Search NASA⌕ Search

SEARCH · Search NASA

Results for “component test capabilities”

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 541 records · Page 30

Performance and analysis of a three-dimensional nonorthogonal laser Doppler anemometer

A three dimensional laser Doppler anemometer with a nonorthogonal third axis coupled by 14 deg was designed and tested. A highly three dimensional flow field of a jet in a crossflow was surveyed to test the three dimensional capability of the instrument. Sample data are presented demonstrating the ability of the 3D LDA to resolve three orthogonal velocity components. Modifications to the optics, signal processing electronics, and data reduction methods are suggested.

Snyder, P. K.↗

Orbital Maneuvering Vehicle (OMV) remote servicing kit

With the design and development of the Orbital Maneuvering Vehicle (OMV) progressing toward an early 1990 initial operating capability (IOC), a new era in remote space operations will evolve. The logical progression to OMV front end kits would make available in situ satellite servicing, repair, and consummables resupply to the satellite community. Several conceptual design study efforts are defining representative kits (propellant tanks, debris recovery, module servicers); additional focus must also be placed on an efficient combination module servicer and consummables resupply kit. A remote servicer kit of this type would be designed to perform many of the early maintenance/resupply tasks in both nominal and high inclination orbits. The kit would have the capability to exchange Orbital Replacement Units (ORUs), exchange propellant tanks, and/or connect fluid transfer umbilicals. Necessary transportation system functions/support could be provided by interfaces with the OMV, Shuttle (STS), or Expendable Launch Vehicle (ELV). Specific remote servicer kit designs, as well as ground and flight demonstrations of servicer technology are necessary to prepare for the potential overwhelming need. Ground test plans should adhere to the component/system/breadboard test philosophy to assure maximum capability of one-g testing. The flight demonstration(s) would most likely be a short duration, Shuttle-bay experiment to validate servicer components requiring a micro-g environment.

Brown, Norman S.↗

Upgraded Near-Field Antenna-Testing System

Upgraded system for near-field testing of large microwave antenna built around commercial automated network analyzer (HP8510) designed for measurement and characterization of microwave circuits and components. Contains highly capable microwave receiver also suitable for far-field and radar-cross-section measurements. Receiver operates in fast-data-acquisition mode at rate of one new data point every millisecond without averaging. Also operates with greater sensitivity by use of averaging feature at time penalty of 0.2 ms per average. Time-domain option added to analyzer enables it to perform time-domain reflectometry.

Kunath, Richard R.↗

Evaluation of Fatigue Crack Growth and Fracture Properties of Cryogenic Model Materials

The criteria used to prevent failure of wind-tunnel models and support hardware were revised as part of a project to enhance the capabilities of cryogenic wind tunnel testing at NASA Langley Research Center. Specifically, damage-tolerance fatigue life prediction methods are now required for critical components, and material selection criteria are more general and based on laboratory test data. The suitability of two candidate model alloys (AerMet 100 and C-250 steel) was investigated by obtaining the fatigue crack growth and fracture data required for a damage-tolerance fatigue life analysis. Finally, an example is presented to illustrate the newly implemented damage tolerance analyses required of wind-tunnel model system components.

Newman, John A.↗

Core Stage TVC Systems Engineering Challenges in Reusing Heritage Hardware

The Space Launch System (SLS) Core Stage (CS) Thrust Vector Control (TVC) system is comprised of 8 mechanical feedback Shuttle heritage Type III TVC actuators and four RS-25 engines, each attached to a Shuttle heritage gimbal block/bearing. Two actuators are used to move each engine in two planes perpendicular to one another (i.e., pitch and yaw). The TVC system design leverages hardware from the Space Shuttle program as well as new hardware designed specifically for the Core Stage. The Space Shuttle heritage hardware directly reused on SLS includes the Orbiter TVC hydraulic servo-actuators (with two slight design modifications), the Orbiter hydraulic circulation pumps, the Orbiter gimbal block/bearing, and the Solid Rocket Booster hydraulic pumps. The Core Auxiliary Power Unit (CAPU) is derived from the Orbiter Auxiliary Power Unit (APU). The Orbiter and Solid Rocket Booster APU turbines are powered by hot gas produced by catalyzed hydrazine decomposition. On the SLS Core Stage, the CAPU turbine is spun using cold gas tapped-off from the RS-25 to CS liquid hydrogen autogenous pressurization line. While direct reuse or slight modification of existing hardware may seem to be a triple-win for a program in cost, schedule, and technical risk mitigation, those benefits can only be realized when its degree of application in a new system is carefully and thoughtfully managed. The heritage hardware reuse should be prescribed within the heritage design capability and reuse environments must lie within the envelope of heritage qualification testing. Despite the significant test and flight experience of the Shuttle heritage hardware components, successful integration with the newly designed CS TVC components and incorporation into the stage design proved to be a challenge which required re-qualification of the heritage hardware as well as thorough integrated testing to support flight certification. Examples of the challenges that were overcome include: re-qualifying heritage hardware to survive new shock and vibration environments, certifying performance of extensively modified heritage hardware, regenerating design insight due to lack of available heritage vendor data, showing compliance to modern structural design standards, translation of heritage requirements for analog avionics to modern digital avionics, and interfacing heritage mechanical hardware with newly designed avionics. This paper is the second installment in a seven-paper series surveying the design, engineering, test validation, and flight performance of the Core Stage Thrust Vector Control system. This paper will discuss several engineering challenges encountered during the development process for SLS CS TVC and how they were successfully overcome to reach flight readiness.

Thrust Vector Control↗

Space Launch System: Core Stage Thrust Vector Control Systems Engineering Challenges in Reusing Heritage Hardware

The Space Launch System (SLS) Core Stage (CS) Thrust Vector Control (TVC) system is comprised of 8 mechanical feedback Shuttle heritage Type III TVC actuators and four RS-25 engines, each attached to a Shuttle heritage gimbal block/bearing. Two actuators are used to move each engine in two planes perpendicular to one another (i.e., pitch and yaw). The TVC system design leverages hardware from the Space Shuttle program as well as new hardware designed specifically for the Core Stage. The Space Shuttle heritage hardware directly reused on SLS includes the Orbiter TVC hydraulic servo-actuators (with two slight design modifications), the Orbiter hydraulic circulation pumps, the Orbiter gimbal block/bearing, and the Solid Rocket Booster hydraulic pumps. The Core Auxiliary Power Unit (CAPU) is derived from the Orbiter Auxiliary Power Unit (APU). The Orbiter and Solid Rocket Booster APU turbines are powered by hot gas produced by catalyzed hydrazine decomposition. On the SLS Core Stage, the CAPU turbine is spun using cold gas tapped-off from the RS-25 to CS liquid hydrogen autogenous pressurization line. While direct reuse or slight modification of existing hardware may seem to be a triple-win for a program in cost, schedule, and technical risk mitigation, those benefits can only be realized when its degree of application in a new system is carefully and thoughtfully managed. The heritage hardware reuse should be prescribed within the heritage design capability and reuse environments must lie within the envelope of heritage qualification testing. Despite the significant test and flight experience of the Shuttle heritage hardware components, successful integration with the newly designed CS TVC components and incorporation into the stage design proved to be a challenge which required re-qualification of the heritage hardware as well as thorough integrated testing to support flight certification. Examples of the challenges that were overcome include: re-qualifying heritage hardware to survive new shock and vibration environments, certifying performance of extensively modified heritage hardware, regenerating design insight due to lack of available heritage vendor data, showing compliance to modern structural design standards, translation of heritage requirements for analog avionics to modern digital avionics, and interfacing heritage mechanical hardware with newly designed avionics. This paper is the second installment in a seven-paper series surveying the design, engineering, test validation, and flight performance of the Core Stage Thrust Vector Control system. This paper will discuss several engineering challenges encountered during the development process for SLS CS TVC and how they were successfully overcome to reach flight readiness.

Thrust Vector Control↗

Environmental Cracking and Irradiation Resistant Stainless Steels by Additive Manufacturing

Metal additive manufacturing (AM), or metal 3D printing is an emergent advanced manufacturing method that can create near net shape geometries directly from computer models. This technology can provide the capability to rapidly fabricate complex parts that may be required to enhance the integrity of reactor internals components. Such opportunities may be observed during a plant refueling outage and AM parts can be rapidly custom designed, manufactured and deployed within the outage interval. Additive manufacturing of stainless steel (SS) components can add business benefits on fast delivery on repair hardware, installation tooling, new design prototypes tests, etc. For the nuclear industry, the supply chain is always an issue for reactor service. AM can provide through-life supply chain (40-60 years) for high-value low-volume components. In the meantime, the capability of generating complex geometries and functional gradient materials will improve the performance, reduce the overall component cost, plant asset management cost and increase the plant reliability by the improvement in materials performance in nuclear environments. While extensive work has been conducted regarding additively manufacturing of austenitic SS parts, most efforts focused only on basic attributes such as porosity, residual stress, basic tensile properties, along with components yield and process monitoring. Little work has been done to define and evaluate the material requirements for nuclear applications. Technical gaps exist, which limit this technology adoption in the nuclear industry, which includes high manufacturing cost, unknown risks, limited nuclear related data, lack of specification and qualification methods, and no prior business experience. The main objective of this program was to generate research data to address all these technical gaps and establish a commercial practice to use AM technology in the nuclear power industry. The detailed objectives are listed as follows: (1) Evaluate nuclear related properties of AM 316L SS, including microstructure, tensile properties, impact toughness, stress corrosion cracking (SCC), corrosion fatigue (CF), irradiation effects, and irradiation assisted stress corrosion cracking (IASCC). (2) Understand the correlations among laser processing, heat treatment, microstructure and SCC/irradiation properties; (3) Optimize and improve the manufacturing process to achieve enhanced nuclear application properties; (4) Fabricate, evaluate, qualify and test a prototype reactor component to demonstrate the commercial viability and cost benefit; (5) Create regulatory approval path and commercialization plans for the production of a commercial reactor component.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Electronic Components and Systems for Cryogenic Space Applications

Electronic components and systems capable of operation at cryogenic temperatures are anticipated in many future NASA space missions such as deep space probes and planetary surface exploration. For example, an unheated interplanetary probe launched to explore the rings of Saturn would reach an average temperature near Saturn of about - 183 C. In addition to surviving the deep space harsh environment, electronics capable of low temperature operation would contribute to improving circuit performance, increasing system efficiency, and reducing payload development and launch costs. Terrestrial applications where components and systems must operate in low temperature environments include cryogenic instrumentation, superconducting magnetic energy storage, magnetic levitation transportation system, and arctic exploration. An on-going research and development program at the NASA Glenn Research Center focuses on the development of reliable electronic devices and efficient power systems capable of surviving in low temperature environments. An overview of the program will be presented in this paper. A description of the low temperature test facilities along with selected data obtained from in-house component testing will also be discussed. Ongoing research activities that are being performed in collaboration with various organizations will also be presented.

Patterson, R. L.↗

A xenon ion propulsion module for enhanced spacecraft capability

Design, development and preliminary testing of a two-engine xenon ion propulsion module for early flight evaluation is described. Extensive use is made of flight spare propellant system components and also engineering model 30-cm ion engines and assemblies originally developed for the Solar Electric Propulsion System program. Significant design features include a redundant propellant feed system that incorporates a novel gas pulse assembly for rapid and completely reliable engine startup, a central neutralizer subsystem with dual neutralizers for redundancy, and major ion engine performance improvements resulting in a nearly doubling of the 30-cm engine thrust in addition to operating on xenon rather than mercury propellant. At a module input power of 10.0 kw, maximum thrust and specific impulse are projected to be 0.4 N and 3,500 sec. respectively, for a total module efficiency of 67 percent. Total mass of the xenon ion module, including the propellant tank, is only 70.2 kg. The technical approaches taken towards developing and integrating the subsystems comprising this ion propulsion module are presented and discussed in detail.

Aston, G.↗

Playing with DIRT: Building the Framework for a Comprehensive In-Situ Soil Materials Testing Database

Long-term, sustainable planetary exploration will require the ability to "live off the land," relying on In-Situ Resource Utilization (ISRU) and In-Situ Construction as core capabilities. Reduction of both risk and launch mass for lunar construction will require evaluating and comparing regolith materials for use as feedstocks suitable for in-situ beneficiation and fabrication of building components. These material assessment capabilities will inform lunar infrastructure design decisions, with co-benefits for terrestrial construction using in-situ materials. Current planetary construction technology development relies on lunar mapping and orbital data, Apollo-era sample analyses, current simulant inventories, and tests conducted using analog site soils. Requirements for lunar infrastructure design decisions and construction systems will be determined based on specific environmental conditions, mission architectures, and the materials available within traverse range of lunar feedstock processing depots. Critical for success is the capability to identify, evaluate and make effective use of a wide range of materials as they are found in-situ on the lunar surface. This paper discusses the development of a digital repository for data on soil and regolith properties, beginning with their structural performance in both cementitious and non-cementitious building material formulations. The Database for In-situ Resource Testing (DIRT) compiles a catalog of raw materials, additives, and formulations, with notations pertaining to material sources and preparation techniques entered via a web-based user interface. Design of consistent data schemas for site-sourced materials evaluation will facilitate linkage with relevant terrestrial and planetary materials databases while enabling guided data input via templates for participation by broader groups of collaborators. Results of these analyses are compiled in a centralized repository to generate insights applicable for regolith resources and landing sites yet to be precisely defined. Collection of thorough records of material characteristics, applications and performance will support innovative construction solutions not only for space infrastructure but for sustainable, resilient design in the terrestrial built environment.

Sarah Joey Seitz↗

ADAPTATION OF CRACK GROWTH DETECTION TECHNIQUES TO US MATERIAL TEST REACTORS

A key component in evaluating the ability of Light Water Reactors to operate beyond 60 years is characterizing the degradation of materials exposed to radiation and various water chemistries. Of particular concern is the response of reactor materials to Irradiation Assisted Stress Corrosion Cracking (IASCC). Some test reactors outside the United States, such as the Halden Boiling Water Reactor (HBWR), have developed techniques to measure crack growth propagation during irradiation. The basic approach is to use a custom-designed compact loading mechanism to stress the specimen during irradiation, while the crack in the specimen is monitored in-situ using the Direct Current Potential Drop (DCPD) method. In 2012 the US Department of Energy commissioned the Idaho National Laboratory and the MIT Nuclear Reactor Laboratory (MIT NRL) to take the basic concepts developed at the HBWR and adapt them to a test rig capable of conducting in-pile IASCC tests in US Material Test Reactors. The first two and half years of the project consisted of designing and testing the loader mechanism, testing individual components of the in-pile rig and electronic support equipment, and autoclave testing of the rig design prior to insertion in the MIT Reactor. The load was applied to the specimen by means of a scissor like mechanism, actuated by a miniature metal bellows driven by pneumatic pressure and sized to fit within the small in-core irradiation volume. In addition to the loader design, technical challenges included developing robust connections to the specimen for the applied current and voltage measurements, appropriate ceramic insulating materials that can endure the LWR environment, dealing with the high electromagnetic noise environment of a reactor core at full power, and accommodating material property changes in the specimen, due primarily to fast neutron damage, which change the specimen resistance without additional crack growth. The project culminated with an in-pile demonstration at the MIT Reactor. The test rig and associated support equipment were used to apply loads to a representative Compact Tensile specimen during one MITR operating cycle, while measuring crack growth using the DCPD method. Although the test period was short (approximately 70 days), and the accumulated neutron dose relatively small, successful operation of the test rig was demonstrated. The specimen was cycled more than 8000 times (more than would be typical for a long term IASCC test), which was sufficient to propagate a crack of over 2 mm.

Crack Growth↗

Application of information technology to the National Launch System

The approach to the development of the Unified Information System (UNIS) to provide in a timely manner all the information required to manage, design, manufacture, integrate, test, launch, operate, and support the Advanced Launch System (NLS), as well as the current and planned capabilities are described. STESYM, the Space Transportation Main Engine (STME) development program, is comprised of a collection of data models which can be grouped into two primary models: the Engine Infrastructure Model (ENGIM) and the Engine Integrated Cast Model (ENGICOM). ENGIM is an end-to-end model of the infrastructure needed to perform the fabrication, assembly, and testing of the STEM program and its components. Together, UNIS and STESYM are to provide NLS managers and engineers with the ability to access various types and files of data quickly and use that data to assess the capabilities of the STEM program.

Mauldin, W. T.↗

DDS-Suite - A Dynamic Data Acquisition, Processing, and Analysis System for Wind Tunnel Testing

Wind Tunnels have optimized their steady-state data systems for acquisition and analysis and even implemented large dynamic-data acquisition systems, however development of near real-time processing and analysis tools for dynamic-data have lagged. DDS-Suite is a set of tools used to acquire, process, and analyze large amounts of dynamic data. Each phase of the testing process: acquisition, processing, and analysis are handled by separate components so that bottlenecks in one phase of the process do not affect the other, leading to a robust system. DDS-Suite is capable of acquiring 672 channels of dynamic data at rate of 275 MB / s. More than 300 channels of the system use 24-bit analog-to-digital cards and are capable of producing data with less than 0.01 of phase difference at 1 kHz. System architecture, design philosophy, and examples of use during NASA Constellation and Fundamental Aerodynamic tests are discussed.

Burnside, Jathan J.↗

Commercial Off-The-Shelf (COTS) Graphics Processing Board (GPB) Radiation Test Evaluation Report

Large round trip communications latency for deep space missions will require more onboard computational capabilities to enable the space vehicle to undertake many tasks that have traditionally been ground-based, mission control responsibilities. As a result, visual display graphics will be required to provide simpler vehicle situational awareness through graphical representations, as well as provide capabilities never before done in a space mission, such as augmented reality for in-flight maintenance or Telepresence activities. These capabilities will require graphics processors and associated support electronic components for high computational graphics processing. In an effort to understand the performance of commercial graphics card electronics operating in the expected radiation environment, a preliminary test was performed on five commercial offthe- shelf (COTS) graphics cards. This paper discusses the preliminary evaluation test results of five COTS graphics processing cards tested to the International Space Station (ISS) low earth orbit radiation environment. Three of the five graphics cards were tested to a total dose of 6000 rads (Si). The test articles, test configuration, preliminary results, and recommendations are discussed.

Salazar, George A.↗

Space Suit Portable Life Support System Test Bed (PLSS 1.0) Development and Testing

A multi-year effort has been carried out at NASA-JSC to develop an advanced extra-vehicular activity Portable Life Support System (PLSS) design intended to further the current state of the art by increasing operational flexibility, reducing consumables, and increasing robustness. Previous efforts have focused on modeling and analyzing the advanced PLSS architecture, as well as developing key enabling technologies. Like the current International Space Station Extra-vehicular Mobility Unit PLSS, the advanced PLSS comprises three subsystems required to sustain the crew during extra-vehicular activity including the Thermal, Ventilation, and Oxygen Subsystems. This multi-year effort has culminated in the construction and operation of PLSS 1.0, a test bed that simulates full functionality of the advanced PLSS design. PLSS 1.0 integrates commercial off the shelf hardware with prototype technology development components, including the primary and secondary oxygen regulators, Ventilation Subsystem fan, Rapid Cycle Amine swingbed carbon dioxide and water vapor removal device, and Spacesuit Water Membrane Evaporator heat rejection device. The overall PLSS 1.0 test objective was to demonstrate the capability of the Advanced PLSS to provide key life support functions including suit pressure regulation, carbon dioxide and water vapor removal, thermal control and contingency purge operations. Supplying oxygen was not one of the specific life support functions because the PLSS 1.0 test was not oxygen rated. Nitrogen was used for the working gas. Additional test objectives were to confirm PLSS technology development components performance within an integrated test bed, identify unexpected system level interactions, and map the PLSS 1.0 performance with respect to key variables such as crewmember metabolic rate and suit pressure. Successful PLSS 1.0 testing completed 168 test points over 44 days of testing and produced a large database of test results that characterize system level and component performance. With the exception of several minor anomalies, the PLSS 1.0 test rig performed as expected; furthermore, many system responses trended in accordance with pre-test predictions.

Watts, Carly↗

Ultra High-Temperature Magnetic Bearing System for s-CO 2 Turbines/Expanders. Phase II Final Report

The goal of this project is to design and construct of a prototype of ultra-high temperature permanent magnet (UHT–PM), which provides the force to pull the spinning shaft towards the target position in the magnetic bearing (MB) clearance circle. This project will develop a PM biased MB system, that is actively cooled to the 550°C maximum operating temperature limit of PMs. The MB test environment will include 700°C, 40krpm and 4,000 psi requirements to advance the supercritical carbon dioxide (sCO 2 ) and other types of Brayton power cycle systems. The proposed novel MBs combine PMs for supplying the bias field and electromagnets for supplying the control field, both in UHT conditions. The choice of material is the UHT Sm-Co PMs (US patent 6,451,132) due to their significantly better thermal stability. The 6-pole homopolar design permits uninterrupted force delivery via a redundant control capability (US patent 7,429,811) even if 3 poles were to fail. The Phase I efforts was to focus on development and demonstration of component technologies and detailed design of full scale components and a demonstrator test rig for use in Phase II. UHT–PMs were produced and used in the MB actuator. The operating environment of the EEC-T550 Sm-Co magnet material could be extended by implementing active cooling. Electromagnetic coils imbedded in the MB provide forces to stabilize the shaft position at the target via sensing and control stages. Shaft position sensors enable to identify the present position of the spinning shaft relative to its desired target position in the MB clearance circle (air gap). Auxiliary bearings will be installed to provide a backup support system in the event that power to the MB fails or there is a fault in a device providing current to the MB coils. For control, reliability and safety reasons the sensors and auxiliary bearings are typically located as close as possible to the MBs, which exposes them to the same environmental conditions. Development of only the actuator for the UHT environment is an inadequate approach to complete integration of the MB system in the turbine/expander. Therefore the UHT sensor and auxiliary bearing were designed and developed along with the actively cooled, UHT-PM actuator development. Although shaft position sensors are commercially available (Kaman) they may cost up to $10k per channel, and a full 5 axis magnetic suspension is best implemented with up to 20 sensors. Once successfully completed, this project will significantly extend the operating environment capability for turbomachinery applications such as nuclear power, concentrated solar thermal, fossil fuel, geothermal, and shipboard propulsion. The further advancement of UHT-PM-MB will also strengthen the US position in PM technology, which is presently heavily dominated by China. The UHT Sm-Co magnet material that will be utilized in the proposed work has been developed exclusively by the project proposer. Many actuators, motors, generators and other MB applications will greatly benefit from the proposed work.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Final Project Report: Composition and Realization of Source-to-Sink High-Performance Flows: File Systems, Storage, Hosts, LAN and WAN

A number of Department of Energy (DOE) science applications, involving exascale computing systems and large experimental facilities, are expected to generate large volumes of data, in the range of petabytes to exabytes, which will be transported over wide-area networks for the purpose of storage, visualization, and analysis. The objectives of this proposal are to (1) develop and test the component technologies and their synthesis methods to achieve source-to-sink high-performance flows, and (2) develop tools that provide these capabilities through simple interfaces to users and applications. In terms of the former, we propose to develop (1) optimization methods that align and transition multiple storage flows to multiple network flows on multicore, multibus hosts; and (2) edge and long-haul network path realization and maintenance using advanced provisioning methods including OSCARS and OpenFlow. We also propose synthesis methods that combine these individual technologies to compose high-performance flows using a collection of constituent storage-network flows, and realize them across the storage and local network connections as well as long-haul connections. We propose to develop automated user tools that profile the hosts, storage systems, and network connections; compose the source-to-sink complex flows; and set up and maintain the needed network connections.

97 MATHEMATICS AND COMPUTING↗

Experiences with SYCL on AMD GPUs with Kokkos

With the recent diversification of the hardware landscape in the high-performance computing (HPC) community, performance-portability solutions are becoming more and more important. One of the most popular choices is Kokkos, which recently became a Linux Foundation project. Most of its development is supported by the US Department of Energy and the French Alternative Energies and Atomic Energy Commission. Kokkos is implemented as a C++ library with multiple backends to support CPUs as well as various GPU architectures. These backends include OpenMP, CUDA, HIP, and also SCYL. This approach enables users to leverage the preferred vendor toolchain for the respective platform (e.g. CUDA, ROCm, OneAPI). The SYCL backend is used to target Intel GPUs, in particular to support the Aurora exascale supercomputer. However, SYCL itself also offers a large degree of portability, and in fact Kokkos’ CI for SYCL has been running on NVIDIA hardware due to a lack of access to Intel GPUs. In this report, we describe our experience with using Kokkos SYCL backend on AMD GPUs targeting the Frontier supercomputer at Oak Ridge National Laboratory. The two major SYCL implementations are DPC++ and AdaptiveCpp. While the Kokkos SYCL backend has been implemented using the former, the latter was the first implementation to target AMD GPUs. We will discuss the experience with both of these SYCL implementations in terms of functionality and performance. Using Kokkos to evaluate SYCL toolchains has a number of benefits. Kokkos’ use of SYCL is fairly complex, exercising features such as graphs, relocatable device functions, atomics – including for non-arithmetic types, as well as pinned and page migratable memory allocations. Kokkos also needs to implement capabilities such as Kokkos’ hierarchical parallelism that are not a straight-forward mapping to SYCL capabilities. Furthermore, a large number of libraries and applications that represent diverse use cases are implemented in Kokkos, providing readily available test cases for a toolchain evaluation. Preliminary results show that support for AMD GPUs in DPC++ is much less mature than for NVIDIA GPUs or Intel GPUs. While the situation has improved significantly over the last year, we still encounter many runtime failures, dispatching problems, and code generation issues. With AdaptiveCpp the challenges arise even earlier in the evaluation process. Since Kokkos’ SYCL implementation is largely focused on supporting Intel GPUs, we opted to leverage SYCL extensions which are available in DPC++ but not in AdaptiveCpp. Furthermore, AdaptiveCpp appears to be less conformant with the SYCL2020 standard which Kokkos relies on. In some cases, we are able to work around the lack of feature support, in other cases we have to disable certain Kokkos capabilities to evaluate the toolchain. Our evaluation will leverage Kokkos’ unit tests to establish basic functionality and feature completeness. We then use simple benchmarks for components of a CG implementation as a measure of usability and performance of the SYCL toolchains.

97 MATHEMATICS AND COMPUTING↗