPUs using this method of offloading. In addition, this approach of automatic offlo- ding expression templates could be applied to other many-core accelerators that provide C++ programming toolkits with the support of C++ template.

97 MATHEMATICS AND COMPUTING↗