Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “caches”

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 127 records · Page 7

Greatly improved cache update times for conditions data with Frontier/Squid

The CMS detector project loads copies of conditions data to over 100,000 computer cores worldwide by using a software subsystem called Frontier. This subsystem translates database queries into HTTP, looks up the results in a central database at CERN, and caches the results in an industry-standard HTTP proxy/caching server called Squid. One of the most challenging aspects of any cache system is coherency, that is, ensuring that changes made to the underlying data get propagated out to all clients in a timely manner. Recently, the Frontier system was enhanced to drastically reduce the time for changes to be propagated everywhere without heavily loading servers. The propagation time is now as low as 15 minutes for some kinds of data and no more than 60 minutes for the rest of the data. This was accomplished by taking advantage of an HTTP and Squid feature called If-Modified-Since. In order to use this feature, the Frontier server sends a Last-Modified timestamp, but since modification times are not normally tracked by Oracle databases, a PL/SQL program was developed to track the modification times of database tables. We discuss the details of this caching scheme and the obstacles overcome including database and Squid bugs.

99 GENERAL AND MISCELLANEOUS↗

Locality-Driven Dynamic GPU Cache Bypassing

This paper presents novel cache optimizations for massively parallel, throughput-oriented architectures like GPUs. Based on the reuse characteristics of GPU workloads, we propose a design that integrates such efficient locality filtering capability into the decoupled tag store of the existing L1 D-cache through simple and cost-effective hardware extensions.

architecture optimization↗

Large Scale Caching and Streaming of Training Data for Online Deep Learning

The training of deep neural network models on large data remains a difficult problem, despite progress towards scalable techniques. In particular, there is a mismatch between the random but predetermined order in which AI flows select training samples and the streaming I/O patterns for which traditional HPC data storage (e.g., parallel file systems) are designed. In addition, as more data are obtained, it is feasible neither simply to train learning models incrementally, due to catastrophic forgetting (i.e., bias towards new samples), nor to train frequently from scratch, due to prohibitive time and/or resource constraints. In this paper, we study data management techniques that combine caching and streaming with rehearsal support in order to enable efficient access to training samples in both offline training and continual learning. We revisit state-of-art streaming approaches based on data pipelines that transparently handle prefetching, caching, shuffling, and data augmentation, and discuss the challenges and opportunities that arise when combining these methods with data-parallel training techniques. We also report on preliminary experiments that evaluate the I/O overheads involved in accessing the training samples from a parallel file system (PFS) under several concurrency scenarios, highlighting the impact of the PFS on the design of the data pipelines.

data pipelines↗

ServiceX A Distributed, Caching, Columnar Data Delivery Service

We will describe a component of the Intelligent Data Delivery Service being developed in collaboration with IRIS-HEP and the LHC experiments. ServiceX is an experiment-agnostic service to enable on-demand data delivery specifically tailored for nearly-interactive vectorized analysis. This work is motivated by the data engineering challenges posed by HL-LHC data volumes and the increasing popularity of python and Spark-based analysis workflows.ServiceX gives analyzers the ability to query events by dataset metadata. It uses containerized transformations to extract just the data required for the analysis. This operation is colocated with the data to avoid transferring unnecessary branches over the WAN. Simple filtering operations are supported to further reduce the amount of data transferred.Transformed events are cached in a columnar datastore to accelerate delivery of subsequent similar requests. ServiceX will learn commonly related columns and automatically include them in the transformation to increase the potential for cache hits by other users.Selected events are streamed to the analysis system using an efficient wire protocol that can be readily consumed by a variety of computational frameworks. This reduces time-to-insight for physics analysis by delegating to ServiceX the complexity of event selection, slimming, reformatting, and streaming.

Galewsky, B.↗

Security in the CernVM File System and the Frontier Distributed Database Caching System

Both the CernVM File System (CVMFS) and the Frontier Distributed Database Caching System (Frontier) distribute centrally updated data worldwide for LHC experiments using http proxy caches. Neither system provides privacy or access control on reading the data, but both control access to updates of the data and can guarantee the authenticity and integrity of the data transferred to clients over the internet. CVMFS has since its early days required digital signatures and secure hashes on all distributed data, and recently Frontier has added X.509-based authenticity and integrity checking. In this paper we detail and compare the security models of CVMFS and Frontier.

Dykstra, D.↗

XRootd, disk-based, caching proxy for optimization of data access, data placement and data replication

Following the success of the XRootd-based US CMS data federation, the AAA project investigated extensions of the federation architecture by developing two sample implementations of an XRootd, disk-based, caching proxy. The first one simply starts fetching a whole file as soon as a file open request is received and is suitable when completely random file access is expected or it is already known that a whole file be read. The second implementation supports on-demand downloading of partial files. Extensions to the Hadoop Distributed File System have been developed to allow for an immediate fallback to network access when local HDFS storage fails to provide the requested block. Both cache implementations are in pre-production testing at UCSD.

Bauerdick, L. A.T.↗

Cache Locality Optimization for Recursive Programs

We present an approach to optimize the cache locality for recursive programs by dynamically splicing--recursively interleaving--the execution of distinct function invocations. By utilizing data effect annotations, we identify concurrency and data reuse opportunities across function invocations and interleave them to reduce reuse distance. We present algorithms that efficiently track effects in recursive programs, detect interference and dependencies, and interleave execution of function invocations using user-level (non-kernel) lightweight threads. To enable multi-core execution, a program is parallelized using a nested fork/join programming model. Our cache optimization strategy is designed to work in the context of a random work stealing scheduler. We present an implementation using the MIT Cilk framework that demonstrates significant improvements in sequential and parallel performance, competitive with a state-of-the-art compile-time optimizer for loop programs and a domain- specific optimizer for stencil programs.

Lifflander, Jonathan↗

CSU Final Report on the Math/CS Institute CACHE: Communication-Avoiding and Communication-Hiding at the Extreme Scale

The CACHE project entails researching and developing new versions of numerical algorithms that result in data reuse that can be scheduled in a communication avoiding way. Since memory accesses take more time than any computation and require the most power, the focus on turning data reuse into data locality is critical to improving performance and reducing power usage in scientific simulations. This final report summarizes the accomplishments at Colorado State University as part of the CACHE project.

97 MATHEMATICS AND COMPUTING↗

Mitigation of cache memory using an embedded hard-core PPC440 processor in a Virtex-5 Field Programmable Gate Array.

Sandia National Laboratories is currently developing new processing and data communication architectures for use in future satellite payloads. These architectures will leverage the flexibility and performance of state-of-the-art static-random-access-memory-based Field Programmable Gate Arrays (FPGAs). One such FPGA is the radiation-hardened version of the Virtex-5 being developed by Xilinx. However, not all features of this FPGA are being radiation-hardened by design and could still be susceptible to on-orbit upsets. One such feature is the embedded hard-core PPC440 processor. Since this processor is implemented in the FPGA as a hard-core, traditional mitigation approaches such as Triple Modular Redundancy (TMR) are not available to improve the processor's on-orbit reliability. The goal of this work is to investigate techniques that can help mitigate the embedded hard-core PPC440 processor within the Virtex-5 FPGA other than TMR. Implementing various mitigation schemes reliably within the PPC440 offers a powerful reconfigurable computing resource to these node-based processing architectures. This document summarizes the work done on the cache mitigation scheme for the embedded hard-core PPC440 processor within the Virtex-5 FPGAs, and describes in detail the design of the cache mitigation scheme and the testing conducted at the radiation effects facility on the Texas A&M campus.

99 GENERAL AND MISCELLANEOUS↗

Towards Transparent Throughput Elasticity for IaaS Cloud Storage: Exploring the Benefits of Adaptive Block-Level Caching

Storage elasticity on IaaS clouds is a crucial feature in the age of data-intensive computing, especially when considering fluctuations of I/O throughput. This paper provides a transparent solution that automatically boosts I/O bandwidth during peaks for underlying virtual disks, effectively avoiding over-provisioning without performance loss. The authors' proposal relies on the idea of leveraging short-lived virtual disks of better performance characteristics (and thus more expensive) to act during peaks as a caching layer for the persistent virtual disks where the application data is stored. Furthermore, they introduce a performance and cost prediction methodology that can be used both independently to estimate in advance what trade-off between performance and cost is possible, as well as an optimization technique that enables better cache size selection to meet the desired performance level with minimal cost. The authors demonstrate the benefits of their proposal both for microbenchmarks and for two real-life applications using large-scale experiments.

Nicolae, Bogdan↗

An integrated GIS/remote sensing data base in North Cache soil conservation district, Utah: A pilot project for the Utah Department of Agriculture's RIMS (Resource Inventory and Monitoring System)

A basic geographic information system (GIS) for the North Cache Soil Conservation District (SCD) was sought for selected resource problems. Since the resource management issues in the North Cache SCD are very complex, it is not feasible in the initial phase to generate all the physical, socioeconomic, and political baseline data needed for resolving all management issues. A selection of critical varables becomes essential. Thus, there are foud specific objectives: (1) assess resource management needs and determine which resource factors ae most fundamental for building a beginning data base; (2) evaluate the variety of data gathering and analysis techniques for the resource factors selected; (3) incorporate the resulting data into a useful and efficient digital data base; and (4) demonstrate the application of the data base to selected real world resoource management issues.

Wheeler, D. J.↗

Formal verification of an MMU and MMU cache

We describe the formal verification of a hardware subsystem consisting of a memory management unit and a cache. These devices are verified independently and then shown to interact correctly when composed. The MMU authorizes memory requests and translates virtual addresses to real addresses. The cache improves performance by maintaining a LRU (least recently used) list from the memory resident segment table.

Schubert, E. T.↗

Predictive Caching Using the TDAG Algorithm

We describe how the TDAG algorithm for learning to predict symbol sequences can be used to design a predictive cache store. A model of a two-level mass storage system is developed and used to calculate the performance of the cache under various conditions. Experimental simulations provide good confirmation of the model.

Laird, Philip↗

Planetary Sample Caching System Design Options

Potential Mars Sample Return missions would aspire to collect small core and regolith samples using a rover with a sample acquisition tool and sample caching system. Samples would need to be stored in individual sealed tubes in a canister that could be transfered to a Mars ascent vehicle and returned to Earth. A sample handling, encapsulation and containerization system (SHEC) has been developed as part of an integrated system for acquiring and storing core samples for application to future potential MSR and other potential sample return missions. Requirements and design options for the SHEC system were studied and a recommended design concept developed. Two families of solutions were explored: 1)transfer of a raw sample from the tool to the SHEC subsystem and 2)transfer of a tube containing the sample to the SHEC subsystem. The recommended design utilizes sample tool bit change out as the mechanism for transferring tubes to and samples in tubes from the tool. The SHEC subsystem design, called the Bit Changeout Caching(BiCC) design, is intended for operations on a MER class rover.

MSR↗

Sealing Station Mechanisms for the Mars 2020 Rover Sample Caching Subsystem

The Sealing Station is composed of two mechanisms: the Gripper Mechanism, for orientation and support of the Sample Tube during the Hermetic Seal activation activity, as well as ejecting Sample Tubes out of the Rover, and the Ram Mechanism, which produces and applies a high force for seal activation. The Ram Mechanism is a compact high force linear actuator that uses a planetary roller screw (PRS) as the output linear stage, tapered roller bearings (TRB) to support the PRS at the high loads, a set of spur gears supported by cylinder roller bearings, and a brushless motor with a planetary gearhead and magnetic detent brake. This paper describes the role of these mechanisms in the Mars 2020 Rover mission as a part of the Sample Caching Subsystem (SCS) and Adaptive Caching Assembly (ACA) and their interactions with the Hermetic Seal and Sample Tube hardware. We report on development testing performed with the PRS that show it has capability of supporting cross-moment loads, something manufacturers of these components do not rate. Also discussed is the interaction of the passive holding torque from magnetic detent brakes and hardstop preload, setting of a very light preload in the TRB, and the complex linear dyno test equipment used, and how that testing could have been improved.

Grimes-York, Jesse↗

Robotics Verification and Validation Strategies for Perseverance Rover Sampling and Caching

The Mars 2020 Sampling and Caching Subsystem(SCS) is the most complex robotic system ever fielded on a MarsRover. It includes a 5 degree-of-freedom Robotic Arm, coringdrill, gas Dust Removal Tool, interfaces for two turret-mountedinstruments, and an Adaptive Caching Assembly (ACA). TheACA is itself a complex robotic system, containing hardware tosupport docking and bit exchange, a 3 degree-of-freedomSample Handling Assembly for manipulating sample tubes,storage for several drill bits and sample tubes, and mechanismsto support observing and sealing samples collected by the drill.To successfully verify and validate the SCS hardware andsoftware and its integration with the Mars 2020 flight systemseveral key strategies were employed.The SCS Verification and Validation (V&V) program utilizedmultiple test venues with tiered levels of fidelity. These includedsimulation and visualization software environments, low fidelitydevelopment testbeds, testbeds with high fidelity SCS hardwareand commercial off-the shelf avionics, integrated systemtestbeds with flight-like avionics, and environmental testbedscapable of simulating Martian surface temperature andpressure. Multiple units of each SCS hardware componentmoved fluidly between test venues to accomplish myriadstandalone and coordinated test objectives. Test preparationand executions were performed by a diverse team of engineerswith training and technical ownership tailored for individualexperience and role. Despite significant differences between testvenues, the SCS V&V team established efficient and consistentprocesses and tools for procedure development, test execution,and data review that enabled personnel, as well as technicalproducts such as sequences and parameter configurations, toflow between venues effectively. A series of benchmark testsprovided evidence of performance consistency as elements weretransferred between venues and as system capability evolved.This paper provides an overview of the SCS V&V program andexplores several overarching strategies that enabled successfuloperation in the face of unprecedented complexity. Keyoutcomes of the SCS validation effort are summarized, alongwith lessons learned and beneficial integrations of validationtool and process innovations into Mars surface operations.

Brooks, Sawyer↗

Simulating Mars: Enabling Testing of the Perseverance Rover Sampling and Caching Subsystem on Earth

The development of the Sampling and Caching Subsystem (SCS) on the JPL Perseverance Rover lies at the intersection of testing, robotics, and geology. The SCS team established three primary system test campaigns and venues to aid in the development of SCS through verification and validation testing – Qualification Model Dirty Testing (QMDT) to provide a venue for testing in a Martian environment, Vehicle System Testbed (VSTB) for testing while integrated with the mobility subsystem on Martian-like terrain, and the Flight Software Testbed (FSWTB) for conducting tests using the flight motor controllers and software system on a hexapod which had the ability to simulate rover tilt. Each venue contributed a vital piece to the SCS building blocks. However, the QMDT venue operating within a 10-ft diameter Thermal Vacuum chamber to simulate Martian environment provided a sui generis opportunity to fine tune the entire sampling and caching process while building the team’s knowledge base about rock drillability, system life, and target selection. On Earth, because Martian rocks are not readily available, the development team must utilize geoanalogs to the rocks and regolith on Mars. Geologists on the team helped establish a set of standard rock types to use for Mars missions, like Basalt, Sandstone, Mudstone, Gypsum, and other related geoanalogs. These geoanalogs are characterized with a standard suite of tests for density, compressibility, and other characteristics to categorize potential drillability. This concept of drillability is what links the geoanalogs on Earth to the samples we collect on Mars. With the simulant characteristics defined, these geoanalog rocks are ready to be drilled into as we do on the Martian surface. A key aspect of interacting with the surface on Mars is rock target identification and selection. The Perseverance robotic system uses the on-board cameras, instrumentation, and software to collect enough information to identify potential scientific targets. With the targets identified, SCS can place the Corer and abrade the surface or collect a sample. For a ground test activity like QMDT, the test team did not have all of the camera and instrumentation systems that the rover does, so the team developed ground test equivalents to process a rock, build a target map, and define the target. The team constructed a Rock Scanning Station to build a 3D point cloud of the rock. This point cloud was then processed and evaluated with predefined and programmed criteria in a Target Downselect Tool. A primary output of the Target Downselect Tool is a defined target that can be uploaded directly to the robotic software system to simulate and build the robotic sequences used in tests. With these insights and programmatic definition of targets, the QMDT test team was able to make the same decisions that the Perseverance surface operations team does. In addition, valuable lessons learned from developing the target selection ground tools and using them were implemented into the tools used for surface operations.

Kim, Junggon↗

The Use of Proxy Caches for File Access in a Multi-Tier Grid Environment

The use of proxy caches has been extensively studied in the HEP environment for efficient access of database data and showed significant performance with only very moderate operational effort at higher grid tiers (T2, T3). In this contribution we propose to apply the same concept to the area of file access and analyze the possible performance gains, operational impact on site services and applicability to different HEP use cases. Base on a proof-of-concept studies with a modified XROOT proxy server we review the cache efficiency and overheads for access patterns of typical ROOT based analysis programs. We conclude with a discussion of the potential role of this new component at the different tiers of a distributed computing grid.

99 GENERAL AND MISCELLANEOUS↗