Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “domain compilation”

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 73 records · Page 4

Towards On-Chip Learning for Low Latency Reasoning with End-to-End Synthesis

The Software Defined Architectures (SODA) Synthesizer is an open-source compiler-based tool able to automatically generate domain-specialized systems targeting Application-Specific Integrated Circuits (ASICs) or Field Programmable Gate Arrays (FPGAs) starting from high-level programming. SODA is composed of a frontend, SODA-OPT, which leverages the multilevel intermediate representation (MLIR) framework to interface with productive programming tools (e.g., machine learning frame-works), identify kernels suitable for acceleration, and perform high-level optimizations, and of a state-of-the-art high-level synthesis backend, Bambu from the PandA framework, to generate custom accelerators. One specific application of the SODA Synthesizer is the generation of accelerators to enable ultra-low latency inference and control on autonomous systems for scientific discovery (e.g., electron microscopes, sensors in particle accelerators, etc.). This paper provides an overview of the flow in the context of the generation of accelerators for edge processing to be integrated in transmission electron microscopy (TEM) devices, focusing on use cases from precision material synthesis. We show the tool in action with an example of design space exploration for inference on reconfigurable devices with a conventional deep neural network model (LeNet). Finally, we discuss the research directions and opportunities enabled by SODA in the area of autonomous control for scientific experimental workflows.

Castellana, Vito G.↗

Ada task scheduling: A focused Ada investigation

The types of control that are important for real time task scheduling are discussed. Some closely related real time issues are mentioned and major committee and research activities in this area are delineated. Although there are some problems with Ada and its real time task scheduling, Ada presents fewer than any known alternative. Ada was designed for the domain of real time embedded systems, but Ada compilers may not contain a level of task scheduling support that is adequate for all real time applications. The question addressed is which implementations of Ada's task scheduling are adequate for effective real time systems for NASA applications.

Legrand, Sue↗

Towards Automated Generation of Chiplet-Based Systems

The Software Defined Architectures (SODA) Synthesizer is an open-source compiler-based tool able to automatically generate domain-specialized systems targeting Application- Specific Integrated Circuits (ASICs) or Field Programmable Gate Arrays (FPGAs) starting from high-level programming. SODA is composed of a high-level frontend, SODA-OPT, which leverages the multilevel intermediate representation (MLIR) framework to interface with productive programming tools (e.g., machine learning frameworks), identify kernels suitable for acceleration, and perform high-level optimizations, and of a state-of-the-art high-level synthesis backend, Bambu from the PandA framework, to generate custom accelerators. One specific application of the SODA Synthesizer is the generation of accelerators to enable ultra-low latency inference and control on autonomous systems for scientific discovery (e.g., electron microscopes, sensors in particle accelerators, etc.). This talk will discuss ongoing work on the SODA synthesizer to enable no-human-in-the-loop generation and design space exploration of the chiplets for highly specialized artificial intelligence accelerators. Connecting these highly specialized chiplets to general-purpose cores or programmable accelerators will allow to quickly deploy autonomous systems for scientific discovery.

Limaye, Ankur M.↗

An Ada programming support environment

The toolset of an Ada Programming Support Environment (APSE) being developed at North American Aircraft Operations (NAAO) of Rockwell International, is described. The APSE is resident on three different hosts and must support developments for the hosts and for embedded targets. Tools and developed software must be freely portable between the hosts. The toolset includes the usual editors, compilers, linkers, debuggers, configuration magnagers, and documentation tools. Generally, these are being supplied by the host computer vendors. Other tools, for example, pretty printer, cross referencer, compilation order tool, and management tools were obtained from public-domain sources, are implemented in Ada and are being ported to the hosts. Several tools being implemented in-house are of interest, these include an Ada Design Language processor based on compilable Ada. A Standalone Test Environment Generator facilitates test tool construction and partially automates unit level testing. A Code Auditor/Static Analyzer permits the Ada programs to be evaluated against measures of quality. An Ada Comment Box Generator partially automates generation of header comment boxes.

Tyrrill, AL↗

Really Embedding Domain-Specific Languages into C++

The following topics are dealt with: program compilers; optimising compilers; parallel processing; software engineering; learning (artificial intelligence); multiprocessing systems; shared memory systems; optimisation; computational complexity; specification languages.

Finkel, Hal J.↗

Using MLIR Framework for Codesign of ML Architectures Algorithms and Simulation Tools

MLIR (Multi-Level Intermediate Representation), is an extensible compiler framework that supports high-level data structures and operation constructs. These higher-level code representations are particularly applicable to the artificial intelligence and machine learning (AI/ML) domain, allowing developers to more easily support upcoming heterogeneous AI/ML accelerators and develop flexible domain specific compilers/frameworks with higher-level intermediate representations (IRs) and advanced compiler optimizations. The result of using MLIR within the LLVM compiler framework is expected to yield significant improvement in the quality of generated machine code, which in turn will result in improved performance and hardware efficiency

97 MATHEMATICS AND COMPUTING↗

Deep-freeze graph training for latent learning

Scientific and engineering advances are primarily driven by multi-tier conceptual constructs and conditional theoretical frameworks. The theories allow predictions of hypothetical system responses, given a set of approximate conditions (ranges of applicability) imposed on latent parameters that cannot be measured directly. Learning to estimate the latent variables (Latent Learning) helps to pinpoint the anticipated range-edge anomalies and improves the confidence in interpretation, interpolation and extrapolation of limited experimental data. Due to high dimensionality and extreme non-linearity of the materials science problems, very large datasets are typically required for conventional data-driven model development. The vital experimental data collection, particularly on microstructural phases, is very challenging, which makes it difficult to compile a high-quality database. Incorporation of the domain knowledge into the computational graph structure, initialization and optimization processes presents a viable mechanism for developing accurate models, with limited datasets. Furthermore, this study successfully utilized the approach to build the Deep Freeze Graph (DeepFreG) by mapping known causality relationships and by digitizing empirical domain knowledge for Latent Learning (LL), with specific applications in materials science.

36 MATERIALS SCIENCE↗

A learning apprentice for software parts composition

An overview of the knowledge acquisition component of the Bauhaus, a prototype computer aided software engineering (CASE) workstation for the development of domain-specific automatic programming systems (D-SAPS) is given. D-SAPS use domain knowledge in the refinement of a description of an application program into a compilable implementation. The approach to the construction of D-SAPS was to automate the process of refining a description of a program, expressed in an object-oriented domain language, into a configuration of software parts that implement the behavior of the domain objects.

Allen, Bradley P.↗

Dual-domain column experiment data for aquifer sediments from Riverton, WY, 2018, examining groundwater mobility and speciation of arsenic

This dataset is a compilation of the data that supported the publication of Kumar et al. (2020) in ES&T, http://dx.doi.org/10.1021/acs.est.9b06502. The data is from a dual domain column experiment examining the impact of redox interfaces on arsenic speciation and mobility in natural groundwater sediments that were collected at the Riverton, WY, US Department of Energy Office of Legacy Management field site. The abstract for the associated publication reads:Groundwater contamination by As from natural and anthropogenic sources is a world-wide concern. Redox heterogeneities over space and time are common and can influence the molecular-level speciation of As, and thus As release/retention, but are largely unexplored. Here we present results from a dual-domain column experiment, with natural organic-rich, fine-grained, and sulfidic sediments embedded as lenses (referred to as ‘reducing lenses’) within natural aquifer sand. We show that redox interfaces in sulfur-rich, alkaline aquifers may release concerning levels of As, even when sediment As concentration is low (<2 mg/kg), due to the formation of mobile thioarsenates at aqueous sulfide:Fe molar ratios <1. In our experiments, this behavior occurred in the aquifer sand between reducing lenses and was attributed to the spreading of sulfidic conditions and subsequent Fe reductive dissolution. In contrast, inside reducing lenses (and some locations in the aquifer) the aqueous sulfide:Fe molar ratios exceeded 1 and aqueous sulfide:As molar ratios exceeded 100, which partitioned As(III)-S to the solid phase (associated with organics or as realgar (As4S4)). These results highlight the importance of thioarsenates in natural sediments and indicate that redox interfaces and sediment heterogeneities could locally degrade groundwater quality, even in aquifers with unconcerning solid phase As concentrations.

54 ENVIRONMENTAL SCIENCES↗

In silico Analysis of Transcription Factor Repertoire and Prediction of Stress Responsive Transcription Factors in Soybean

Sequence-specific DNA-binding transcription factors (TFs) are often termed as ‘master regulators’ which bind to DNA and either activate or repress gene transcription. We have computationally analysed the soybean genome sequence data and constructed a proper set of TFs based on the Hidden Markov Model profiles of DNA-binding domain families. Within the soybean genome, we identified 4342 loci encoding 5035 TF models which grouped into 61 families. We constructed a database named SoybeanTFDB (http://soybeantfdb.psc.riken.jp) containing the full compilation of soybean TFs and significant information such as: functional motifs, full-length cDNAs, domain alignments, promoter regions, genomic organization and putative regulatory functions based on annotations of gene ontology (GO) inferred by comparative analysis with Arabidopsis. With particular interest in abiotic stress signalling, we analysed the promoter regions for all of the TF encoding genes as a means to identify abiotic stress responsive cis-elements as well as all types of cis-motifs provided by the PLACE database. SoybeanTFDB enables scientists to easily access cis-element and GO annotations to aid in the prediction of TF function and selection of TFs with functions of interest. This study provides a basic framework and an important user-friendly public information resource which enables analyses of transcriptional regulation in soybean.

59 BASIC BIOLOGICAL SCIENCES↗

Global Transformations for Legacy Parallel Applications via Structural Analysis and Rewriting

Performance and scalability optimization of large HPC applications is currently a labor-intensive, manual process with very low productivity. Major difficulties come from the disaggregated environment for HPC application development: the compiler is only involved in local decisions (core or multithreaded domain), while a library-based, communication-oriented programming model realizes whole-machine parallelism. Realizing any major global change in such a disaggregated environment is very difficult and involves changing large portions of the source code. We present semi-automated techniques, based on structural analysis and rewriting, for performing global transformations on an HPC application source code. We present two case studies using the Self-Consistent Field (SCF) standalone benchmark as well as the Coupled Cluster (CCSD) module (2.9 million lines of Fortran code), a key module of the NWChem computational chemistry application. We demonstrate how structural rewriting techniques can be used to automate transformations that affect multiple sections of the application's source code. We show that the transformations can be applied in a systematic fashion across the source code bases with minimal manual effort. These transformations improve the scalability of the SCF benchmark by more than two orders of magnitude and the performance of the full CCSD module by a factor of 4.0.

legacy parallel applications↗

Testing New Programming Paradigms with NAS Parallel Benchmarks

Over the past decade, high performance computing has evolved rapidly, not only in hardware architectures but also with increasing complexity of real applications. Technologies have been developing to aim at scaling up to thousands of processors on both distributed and shared memory systems. Development of parallel programs on these computers is always a challenging task. Today, writing parallel programs with message passing (e.g. MPI) is the most popular way of achieving scalability and high performance. However, writing message passing programs is difficult and error prone. Recent years new effort has been made in defining new parallel programming paradigms. The best examples are: HPF (based on data parallelism) and OpenMP (based on shared memory parallelism). Both provide simple and clear extensions to sequential programs, thus greatly simplify the tedious tasks encountered in writing message passing programs. HPF is independent of memory hierarchy, however, due to the immaturity of compiler technology its performance is still questionable. Although use of parallel compiler directives is not new, OpenMP offers a portable solution in the shared-memory domain. Another important development involves the tremendous progress in the internet and its associated technology. Although still in its infancy, Java promisses portability in a heterogeneous environment and offers possibility to "compile once and run anywhere." In light of testing these new technologies, we implemented new parallel versions of the NAS Parallel Benchmarks (NPBs) with HPF and OpenMP directives, and extended the work with Java and Java-threads. The purpose of this study is to examine the effectiveness of alternative programming paradigms. NPBs consist of five kernels and three simulated applications that mimic the computation and data movement of large scale computational fluid dynamics (CFD) applications. We started with the serial version included in NPB2.3. Optimization of memory and cache usage was applied to several benchmarks, noticeably BT and SP, resulting in better sequential performance. In order to overcome the lack of an HPF performance model and guide the development of the HPF codes, we employed an empirical performance model for several primitives found in the benchmarks. We encountered a few limitations of HPF, such as lack of supporting the "REDISTRIBUTION" directive and no easy way to handle irregular computation. The parallelization with OpenMP directives was done at the outer-most loop level to achieve the largest granularity. The performance of six HPF and OpenMP benchmarks is compared with their MPI counterparts for the Class-A problem size in the figure in next page. These results were obtained on an SGI Origin2000 (195MHz) with MIPSpro-f77 compiler 7.2.1 for OpenMP and MPI codes and PGI pghpf-2.4.3 compiler with MPI interface for HPF programs.

Jin, H.↗

An MLIR-based Compiler Flow for System-Level Design and Hardware Acceleration

The generation of custom hardware accelerators for applications implemented within high-level productive programming frameworks requires considerable manual effort. To automate this process, we introduce \sodaopt, a compiler tool that extends the MLIR infrastructure. \sodaopt automatically searches, outlines, tiles, and pre-optimizes relevant code regions to generate high-quality accelerators through high-level synthesis. \sodaopt can support any high-level programming framework and domain-specific language that interface with the MLIR infrastructure. By leveraging MLIR, \sodaopt solves compiler optimization problems with specialized abstractions. Backend synthesis tools connect to \sodaopt through progressive intermediate representation lowerings. \sodaopt interfaces to a design space exploration engine to identify the combination of compiler optimization passes and options that provides high-performance generated designs for different backends and targets. We demonstrate the practical applicability of the compilation flow by exploring the automatic generation of accelerators for deep neural networks operators outlined at arbitrary granularity and by combining outlining with tiling on large convolution layers. Experimental results with kernels from the PolyBench benchmark show that \sodaopt high-level optimizations improve execution delays of synthesized accelerators up to 60x. We also show that for the selected kernels, our solution outperforms the current of state-of-the art in more than 70% of the benchmarks and provides better average speedup in 55% of them.

Bohm Agostini, Nicolas↗

ATK Launch Systems Engineering NASA Programs Engineering Examples

This presentation provides an overview of the work done at ATK Launch Systems with and indication of how engineering knowledge can be applied to several real world problems. All material in the presentation has been screened to meet ITAR restrictions. The information provided is a compilation of general engineering knowledge and material available in the public domain. The presentation provides an overview of ATK Launch Systems and NASA programs. Some discussion is provided about the types of engineering conducted at the Promontory plant with added detail about RSRM nozzle engineering. Some brief examples of examples of nozzle technical issues with regard to adhesives and phenolics are shared. These technical issue discussions are based on material available in the public domain.

Richardson, David↗

ROSE

Developed at Lawrence Livermore National Laboratory (LLNL), ROSE is an open source compiler infrastructure to build source-to-source program transformation and analysis tools for large-scale C (C89 to C23), C++ (C++98 to C++23), UPC, Fortran (Fortran4, 66, 77, 95, 2003), OpenMP, Java, Python, and Binary applications. ROSE users range from experienced compiler researchers to library and tool developers who may have minimal compiler experience. ROSE is particularly well suited for building custom tools for static analysis, program optimization, arbitrary program transformation, domain-specific optimizations, complex loop optimizations, performance analysis, and cyber-security. ROSE is: A library (and set of associated tools) to quickly and easily apply compiler techniques to one's code in order to improve application performance and developer productivity. A research and development compiler infrastructure for for writing custom source-to-source translators to perform source code transformations, analysis, and optimizations. Is

Pinnow, NathanT [Lawrence Livermore National Labor↗

Research in mathematical theory of computation

Research progress in the following areas is reviewed: (1) new version of computer program LCF (logic for computable functions) including a facility to search for proofs automatically; (2) the description of the language PASCAL in terms of both LCF and in first order logic; (3) discussion of LISP semantics in LCF and attempt to prove the correctness of the London compilers in a formal way; (4) design of both special purpose and domain independent proving procedures specifically program correctness in mind; (5) design of languages for describing such proof procedures; and (6) the embedding of ideas in the first order checker.

Mccarthy, J.↗

NASA's Integrated Instrument Simulator Suite for Atmospheric Remote Sensing from Spaceborne Platforms (ISSARS) and Its Role for the ACE and GPM Missions

Forward simulation is an indispensable tool for evaluation of precipitation retrieval algorithms as well as for studying snow/ice microphysics and their radiative properties. The main challenge of the implementation arises due to the size of the problem domain. To overcome this hurdle, assumptions need to be made to simplify compiles cloud microphysics. It is important that these assumptions are applied consistently throughout the simulation process. ISSARS addresses this issue by providing a computationally efficient and modular framework that can integrate currently existing models and is also capable of expanding for future development. ISSARS is designed to accommodate the simulation needs of the Aerosol/Clouds/Ecosystems (ACE) mission and the Global Precipitation Measurement (GPM) mission: radars, microwave radiometers, and optical instruments such as lidars and polarimeter. ISSARS's computation is performed in three stages: input reconditioning (IRM), electromagnetic properties (scattering/emission/absorption) calculation (SEAM), and instrument simulation (ISM). The computation is implemented as a web service while its configuration can be accessed through a web-based interface.

real-aperture instruments↗

The Worldwide C3S CORDEX Grand Ensemble: A Major Contribution to Assess Regional Climate Change in the IPCC AR6 Atlas

The collaboration between the Coordinated Regional Climate Downscaling Experiment (CORDEX) and the Earth System Grid Federation (ESGF) provides open access to an unprecedented ensemble of regional climate model (RCM) simulations, across the 14 CORDEX continental-scale domains, with global coverage. These simulations have been used as a new line of evidence to assess regional climate projections in the latest contribution of the Working Group I (WGI) to the IPCC Sixth Assessment Report (AR6), particularly in the regional chapters and the Atlas. Here, we present the work done in the framework of the Copernicus Climate Change Service (C3S) to assemble a consistent worldwide CORDEX grand ensemble, aligned with the deadlines and activities of IPCC AR6. This work addressed the uneven and heterogeneous availability of CORDEX ESGF data by supporting publication in CORDEX domains with few archived simulations and performing quality control. It also addressed the lack of comprehensive documentation by compiling information from all contributing regional models, allowing for an informed use of data. In addition to presenting the worldwide CORDEX dataset, we assess here its consistency for precipitation and temperature by comparing climate change signals in regions with overlapping CORDEX domains, obtaining overall coincident regional climate change signals. The C3S CORDEX dataset has been used for the assessment of regional climate change in the IPCC AR6 (and for the interactive Atlas) and is available through the Copernicus Climate Data Store (CDS).

54 ENVIRONMENTAL SCIENCES↗