Search NASA⌕ Search

SEARCH · Search NASA

Results for “computer languages”

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 487 records · Page 27
All sourcesNASA NTRSDOE OSTI
NASA NTRS1979

Interpretive computer simulator for the NASA Standard Spacecraft Computer-2 (NSSC-2)

An Interpretive Computer Simulator (ICS) for the NASA Standard Spacecraft Computer-II (NSSC-II) was developed as a code verification and testing tool for the Annular Suspension and Pointing System (ASPS) project. The simulator is written in the higher level language PASCAL and implented on the CDC CYBER series computer system. It is supported by a metal assembler, a linkage loader for the NSSC-II, and a utility library to meet the application requirements. The architectural design of the NSSC-II is that of an IBM System/360 (S/360) and supports all but four instructions of the S/360 standard instruction set. The structural design of the ICS is described with emphasis on the design differences between it and the NSSC-II hardware. The program flow is diagrammed, with the function of each procedure being defined; the instruction implementation is discussed in broad terms; and the instruction timings used in the ICS are listed. An example of the steps required to process an assembly level language program on the ICS is included. The example illustrates the control cards necessary to assemble, load, and execute assembly language code; the sample program to to be executed; the executable load module produced by the loader; and the resulting output produced by the ICS.

Smith, R. S.↗
DOE OSTI2026

HARMONY: Large-Scale Architecture Search for Efficient Hybrid Language Models

As large language models scale to trillions of parameters, their computational and memory requirements present critical challenges for efficient training and deployment. While Mixture of Experts (MoE) architectures enable efficient scaling through sparse parameter activation, and state-space models like Mamba offer linear-time complexity, principled methods for combining these paradigms remain undeveloped. We introduce HARMONY (Hybrid Architecture Research for Mamba, Optimized with Neural efficiencY), a multi-objective evolutionary neural architecture search framework for discovering efficient hybrid language models that integrate Transformer attention mechanisms, Mixture-of-Experts routing, and Mamba state-space components. Through large-scale distributed search using 16,384 MI250X GPUs on the Frontier supercomputer, HARMONY explores a comprehensive design space encompassing six attention variants (MHA, MQA, GQA, MLA, SWA, and Mamba-2), variable MoE configurations with both routed and shared experts, and extensive Mamba hyperparameters. Our framework discovers heterogeneous architectures that balance training performance with computational efficiency through multi-objective optimization incorporating latency penalties and fitness-based selection. Analysis of discovered architectures reveals that optimal hybrid designs favor heterogeneous component mixing rather than homogeneous patterns, with Mamba-2 and Multi-Head Latent Attention (MLA) emerging as preferred mechanisms. Discovered architectures demonstrate superior training efficiency: our best configuration achieves a final perplexity of 1.0874 with 2.38B parameters while processing 4,320 tokens/second, outperforming significantly larger manually designed models. Full-scale evaluation shows HARMONY's top architectures achieve better loss trajectories than equivalently-sized models using state-of-the-art configurations including Mixtral, Jamba, and Samba. Additionally, we demonstrate 91% weak scaling efficiency when training discovered 36B-parameter models across 1,024 GPUs. HARMONY is released as an open framework with comprehensive tools for building and training hybrid models using expert-data-pipeline parallelism, democratizing access to automated architecture design for next-generation language models.

Herron, Emily [ORNL] (ORCID:0000000273008172)↗
NASA NTRS1974

A computer program for plotting stress-strain data from compression, tension, and torsion tests of materials

A computer program for plotting stress-strain curves obtained from compression and tension tests on rectangular (flat) specimens and circular-cross-section specimens (rods and tubes) and both stress-strain and torque-twist curves obtained from torsion tests on tubes is presented in detail. The program is written in FORTRAN 4 language for the Control Data 6000 series digital computer with the SCOPE 3.0 operating system and requires approximately 110000 octal locations of core storage. The program has the capability of plotting individual strain-gage outputs and/or the average output of several strain gages and the capability of computing the slope of a straight line which provides a least-squares fit to a specified section of the plotted curve. In addition, the program can compute the slope of the stress-strain curve at any point along the curve. The computer program input and output for three sample problems are presented.

Greenbaum, A.↗
DOE OSTI2024

Complexity-calibrated benchmarks for machine learning reveal when prediction algorithms succeed and mislead

Abstract Recurrent neural networks are used to forecast time series in finance, climate, language, and from many other domains. Reservoir computers are a particularly easily trainable form of recurrent neural network. Recently, a “next-generation” reservoir computer was introduced in which the memory trace involves only a finite number of previous symbols. We explore the inherent limitations of finite-past memory traces in this intriguing proposal. A lower bound from Fano’s inequality shows that, on highly non-Markovian processes generated by large probabilistic state machines, next-generation reservoir computers with reasonably long memory traces have an error probability that is at least $$\sim 60\%$$ ∼ 60 % higher than the minimal attainable error probability in predicting the next observation. More generally, it appears that popular recurrent neural networks fall far short of optimally predicting such complex processes. These results highlight the need for a new generation of optimized recurrent neural network architectures. Alongside this finding, we present concentration-of-measure results for randomly-generated but complex processes. One conclusion is that large probabilistic state machines—specifically, large $$\epsilon$$ ϵ -machines—are key to generating challenging and structurally-unbiased stimuli for ground-truthing recurrent neural network architectures.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

ChatHPC: Building the Foundations for a Productive and Trustworthy AI-Assisted HPC Ecosystem

ChatHPC democratizes large language models for the high-performance computing (HPC) community by providing the infrastructure, ecosystem, and knowledge needed to apply modern generative AI technologies to rapidly create specific capabilities for critical HPC components while using relatively modest computational resources. Our divide-and-conquer approach focuses on creating a collection of reliable, highly specialized, and optimized AI assistants for HPC based on the cost-effective and fast Code Llama fine-tuning processes and expert supervision. We target major components of the HPC software stack, including programming models, runtimes, I/O, tooling, and math libraries. Thanks to AI, ChatHPC provides a more productive HPC ecosystem by boosting important tasks related to portability, parallelization, optimization, scalability, and instrumentation, among others. With relatively small datasets (on the order of KB), the AI assistants, which are created in a few minutes by using one node with two NVIDIA H100 GPUs and the ChatHPC library, can create new capabilities with Meta’s 7-billion parameter Code Llama base model to produce high-quality software with a level of trustworthiness of up to 90% higher than the 1.8-trillion parameter OpenAI ChatGPT-4o model for critical programming tasks in the HPC software stack.

Young, Aaron [ORNL] (ORCID:0000000254484667)↗
NASA NTRS1969

Dynamic characteristics of a variable-mass flexible missile: Dynamics of a two-stage variable-mass flexible rocket

The dynamic characteristics of two-stage slender elastic body were investigated. The first stage, containing a solid-fuel rocket, possesses variable mass while the second stage, envisioned as a flexible case, contains packaged instruments of constant mass. The mathematical formulation was in terms of vector equations of motion transformed by a variational principle into sets of scalar differential equations in terms of generalized coordinates. Solutions to the complete equations were obtained numerically by means of finite difference techniques. The problem has been programmed in the FORTRAN 4 language and solved on an IBM 360/50 computer. Results for limited cases are presented showing the nature of the solutions.

Meirovitch, L.↗
NASA NTRS1975

Performance evaluation using SYSTID time domain simulation

This paper reviews the SYSTID technique for performance evaluation of communication systems using time-domain computer simulation. An example program illustrates the language. The inclusion of both Gaussian and impulse noise models make accurate simulation possible in a wide variety of environments. A very flexible postprocessor makes possible accurate and efficient performance evaluation.

Tranter, W. H.↗
NASA NTRS1978

MANART user's guide

This manual contains a comprehensive description of the MANART conversion program. This program is written in XDS Sigma 5 FORTRAN IV language for execution on the Sigma 5 computer. A general introduction is presented to familiarize the reader with the basic concepts of MANART. The program's operation is described. The inputs and outputs of the program are discussed with special emphasis on the control cards. The error messages output by the program are listed, with an explanation of each.

Kuelthau, R. L.↗
NASA NTRS1986

High-Level Data-Abstraction System

Communication with data-base processor flexible and efficient. High Level Data Abstraction (HILDA) system is three-layer system supporting data-abstraction features of Intel data-base processor (DBP). Purpose of HILDA establishment of flexible method of efficiently communicating with DBP. Power of HILDA lies in its extensibility with regard to syntax and semantic changes. HILDA's high-level query language readily modified. Offers powerful potential to computer sites where DBP attached to DEC VAX-series computer. HILDA system written in Pascal and FORTRAN 77 for interactive execution.

Fishwick, P. A.↗
NASA NTRS1986

Image processing methods in two and three dimensions used to animate remotely sensed data

Image processing methods and software used to animate nonimaging remotely sensed data on cloud cover are described. Three FORTRAN programs were written in the VICAR2/TAE image processing domain to perform 3D perspective rendering, to interactively select parameters controlling the projection, and to interpolate parameter sets for animation images between key frames. Operation of the 3D programs and transferring the images to film is automated using executive control language and custom hardware to link the computer and camera.

Hussey, K. J.↗
NASA NTRS1990

Software For Least-Squares And Robust Estimation

GAUSSFIT computer program includes full-featured programming language facilitating creation of mathematical models solving least-squares and robust-estimation problems. Programming language designed to make it easy to specify complex reduction models. Written in 100 percent C language.

Jeffreys, William H.↗
NASA NTRS1990

An innovative on-board processor for lightsats

The Applied Physics Laboratory (APL) has developed a flightworthy custom microprocessor that increases capability and reduces development costs of lightsat science instruments. This device, called the FRISC (FORTH Reduced Instruction Set Computer), directly executes the high-level language called FORTH, which is ideally suited to the multitasking control and data processing environment of a spaceborne instrument processor. The FRISC will be flown as the onboard processor in the Magnetic Field Experiment on the Freja satllite. APL has achieved a significant increase in onboard processing capability with no increase in cost when compared to the magnetometer instrument on Freja's predecessor, the Viking satellite.

Henshaw, R. M.↗
NASA NTRS1990

Software breadboard study

The overall goal of this study was to develop new concepts and technology for the Comet Rendezvous Asteroid Flyby (CRAF), Cassini, and other future deep space missions which maximally conform to the Functional Specification for the NASA X-Band Transponder (NXT), FM513778 (preliminary, revised July 26, 1988). The study is composed of two tasks. The first task was to investigate a new digital signal processing technique which involves the processing of 1-bit samples and has the potential for significant size, mass, power, and electrical performance improvements over conventional analog approaches. The entire X-band receiver tracking loop was simulated on a digital computer using a high-level programming language. Simulations on this 'software breadboard' showed the technique to be well-behaved and a good approximation to its analog predecessor from threshold to strong signal levels in terms of tracking-loop performance, command signal-to-noise ratio and ranging signal-to-noise ratio. The successful completion of this task paves the way for building a hardware breadboard, the recommended next step in confirming this approach is ready for incorporation into flight hardware. The second task in this study was to investigate another technique which provides considerable simplification in the synthesis of the receiver first LO over conventional phase-locked multiplier schemes and in this approach, provides down-conversion for an S-band emergency receive mode without the need of an additional LO. The objective of this study was to develop methodology and models to predict the conversion loss, input RF bandwidth, and output RF bandwidth of a series GaAs FET sampling mixer and to breadboard and test a circuit design suitable for the X and S-band down-conversion applications.

Nuckolls, C.↗
NASA NTRS1994

Program Helps Design Tests Of Developmental Software

Computer program called "A Formal Test Representation Language and Tool for Functional Test Designs" (TRL) provides automatic software tool and formal language used to implement category-partition method and produce specification of test cases in testing phase of development of software. Category-partition method useful in defining input, outputs, and purpose of test-design phase of development and combines benefits of choosing normal cases having error-exposing properties. Traceability maintained quite easily by creating test design for each objective in test plan. Effort to transform test cases into procedures simplified by use of automatic software tool to create cases based on test design. Method enables rapid elimination of undesired test cases from consideration and facilitates review of test designs by peer groups. Written in C language.

Hops, Jonathan↗
NASA NTRS1989

An error-resistant linguistic protocol for air traffic control

The research results described here are intended to enhance the effectiveness of the DATALINK interface that is scheduled by the Federal Aviation Administration (FAA) to be deployed during the 1990's to improve the safety of various aspects of aviation. While voice has a natural appeal as the preferred means of communication both among humans themselves and between humans and machines as the form of communication that people find most convenient, the complexity and flexibility of natural language are problematic, because of the confusions and misunderstandings that can arise as a result of ambiguity, unclear reference, intonation peculiarities, implicit inference, and presupposition. The DATALINK interface will avoid many of these problems by replacing voice with vision and speech with written instructions. This report describes results achieved to date on an on-going research effort to refine the protocol of the DATALINK system so as to avoid many of the linguistic problems that still remain in the visual mode. In particular, a working prototype DATALINK simulator system has been developed consisting of an unambiguous, context-free grammar and parser, based on the current air-traffic-control language and incorporated into a visual display involving simulated touch-screen buttons and three levels of menu screens. The system is written in the C programming language and runs on the Macintosh II computer. After reviewing work already done on the project, new tasks for further development are described.

Cushing, Steven↗