Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “performance portable”

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 1,027 records · Page 57

Optical Breath Gas Extravehicular Activity Sensor for the Advanced Portable Life Support System

The function of the infrared gas transducer used during extravehicular activity (EVA) in the current space suit is to measure and report the concentration of carbon dioxide (CO2) in the ventilation loop. The next generation portable life support system (PLSS) requires highly accurate CO2 sensing technology with performance beyond that presently in use on the International Space Station extravehicular mobility unit (EMU). Further, that accuracy needs to be provided over the full operating pressure range of the suit (3 to 25 psia). Accommodation within space suits demands that optical sensors meet stringent size, weight, and power requirements. A laser diode (LD) sensor based on infrared absorption spectroscopy is being developed for this purpose by Vista Photonics, Inc. Version 1.0 prototype devices were delivered to NASA Johnson Space Center (JSC) in September 2011. The prototypes were upgraded with more sophisticated communications and faster response times to version 2.0 and delivered to JSC in July 2012. The sensors incorporate a laser diode based CO2 channel that also includes an incidental water vapor (humidity) measurement. The prototypes are controlled digitally with an field-programmable gate array microcontroller architecture. Based on the results of the iterative instrument development, further prototype development and testing of instruments were performed leveraging the lessons learned where feasible. The present development extends and upgrades the earlier hardware for the advanced PLSS 2.5 prototypes for testing at JSC. The prototypes provide significantly enhanced accuracy for water vapor measurement and eliminate wavelength drift affecting the earlier versions. Various improvements to the electronics and gas sampling are currently being advanced including the companion development of engineering development units that will ultimately be capable of radiation tolerance. The combination of low power electronics with the performance of a long wavelength laser spectrometer enables multi-gas sensors with significantly increased performance over that presently offered in the EMU.

Wood, William R.↗

Performance of the Widely-Used CFD Code OVERFLOW on the Pleides Supercomputer

Computational performance studies were made for NASA's widely used Computational Fluid Dynamics code OVERFLOW on the Pleiades Supercomputer. Two test cases were considered: a full launch vehicle with a grid of 286 million points and a full rotorcraft model with a grid of 614 million points. Computations using up to 8000 cores were run on Sandy Bridge and Ivy Bridge nodes. Performance was monitored using times reported in the day files from the Portable Batch System utility. Results for two grid topologies are presented and compared in detail. Observations and suggestions for future work are made.

Widely Used↗

Portable C++ Code that can Look and Feel Like Fortran Code with Yet Another Kernel Launcher (YAKL)

This paper introduces the Yet Another Kernel Launcher (YAKL) C++ portability library, which strives to enable user-level code with the look and feel of Fortran code. The intended audience includes both C++ developers and Fortran developers unfamiliar with C++. The C++ portability approach is briefly explained, YAKL’s main features are described, and code examples are given that demonstrate YAKL’s usage. YAKL fills a niche capability important particularly to scientific applications seeking to port Fortran code quickly to a portable C++ library. YAKL places heavy emphasis on simplicity, readability, and productivity with performance mainly emphasizing Graphics Processing Units (GPUs). Central to YAKL’s ability to allow Fortran-like user-level code are three features: (1) a multi-dimensional Array class that allows Fortran behavior; (2) a limited library of Fortran intrinsic functions; and (3) an efficient pool allocator that transparently enables cheap frequent allocations and deallocations of YAKL Arrays. While YAKL allows Fortran-style code, it also allows Arrays that exhibit C-like behavior as well, including row-major index ordering and lower bounds of “0”. YAKL currently supports CPUs, CPU threading, and Nvidia, AMD, and Intel GPUs.

97 MATHEMATICS AND COMPUTING↗

A Simple and Portable Multi-Channel Pyrometer allowing temperature measurements down to 800 K on the microsecond scale

The measurement of transient temperatures less than 1000 K for samples in laser-heated diamond anvil cells remains a challenge. Here we present a design and performance characteristics for a multi-channel pyrometer that works in the NIR from 1200 to 2000 nm. It has a relatively small footprint, is portable, only requires low voltage power supplies and can report temperatures down to 800 K on the millisecond scale or faster. A single data point without averaging can be acquired in 14 microseconds (sampling rate of 70'000 times/s). In conjunction with a diamond anvil cell, the system still works down to about 830 K. The pyrometer has been interfaced several times with the laser heating system at the High Pressure Collaborative Access Team (HPCAT) at the Advanced Photon Source in Chicago with good results.

Montgomery, Jeffrey M.↗

Neutron tomography of axially symmetric objects using 14 MeV neutrons from a portable neutron generator

In nuclear boiling water reactor cores, the distribution of water and steam (void) is essential for both safety and efficiency reasons. In order to enhance predictive capabilities, void distribution assessment is performed in two-phase test-loops under reactor-relevant conditions. This article proposes the novel technique of fast-neutron tomography using a portable deuterium-tritium neutron generator to determine the time-averaged void distribution in these loops. Fast neutrons have the advantage of high transmission through the metallic structures and pipes typically concealing a thermal-hydraulic test loop, while still being fairly sensitive to the water/void content. However, commercially available fast-neutron generators also have the disadvantage of a relatively low yield and fast-neutron detection also suffers from relatively low detection efficiency. Fortunately, some loops are axially symmetric, a property which can be exploited to reduce the amount of data needed for tomographic measurement, thus limiting the interrogation time needed. In this article, three axially symmetric test objects depicting a thermal-hydraulic test loop have been examined; steel pipes with outer diameter 24 mm, thickness 1.5 mm, and with three different distributions of the plastic material POM inside the pipes. Data recorded with the FANTOM fast-neutron tomography instrument have been used to perform tomographic reconstructions to assess their radial material distribution. Here, a dedicated tomographic algorithm that exploits the symmetry of these objects has been applied, which is described in the paper. Results are demonstrated in 20 rixel (radial pixel) reconstructions of the interior constitution and 2D visualization of the pipe interior is demonstrated. The local POM attenuation coefficients in the rixels were measured with errors (RMS) of 0.025, 0.020, and 0.022 cm{sup −1}, solid POM attenuation coefficient. The accuracy and precision is high enough to provide a useful indication on the flow mode, and a visualization of the radial material distribution can be obtained. A benefit of this system is its potential to be mounted at any axial height of a two-phase test section without requirements for pre-fabricated entrances or windows. This could mean a significant increase in flexibility of the void distribution assessment capability at many existing two-phase test loops.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

SU-E-T-673: Recent Developments and Comprehensive Validations of a GPU-Based Proton Monte Carlo Simulation Package, GPMC

Purpose: A GPU-based Monte Carlo (MC) simulation package gPMC has been previously developed and high computational efficiency was achieved. This abstract reports our recent improvements on this package in terms of accuracy, functionality, and code portability. Methods: In the latest version of gPMC, nuclear interaction cross section database was updated to include data from TOPAS/Geant4. Inelastic interaction model, particularly the proton scattering angle distribution, was updated to improve overall simulation accuracy. Calculation of dose averaged LET (LETd) was implemented. gPMC was ported onto an OpenCL environment to enable portability across different computing devices (GPUs from different vendors and CPUs). We also performed comprehensive tests of the code accuracy. Dose from electro-magnetic (EM) interaction channel, primary and secondary proton doses and fluences were scored and compared with those computed by TOPAS. Results: In a homogeneous water phantom with 100 and 200 MeV beams, mean dose differences in EM channel computed by gPMC and by TOPAS were 0.28% and 0.65% of the corresponding maximum dose, respectively. With the Geant4 nuclear interaction cross section data, mean difference of primary proton dose was 0.84% for the 200 MeV case and 0.78% for the 100 MeV case. After updating inelastic interaction model, maximum difference of secondary proton fluence and dose were 0.08% and 0.5% for the 200 MeV beam, and 0.04% and 0.2% for the 100 MeV beam. In a test case with a 150MeV proton beam, the mean difference between LETd computed by gPMC and TOPAS was 0.96% within the proton range. With the OpenCL implementation, gPMC is executable on AMD and Nvidia GPUs, as well as on Intel CPU in single or multiple threads. Results on different platforms agreed within statistical uncertainty. Conclusion: Several improvements have been implemented in the latest version of gPMC, which enhanced its accuracy, functionality, and code portability.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Optimizing Irregular Communication with Neighborhood Collectives and Locality-Aware Parallelism

Irregular communication often limits both the performance and scalability of parallel applications. Typically, applications individually implement irregular communication as point-to-point, and any optimizations are integrated directly into the application. As a result, these optimizations lack portability. It is difficult to optimize point-to-point messages within MPI, as the interface for single messages provides no information on the collection of all communication to be performed. However, the persistent neighbor collective API, released in the MPI 4 standard, provides an interface for portable optimizations of irregular communication within MPI libraries. This paper presents methods for implementing existing optimizations for irregular communication within neighborhood collectives, analyzes the impact of replacing point-to-point communication in existing codebases such as Hypre BoomerAMG with neighborhood collectives, and finally shows up to a 1.38x speedup on sparse matrix-vector multiplication communication within a BoomerAMG solve through the use of our optimized neighbor collectives. Here, the authors analyze three implementations of persistent neighborhood collectives for Alltoallv: an unoptimized wrapper of standard point-to-point communication, and two locality-aware aggregating methods. The second locality-aware implementation exposes an non-standard interface to perform additional optimization, and the authors present the additional 0.07x speedup from the extended interface. All optimizations are available in an open-source codebase, MPI Advance, which sits on top of MPI, allowing for optimizations to be added into existing codebases regardless of the system MPI install.

AMG↗

Extending SEER for Extreme Heterogeneity

Heterogeneous and multi-device nodes are increasingly common in high-performance computing and data centers, yet existing programming models often lack simple, transparent, and portable support for these diverse architectures. The main contribution of this work is the development of novel SEER capabilities to address this challenge by providing a descriptive programming model that allows applications to seamlessly leverage heterogeneous nodes across various device types. SEER uses efficient memory management and can select the proper device[s] depending on the computational cost of the applications. This is completely transparent to the programmer, thereby providing a highly productive programming environment. Integrating extreme heterogeneity into the SEER library as shown with the use of NVIDIA and AMD GPUs simultaneously allows it to expand and exploit the performance possibilities. Our analysis based on the well-known Conjugate Gradient algorithm reports accelerations above 1.5 × on computationally demanding steps of such an algorithm by using both architectures simultaneously.

Teranishi, Keita [ORNL] (ORCID:0000000166472690)↗

Evaluating Mass Analyzers as Candidates for Small, Portable, Rugged Single Point Mass Spectrometers for Analysis of Permanent Gases

For Space Shuttle launch safety, there is a need to monitor the concentration of H2, He, O2 and Ar around the launch vehicle. Currently a large mass spectrometry system performs this task, using long transport lines to draw in samples. There is great interest in replacing this stationary system with several miniature, portable, rugged mass spectrometers which act as point sensors which can be placed at the sampling point. Five commercial and two non-commercial analyzers are evaluated. The five commercial systems include the Leybold Inficon XPR-2 linear quadrupole, the Stanford Research (SRS-100) linear quadrupole, the Ferran linear quadrupole array, the ThermoQuest Polaris-Q quadrupole ion trap, and the IonWerks Time-of-Flight (TOF). The non-commercial systems include a compact double focusing sector (CDFMS) developed at the University of Minnesota, and a quadrupole ion trap (UF-IT) developed at the University of Florida. The System Volume is determined by measuring the entire system volume including the mass analyzer, its associated electronics, the associated vacuum system, the high vacuum pump and rough pump. Also measured are any ion gauge controllers or other required equipment. Computers are not included. Scan Time is the time required for one scan to be acquired and the data to be transferred. It is determined by measuring the time required acquiring a known number of scans and dividing by said number of scans. Limit of Detection is determined first by performing a zero-span calibration (using a 10-point data set). Then the limit of detection (LOD) is defined as 3 times the standard deviation of the zero data set. (An LOD of 10 ppm or less is considered acceptable.)

Arkin, C. Richard↗

Raman spectroscopy coupled with ambient ionization mass spectrometry: A forensic laboratory investigation into rapid and simple dual instrumental analysis techniques

We present a laboratory experiment, developed around a pair of instrumental analysis techniques, Raman spectroscopy and ambient ionization mass spectrometry (MS), conducted by senior chemistry undergraduate students for bulk analysis of over-the-counter drugs using benchtop versions of the two instruments, as well as trace analysis of illicit drugs utilizing the corresponding portable instruments. The identification and confirmation of seized drugs is a routine operation in forensic laboratories that can be simplified using these two instrumental techniques in tandem. Bulk samples are readily analyzed using conventional Raman spectroscopy followed by paper cone spray ionization (PCSI) MS, while trace analysis is allowed by a simple single-substrate dual-analyzer method using surface enhanced Raman spectroscopy (SERS) from paper followed by paper spray ionization (PSI) MS. The introduction of these novel strategies to the students exposed them to potential career paths in forensic science by means of a hands-on experience that provided fundamental knowledge about this set of analytical methods, including intercomparisons of their performance, as well as practical issues like costs, regulations and databases. The use of portable instrumentation represents an additional learning experience for the students, considering that it is cutting-edge technology that is just now being transferred from academic to industrial labs and that it allows a valuable discussion of the advantages and disadvantages of bringing the laboratory to the field.

47 OTHER INSTRUMENTATION↗

CommBench: Micro-Benchmarking Hierarchical Networks with Multi-GPU, Multi-NIC Nodes

Modern high-performance computing systems have multiple GPUs and network interface cards (NICs) per node. The resulting network architectures have multilevel hierarchies of subnetworks with different interconnect and software technologies. These systems offer multiple vendor-provided communication capabilities and library implementations (IPC, MPI, NCCL, RCCL, OneCCL) with APIs providing varying levels of performance across the different levels. Understanding this performance is currently difficult because of the wide range of architectures and programming models (CUDA, HIP, OneAPI). We present CommBench, a library with cross-system portability and a high-level API that enables developers to easily build microbenchmarks relevant to their use cases and gain insight into the performance (bandwidth & latency) of multiple implementation libraries on different networks. We demonstrate CommBench with three sets of microbenchmarks that profile the performance of six systems. Our experimental results reveal the effect of multiple NICs on optimizing the bandwidth across nodes and also present the performance characteristics of four available communication libraries within and across nodes of NVIDIA, AMD, and Intel GPU networks.

Hidayetoglu, Mert↗

Numerical eigen-spectrum slicing, accurate orthogonal eigen-basis, and mixed-precision eigenvalue refinement using OpenMP data-dependent tasks and accelerator offload

Performing a variety of numerical computations efficiently and, at the same time, in a portable fashion requires both an overarching design followed by a number of implementation strategies. All of these are exemplified below as we present transitioning the PLASMA numerical library from relying on dependence-driven large tasks to achieving utilization of fine grain tasking and offload to hardware accelerators while keeping its core dependence sets: OpenMP source code pragmas and runtime for most system-level functionality and basic low-level numerical kernels provided directly by hardware vendors or open source projects with vendor contributions. We also present new algorithmic methods and their efficient parallel implementations including fine grained tasking for eigen-spectrum slicing and offload for mixed-precision eigenvalue refinement. We provide performance, scaling, and numerical results showing sizable gains over the available solutions from either the open source and vendor-provided packages.

Luszczek, Piotr↗

Evaluating Mass Analyzers as Candidates for Small, Portable, Rugged Single Point Mass Spectrometers for Analysis of Permanent Gases

For Space Shuttle launch safety, there is a need to monitor the concentration Of H2, He, O2, and Ar around the launch vehicle. Currently a large mass spectrometry system performs this task, using long transport lines to draw in samples. There is great interest in replacing this stationary system with several miniature, portable, rugged mass spectrometers which act as point sensors which can be placed at the sampling point. Five commercial and two non-commercial analyzers are evaluated. The five commercial systems include the Leybold Inficon XPR-2 linear quadrupole, the Stanford Research (SRS-100) linear quadrupole, the Ferran linear quadrupole array, the ThermoQuest Polaris-Q quadrupole ion trap, and the IonWerks Time-of-Flight (TOF). The non-commercial systems include a compact double focusing sector (CDFMS) developed at the University of Minnesota, and a quadrupole ion trap (UF-IT) developed at the University of Florida.

Arkin, C. Richard↗

...And Eat it Too: High Read Performance in Write-Optimized HPC I/O Middleware File Formats

As HPC applications run on increasingly high process counts on larger and larger machines, both the frequency of checkpoints needed for fault tolerance and the resolution and size of Data Analysis Dumps are expected to increase proportionally. In order to maintain an acceptable ratio of time spent performing useful computation work to time spent performing I/O, write bandwidth to the underlying storage system must increase proportionally to this increase in the checkpoint and computation size. Unfortunately, popular scientific self-describing file formats such as netCDF and HDF5 are designed with a focus on portability and flexibility. Extra care and careful crafting of the output structure and API calls is required to optimize for write performance using these APIs. To provide sufficient write bandwidth to continue to support the demands of scientific applications, the HPC community has developed a number of I/O middleware layers, that structure output into write-optimized file formats. However, the obvious concern with any write optimized file format would be a corresponding penalty on reads. In the log-structured filesystem, for example, a file generated by random writes could be written efficiently, but reading the file back sequentially later would result in very poor performance. Simulation results require efficient read-back for visualization and analytics, and though most checkpoint files are never used, the efficiency of a restart is very important in the face of inevitable failures. The utility of write speed improving middleware would be greatly diminished if it sacrificed acceptable read performance. In this paper we examine the read performance of two write-optimized middleware layers on large parallel machines and compare it to reading data natively in popular file formats. The two I/O middleware layer examined in this paper are the Adaptable IO System (ADIOS), a library-based approach developed at Oak Ridge National Laboratory to provide a high-level IO API that can be used in place of netCDF or HDF5 to do much more aggressive write-behind and efficient reordering of data locations within the file; and the Parallel Log-structured Filesystem (PLFS), a stackable FUSE filesystem approach developed at Los Alamos National Laboratory that decouples concurrent writes to improve the speed of checkpoints. Since ADIOS is an I/O componentization that affords selection of different I/O methods at or during runtime, through a single API, users can have access to MPI-IO, Posix-IO, HDF5, netCDF, and staging methods. The ADIOS BP file format is a new log-file format that has a superset of the features of both HDF5 and netCDF, but is designed to be portable and flexible while being optimized for writing. PLFS, on the other hand, is mounted as a stackable filesystem on top of an existing parallel filesystem. Reads or writes to the PLFS filesystem are transparently translated into operations on per-process log files stored in the underlying parallel filesystem. Since PLFS performs this translation without application modification, users can write in HDF5, netCDF, or app-specific file formats and PLFS will store the writes in a set of efficiently written log-formatted files, while presenting the user with a logical 'flat' file on reads. Despite their different approaches, the commonality behind both of these middleware systems is that they both write to a log file format. As shown in previous publications, writes are fully optimized in both systems, sometimes resulting in 100x improvements over writing data in popular file formats. But as mentioned above, writing is only one part of the story. In this paper we examine the read performance of our middleware layers on large parallel machines and compare these to reading data either natively or from other popular file formats. We compare the reading performance in two different scenarios: (1) Reading back restarts from the same number of processors as wrote the data and (2) Reading back restart data from a different number of processors. We observe that not only can write-optimized I/O middleware be built to not penalize read speeds, but for important workloads, techniques that improve write performance can, perhaps counterintuitively, improve read speeds over reading from a contiguously organized file format. In the remainder of this paper, we investigate this further through case studies of PLFS and ADIOS on simulation checkpoint restarts.

99 GENERAL AND MISCELLANEOUS↗

Resilience Design Patterns - A Structured Approach to Resilience at Extreme Scale (version 1.0)

Reliability is a serious concern for future extreme-scale high-performance computing (HPC) systems. Projections based on the current generation of HPC systems and technology roadmaps suggest that very high fault rates in future systems. The errors resulting from these faults will propagate and generate various kinds of failures, which may result in outcomes ranging from result corruptions to catastrophic application crashes. Practical limits on power consumption in HPC systems will require future systems to embrace innovative architectures, increasing the levels of hardware and software complexities. The resilience challenge for extreme-scale HPC systems requires management of various hardware and software technologies that are capable of handling a broad set of fault models at accelerated fault rates. These techniques must seek to improve resilience at reasonable overheads to power consumption and performance. While the HPC community has developed various solutions, application-level as well as system-based solutions, the solution space of HPC resilience techniques remains fragmented. There are no formal methods and metrics to investigate and evaluate resilience holistically in HPC systems that consider impact scope, handling coverage, and performance & power eciency across the system stack. Additionally, few of the current approaches are portable to newer architectures and software ecosystems, which are expected to be deployed on future systems. In this document, we develop a structured approach to the management of HPC resilience based on the concept of resilience-based design patterns. A design pattern is a general repeatable solution to a commonly occurring problem. We identify the commonly occurring problems and solutions used to deal with faults, errors and failures in HPC systems. The catalog of resilience design patterns provides designers with reusable design elements. We define a design framework that enhances our understanding of the important constraints and opportunities for solutions deployed at various layers of the system stack. The framework may be used to establish mechanisms and interfaces to coordinate flexible fault management across hardware and software components. The framework also enables optimization of the cost-benefit trade-os among performance, resilience, and power consumption. The overall goal of this work is to enable a systematic methodology for the design and evaluation of resilience technologies in extreme-scale HPC systems that keep scientific applications running to a correct solution in a timely and cost-ecient manner in spite of frequent faults, errors, and failures of various types.

97 MATHEMATICS AND COMPUTING↗

Effectiveness of Cool Roof Coatings with Ceramic Particles

Liquid applied coatings promoted as cool roof coatings, including several with ceramic particles, were tested at Oak Ridge National Laboratory (ORNL), Oak Ridge, Tenn., for the purpose of quantifying their thermal performances. Solar reflectance measurements were made for new samples and aged samples using a portable reflectometer (ASTM C1549, Standard Test Method for Determination of Solar Reflectance Near Ambient Temperature Using a Portable Solar Reflectometer) and for new samples using the integrating spheres method (ASTM E903, Standard Test Method for Solar Absorptance, Reflectance, and Transmittance of Materials Using Integrating Spheres). Thermal emittance was measured for the new samples using a portable emissometer (ASTM C1371, Standard Test Method for Determination of Emittance of Materials Near Room 1 Proceedings of the 2011 International Roofing Symposium Temperature Using Portable Emissometers). Thermal conductivity of the coatings was measured using a FOX 304 heat flow meter (ASTM C518, Standard Test Method for Steady-State Thermal Transmission Properties by Means of the Heat Flow Meter Apparatus). The surface properties of the cool roof coatings had higher solar reflectance than the reference black and white material, but there were no significant differences among coatings with and without ceramics. The coatings were applied to EPDM (ethylene propylene diene monomer) membranes and installed on the Roof Thermal Research Apparatus (RTRA), an instrumented facility at ORNL for testing roofs. Roof temperatures and heat flux through the roof were obtained for a year of exposure in east Tennessee. The field tests showed significant reduction in cooling required compared with the black reference roof (~80 percent) and a modest reduction in cooling compared with the white reference roof (~33 percent). The coating material with the highest solar reflectivity (no ceramic particles) demonstrated the best overall thermal performance (combination of reducing the cooling load cost and not incurring a large heating penalty cost) and suggests solar reflectivity is the significant characteristic for selecting cool roof coatings.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Challenges and Opportunities of User-Level File Systemsfor HPC

The performance gap between magnetic disks and data processing on HPC systems has become that huge that an efficient data processing can only be achieved by introducing non-volatile memory (NVRAM) as a new storage tier. Although the benefits of hierarchical storage have been adequately demonstrated to the point that the newest leadership class HPC systems will employ burst buffers, critical questions remain for supporting hierarchical storage systems, including: How should we present hierarchical storage systems to user applications, such that they are easy to use and that application code is portable across systems? How should we manage data movement through a storage hierarchy for best performance and resilience of data? How do the particular I/O use cases mandate the way we manage data? There have been many efforts to explore this space in the form of file systems, with increasingly more implemented at the user level. This is because it is relatively easy to swap in new, specialized user-level file systems for use by applications on a case-by-case basis, as opposed to the current mainstream approach of using general-purpose, system-level file systems which may not be optimized for HPC workloads and must be installed by administrators. In contrast, file systems at the user level can be tailored for specific HPC workloads for high performance and can be used by applications without administrator intervention. Many user-level file system developers have found themselves “having to reinvent the wheel” to implement various optimizations in their file systems. Thus, a main goal of this meeting was to bring together experts in I/O performance, file systems, and storage, and collectively explore the space of current and future problems and solutions for I/O on hierarchical storage systems in order to begin a community effort in enabling user-level file system support for HPC systems. We had a lively week of learning about each other’s approaches as well as unique I/O use cases that can influence the design of a community-driven file and storage system standards. The agenda for this meeting contained talks from participants on the following high level topics: HPC storage and I/O support today; what do HPC users need for I/O; existing userlevel file system efforts; object stores and other alternative storage systems; and components for building user-level file systems. The talks were short and intended to be conversation starters for more in-depth discussions with the whole group. The participants engaged in lengthy discussions on various questions that arose from the talks including: Are we ready to program to a memory hierarchy versus block devices? Are the needs of HPC users reflected in our existing file systems and storage systems? Should we drop or keep POSIX moving forward? What do we mean when we say “user-level file system”? Do we all mean the same thing? How should the IO 500 benchmark be defined so it is fair and useful? and How are stage-in and stage-out actually going to work? The report for this seminar contains a record of the talks from the participants as well as the resulting discussions. Our hope is that the effort initiated during this seminar will result in long-term collaborations that will benefit the HPC community as a whole.

97 MATHEMATICS AND COMPUTING↗

Resilience Design Patterns - A Structured Approach to Resilience at Extreme Scale (version 1.1)

Reliability is a serious concern for future extreme-scale high-performance computing (HPC) systems. Projections based on the current generation of HPC systems and technology roadmaps suggest the prevalence of very high fault rates in future systems. The errors resulting from these faults will propagate and generate various kinds of failures, which may result in outcomes ranging from result corruptions to catastrophic application crashes. Therefore the resilience challenge for extreme-scale HPC systems requires management of various hardware and software technologies that are capable of handling a broad set of fault models at accelerated fault rates. Also, due to practical limits on power consumption in HPC systems future systems are likely to embrace innovative architectures, increasing the levels of hardware and software complexities. As a result the techniques that seek to improve resilience must navigate the complex trade-off space between resilience and the overheads to power consumption and performance. While the HPC community has developed various resilience solutions, application-level techniques as well as system-based solutions, the solution space of HPC resilience techniques remains fragmented. There are no formal methods and metrics to investigate and evaluate resilience holistically in HPC systems that consider impact scope, handling coverage, and performance & power efficiency across the system stack. Additionally, few of the current approaches are portable to newer architectures and software environments that will be deployed on future systems. In this document, we develop a structured approach to the management of HPC resilience using the concept of resilience-based design patterns. A design pattern is a general repeatable solution to a commonly occurring problem. We identify the commonly occurring problems and solutions used to deal with faults, errors and failures in HPC systems. Each established solution is described in the form of a pattern that addresses concrete problems in the design of resilient systems. The complete catalog of resilience design patterns provides designers with reusable design elements. We also define a framework that enhances a designer's understanding of the important constraints and opportunities for the design patterns to be implemented and deployed at various layers of the system stack. This design framework may be used to establish mechanisms and interfaces to coordinate flexible fault management across hardware and software components. The framework also supports optimization of the cost-benefit trade-offs among performance, resilience, and power consumption. The overall goal of this work is to enable a systematic methodology for the design and evaluation of resilience technologies in extreme-scale HPC systems that keep scientific applications running to a correct solution in a timely and cost-efficient manner in spite of frequent faults, errors, and failures of various types.

97 MATHEMATICS AND COMPUTING↗