Sandia National Laboratories

Fault-Tolerant Spaceborne Computing Employing New Technologies

Presentations
Fault-Tolerant Spaceborne Workshop Home Page
Zettaflops.org
Computer Science Research Institute
Sandia National Laboratories

Collected 2008 Presentations

Where the author has given permission, the titles for Wedndesday and Thursday talks are hyperlinks to the presentation documents. Where this has not been possible, there is often a parenthetical note explaining why.

Richard Berger BAE Spaceborne Computing Directions Based on New and Emerging Technologies (presentation proprietary)
Abstract: The range of trades that need to be considered for spaceborne operation and methods of radiation hardening will be discussed. BAE Systems' current and upcoming range of products how they come together into a spacecraft architecture example will be presented. Some of the emerging technology areas in which BAE Systems is working at the R&D level will also be addressed.
Lewis Cohn DTRA DTRA Radiation Hardened Microelectronics Program: RH 90nm Technology Development Projects
Abstract: This briefing will address a number of different projects being managed by DTRA to develop and demonstrate radiation hardened 90nm microelectronics technology. The discussion will provide current status on three programs that include: Boeing Radiation Hardened by Design Technology Demonstration Honeywell Radiation Hardening Process Development and Demonstration BAE Systems Radiation Hardening Process Development and Demonstration In addition, a number of other projects that are developing enabling technologies to support the hardening of 90nm technology will be identified.
Steve Crago ISI-East OPERA Software
Abstract: The OPERA software roadmap includes tools that supplement the commercial software chain with tools and libraries important to government customers. Libraries include MPI and VSIPL. The roadmap includes tools that will provide a base for fault tolerant software, including run-time monitor and dynamic resource management software.
G. Cieslewski, A. Jacobs, C. Conger, and A. George UFL Advanced Space Computing with System-Level Fault Tolerance
Abstract: Increasing demand for high-performance computing in space, coupled with limitations of conventional device-level methods for SEU mitigation, are driving innovations in advanced space computing with system-level fault tolerance supporting COTS technologies. This presentation will highlight research activities at the University of Florida from two on-going projects on this path, the NASA Dependable Multiprocessor (NMP ST-8) and the reconfigurable fault tolerance (RFT) framework.
Jamison Collins Intel EXOCHI: Architecture and Programming Environment for a Heterogeneous Multi-core Multithreaded System
Abstract: To maximize performance and power efficiency, future multi-core architectures may be heterogeneous, incorporating some accelerator cores alongside the IA cores. EXOCHI provides the familiar shared virtual memory heterogeneous multi-threaded programming paradigm for these accelerators using novel CPU instruction set extensions and software tool chains with an Intel(r) Architecture (IA) look-n-feel. This talk presents these hardware and software extensions, and describes our experiences and performance improvements when targeting a number of computationally intensive media kernels with EXOCHI.
David Czajkowski Space Micro Rad Hard High Performance Computing
Abstract: Space Micro will discuss our recently developed high performance radiation hardening techniques for modern microprocessors and FPGAs. These techniques have resulted in new computer platforms that are capable of 4,000 MIPS and FPGA computing platforms providing 4 Gbps throughput with 30 million available gates. These platforms are fully radiation hardened.
Erik DeBenedictis Sandia The End of CMOS Scaling Will Be Good for Space Computing
Paul Dodd Sandia Radiation Effects Issues and Trends for High Performance Space Computing
Abstract: In this talk we discuss radiation effects in advanced electronics, including the effects observed in current devices, trends and mitigation techniques for highly scaled technologies, and challenges and opportunities for new technology insertion in spaceborne computing assets.
Jeff Draper ISI Radiation Effects Challenges in 90nm Commercial-Density SRAMs
Abstract: Recent experiments conducted at USC as part of DARPA's Radiation-Hardening-by-Design (RHBD) program show that RHBD techniques can be applied with promising results for sub-100nm technologies. SEE and TID results for two 90nm prototype SRAM devices demonstrate the feasibility of such an approach.
Dan Elftmann Achronix Achronix Reconfigurable High Speed Radiation Hardened FPGA Technology (presentation proprietary)
Abstract: Achronix will discuss the fundamental mechanism of data communication in their picoPIPE based FPGA logic fabric. From here the discussion will move to the Single Event Effects (SEE) mitigation strategy and initial test chip results. This SEE mitigation methodology will be leveraged to develop a High Speed Radiation Hardened FPGA Technology on the BAE Systems RH15 150 nm process technology with a hardened configuration memory based on the BAE Systems hardened SRAM cell.
Joe Fabula Xilinx Current Status of the SIRF Program
Abstract: Xilinx details TID measurements made on standard transistors fabricated in current commercial 90 nm and 65 nm CMOS foundry processes. The performance of some deep sub-micron processes to total ionizing dose exposure was found to be surprisingly robust.
Mitch Fletcher Honeywell Lessons Learned and Ideal Architecture
Stephen Gooch Wind River Unsupervised Multi-OS Solutions for Single, Multi-Core, and Many-Core Processors
Abstract: Unsupervised multi-OS solutions to achieve real-time determinism of a real-time operating system and application support of Linux on the same processor, best of both worlds using two different OSs. Simplified migration of existing software or hardware for when it needs to be combined with new software that targets a different OS or OS version. Replacing separate boards/chips with a single processor reduces system cost and/or boosts performance. Partitioning safely separating applications with different levels of security, reliability, performance, or IP protection is required. It can also be used as an integration model and GPL containment.
Jeff Kalb Sandia NNSA/NA-22 Space Architecture (Government, FFRDC, and SETA participants only, but no prearrangements needed)
Zbigniew Kalbarczyk UIUC Experimental Evaluation: Assessment of Reliability Metrics
Andrew Keys, Michael Brieden, Joe Coughlan NASA MSFC, JSC, ARC

Radiation Hardened Electronics for Space Environments (first part of the presentation)
Processors and Avionics for Support of Human Rated Missions (second part of presentation)

Abstract: The development of a manned flight system requires a different set of considerations than those made during the development of a traditional unmanned space flight system. NASA's Constellation Program is now in the process of developing the next generation of manned spacecraft designed for missions to low earth orbit, to the surface of the Moon, and eventually to Mars. These spacecraft will apply advanced avionics hardware, software and architecture design to meet the stringent requirements of manned spaceflight. In this talk, component, software, and distributed systems issues will be addressed in the context of high reliability, redundancy, fault protection, environmental, complex distributed systems integration, software reliability, and upgradeability over a period of decades as newer technology evolves and is introduced into the system.
Tom Majumder AFRL WPAFB Real-time Wide Area ISR System Development Utilizing Synthetic Aperture Radar (SAR) (presentation export controlled)
Abstract: In this talk, we will present real-time Wide Area ISR System Development Utilizing Synthetic Aperture Radar (SAR). We will discuss current computing technologies such as multicore system, FPGA and multiprocessor that enables real-time SAR data exploitation. We also discuss huge amount of data that SAR system produces and how to process these data in real-time. SAR data exploitation such as Video SAR production utilizing massively parallel multiprocessor system will be illustrated. Sensors' Directorate GOTCHA Radar Exploitation Program (GREP) will be introduced.
Michael Malone Draper Laboratory On-Board Processing Expandable Reconfigurable Architecture (OPERA) Program Overview
Abstract: By purchasing commercial state-of-the-art multicore parallel processor intellectual property and converting it to use radiation hardened by design libraries, the government plans to provide a radiation hardened 90 nm CMOS, 49 core, 70 GOPS, 10 Gbps throughput, TRL 6 general purpose processor with complete intellectual property to the space computing industry by the end of 2010. This presentation will discuss the OPERA hardware architecture and intellectual property rollout plan details.
Ray McConnell ClearSpeed Heterogeneous Computing techniques applied to high performance processing and interconnect fabric for SIMD devices
Abstract: A discussion of the unique heterogeneous computing techniques on ClearSpeed Technology's high performance MTAP and ClearConnectTM interconnect fabric silicon IP
Michael McDougall GrammaTech Verifying Software for Multicore Systems
Abstract: Software running on multi-core systems presents new challenges to software verification. These new challenges will require adapting traditional verification techniques to identify software flaws exposed by the fine-grain parallelism expected in multi-core applications. This talk will discuss the challenges for multi-core verification in general, and focus on how reordered memory operations can invalidate concurrent algorithms previously considered to be correct.
Scott Michel Aerospace Multicore Software And Application Development For Spaceborne Computing
Abstract: Parallel computing architectures, which were actively developed and researched in the 80's and early 90's MPP era, are today's and the foreseeable future's dominant computing platform. As was true during the MPP era, software and application development lags behind architectural innovation. This lag is colloquially known as the "Programmability Gap" in multicore. Numerous efforts are underway to bridge the programmability gap for the various multicore architectural styles, ranging from drop-in replacement libraries, class hierarchies to assist developers to manage threads, to new languages and compiler techniques. This talk introduces the programmability gap, its challenges and a representative spectrum of applied and speculative research, with emphasis on USAF SMC-related spaceborne computing issues.
Kirk Reinholtz JPL Managing Complexity in Next Generation Robotic Spacecraft
Abstract: Future spacecraft avionics hardware will draw on ever larger logic chips and processors that will make it progressively harder for engineers to understand and validate their designs, and likewise, harder to write sound software, resulting in greater system cost and/or risk. This talk describes how complexity fundamentally limits the capabilities of the spacecraft we deliver, and presents some hardware do's and don't do's that will improve our ability to write software that fully utilizes the power of advanced hardware capabilities.
Kirk Reinholtz (for Lorraine Fesq) JPL Preliminary Findings from NASA SMD/PMD Planetary Spacecraft Fault Management Workshop
Abstract: This talk will present preliminary findings of the "NASA Science Mission Directorate, Planetary Science Division's Planetary Spacecraft Fault Management Workshop," which was held on April 14-16, 2008. The Workshop provided a forum for the aerospace community to expose interesting Fault Management case studies from recent missions, to discuss and characterize issues plaguing this discipline, and to view future directions via posters and invited speakers from academia.
Rick Ridgley and Dave Davis Federal Government and AF/SMC R&D Roadmaps to Address Spaceborne Computing (US Citizens only)
Abstract: DoD needs for radiation hardened microelectronics for spaceborne assets are determined and addressed through a formal process established by USD (AT&L). The Director, Defense Research and Engineering (DDR&E) chairs the Radiation Hardened Electronics Oversight Council (RHOC), which validates:
  • the threat;
  • requirements for hardening from the Joint Chiefs of Staff (JCS) and Services;
  • Program Office technology needs;
  • the technology roadmap with milestones and funding levels tied directly to technology need dates for DoD systems.
This talk will address the DoD "corporate management" of radiation hardened microelectronics R&D; the approach used to determine performance needs; the radiation hardness objectives for future R&D programs; space community R&D technology roadmaps to address Program Office needs, and the key metrics for determining the R&D path ahead.
Jack Stocky NASA JPL On-Board Science Data Processing for NASA Science Missions
Abstract: Improved, on-board processing capability can significantly improve the science return from NASA's robotic missions. Hardware and software fault tolerance is necessary to permit the processing capability of COTS hardware to be available for on-board science data processing. The ST 8 Dependable Multiprocessor Validation Experiment is one of NASA's first steps toward realizing this capability.
Ian Troxel Seakr Achieving Fault-Tolerant Spaceborne Computing with Commercial Components
Abstract: Commercial off the shelf (COTS) components have become a means by which cost-effective processing systems are being deployed in spacecraft to meet ever-challenging mission requirements. However, a difficult aspect of infusing COTS devices into aerospace systems is overcoming faults brought about by the harsh radiation environment in which they operate. This presentation will provide an overview of the methods used to achieve the required level of fault tolerance in numerous COTS-based data recorder and onboard processing missions developed at SEAKR.

Contact: Erik DeBenedictis

Top of page