Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “BURST BUFFER”

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 91 records · Page 5

Tuning HDF5 subfiling performance on parallel file systems

Subfiling is a technique used on parallel file systems to reduce locking and contention issues when multiple compute nodes interact with the same storage target node. Subfiling provides a compromise between the single shared file approach that instigates the lock contention problems on parallel file systems and having one file per process, which results in generating a massive and unmanageable number of files. In this paper, we evaluate and tune the performance of recently implemented subfiling feature in HDF5. In specific, we explain the implementation strategy of subfiling feature in HDF5, provide examples of using the feature, and evaluate and tune parallel I/O performance of this feature with parallel file systems of the Cray XC40 system at NERSC (Cori) that include a burst buffer storage and a Lustre disk-based storage. We also evaluate I/O performance on the Cray XC30 system, Edison, at NERSC. Our results show performance benefits of 1.2X to 6X performance advantage with subfiling compared to writing a single shared HDF5 file. We present our exploration of configurations, such as the number of subfiles and the number of Lustre storage targets to storing files, as optimization parameters to obtain superior I/O performance. Based on this exploration, we discuss recommendations for achieving good I/O performance as well as limitations with using the subfiling feature.

97 MATHEMATICS AND COMPUTING↗

Computing erasure metadata and data layout prior to storage using a processing platform

Techniques are provided for computing data and metadata layout prior to storage in a storage system using a processing platform. An exemplary processing platform comprises one or more of a compute node and a burst buffer appliance. The processing platform communicates with a plurality of the compute nodes over a network, wherein a plurality of applications executing on the plurality of compute nodes generate a plurality of data objects; computes erasure metadata for one or more of the data objects on at least one of the compute nodes; and provides the erasure metadata with the corresponding one or more data objects to a storage system. The processing platform optionally determines a full set of the data objects to be stored and queries the storage system to determine an anticipated layout of the full set of the data objects to be stored. The anticipated layout allows special handling, for example, for small files and large files that are identified based on predefined criteria.

96 KNOWLEDGE MANAGEMENT AND PRESERVATION↗

Orchestrating Fault Prediction with Live Migration and Checkpointing

Checkpoint/Restart (C/R) is widely used to provide fault tolerance on High-Performance Computing (HPC) systems. However, Parallel File System (PFS) overhead and failure uncertainty cause significant application overhead. This paper develops an adaptive multi-level C/R model that incorporates a failure prediction and analysis model, which orchestrates failure prediction, checkpointing, checkpoint frequency, and proactive live migration along with the additional benefit of Burst Buffers (BB). It effectively reduces the overheads due to failures, checkpointing, and recovery. Simulation results for the Summit supercomputer yield a reduction of ~20%-86% in application overhead due to BBs, orchestrated failure prediction, and migration. We also observe a ~29% decrease in checkpoint writes to BBs, which can increase the longevity of the BB storage devices.

Behera, Subhendu↗

Architecture and Performance of Perlmutter’s 35 PB ClusterStor E1000 All-Flash File System

NERSC's newest system, Perlmutter, features a 35 PB all-flash Lustre file system built on HPE Cray ClusterStor E1000. We present its architecture, early performance figures, and performance considerations unique to this architecture. We demonstrate the performance of E1000 OSSes through low-level Lustre tests that achieve over 90% of the theoretical bandwidth of the SSDs at the OST and LNet levels. We also show end-to-end performance for both traditional dimensions of I/O performance (peak bulk-synchronous bandwidth) and non-optimal workloads endemic to production computing (small, incoherent I/Os at random offsets) and compare them to NERSC's previous system, Cori, to illustrate that Perlmutter achieves the performance of a burst buffer and the resilience of a scratch file system. Finally, we discuss performance considerations unique to all-flash Lustre and present ways in which users and HPC facilities can adjust their I/O patterns and operations to make optimal use of such architectures.

Lockwood, Glenn K↗

Accelerating Application Bulk Synchronous Writes in HPC Environments

High-bandwidth storage tiers are becoming more common for their capability to absorb high-rate, bursty I/Os. Notably, the designs of these fast storage tiers differ from system to system. The variation of these layers and non-uniform methods of access can pose chal- lenges for applications seeking to run at multiple HPC facilities. Therefore, in this work, we present Spectral, a rapid-output ab- straction library to accelerate application, bulk-synchronous writes on HPC systems. We design Spectral to enable applications to use high-bandwidth storage, such as node-local storage and dis- tributed, write-caches (e.g., burst buffers) transparently without requiring modifications to the application or file system source code. The key idea is to allow applications to spend most of the time performing productive work and to not require any source code changes for maximum portability on different HPC archi- tectures. Spectral internally re-routes write-only files through available, high-performance I/O resources before ultimately mi- grating them to the shared global parallel file system. For instance, on Summit, Spectral transparently places application outputs on node-local storage and then utilizes asynchronous migration to the center-wide GPFS file system. We evaluate Spectral on the Summit HPC system (1024 nodes) using the IOR benchmark and real scientific applications. Spectral shows linear performance scaling, improving application write performance by over an order of magnitude when compared to GPFS.

Khan, Awais↗

Ad Hoc File Systems for High-Performance Computing

Storage backends of parallel compute clusters are still based mostly on magnetic disks, while newer and faster storage technologies such as flash-based SSDs or non-volatile random access memory (NVRAM) are deployed within compute nodes. Including these new storage technologies into scientific workflows is unfortunately today a mostly manual task, and most scientists therefore do not take advantage of the faster storage media. One approach to systematically include nodelocal SSDs or NVRAMs into scientific workflows is to deploy ad hoc file systems over a set of compute nodes, which serve as temporary storage systems for single applications or longer-running campaigns. This paper presents results from the Dagstuhl Seminar 17202 “Challenges and Opportunities of User-Level File Systems for HPC” and discusses application scenarios as well as design strategies for ad hoc file systems using node-local storage media. The discussion includes open research questions, such as how to couple ad hoc file systems with the batch scheduling environment and how to schedule stage-in and stage-out processes of data between the storage backend and the ad hoc file systems. Also presented are strategies to build ad hoc file systems by using reusable components for networking and how to improve storage device compatibility. Various interfaces and semantics are presented, for example those used by the three ad hoc file systems BeeOND, GekkoFS, and BurstFS. Their presentation covers a range from file systems running in production to cutting-edge research focusing on reaching the performance limits of the underlying devices.

97 MATHEMATICS AND COMPUTING↗

Programming Abstractions for Managing Workflows on Tiered Storage Systems

Scientific workflows in High Performance Computing (HPC) environments are processing large amounts of data. The storage hierarchy on HPC systems is getting deeper, driven by new technologies (NVRAMs, SSDs, etc.) There is a need for new programming abstractions that allow users to seamlessly manage data at the workflow level on multi-tiered storage systems, and provide optimal workflow performance and use of storage resources. In previous work, we introduced a software architecture Managing Data on Tiered Storage for Scientific Workflows (MaDaTS) that used a Virtual Data Space (VDS) abstraction to hide the complexities of the underlying storage system while allowing users to control data management strategies. In this article, we detail the data-centric programming abstractions that allow users to manage a workflow around its data on the storage layer. The programming abstractions simplify data management for scientific workflows on multi-tiered storage systems, without affecting workflow performance or storage capacity. We measure the overheads and effectiveness introduced by the programming abstractions of MaDaTS. Our results show that these abstractions can optimally use the storage capacity in lesser capacity storage tiers, and simplify data management without adding any performance overheads.

96 KNOWLEDGE MANAGEMENT AND PRESERVATION↗

Processor Reformats Data For Transmission In Bursts

Data-processor-and-buffer electronic system receives audio signals digitized in first standard format at relatively low data rate, rearranges data for transmission in bursts in second standard format at relatively high rate, stores second-format bursts, and releases them at higher rate upon request. Conceived for asynchronous, one-way transmission of digitized speech in outer-space communications, concept of system applied in other digital communication systems in which data transmitted from low-rate sources to high-rate sinks not synchronized with sources.

Steele, Glen F.↗

Detection of individual atoms in helium buffer gas and observation of their real-time motion

Single atoms are detected and their motion measured for the first time to our knowledge by the fluorescence photon-burst method in the presence of large quantities of buffer gas. A single-clipped digital correlator records the photon burst in real time and displays the atom's transit time across the laser beam. A comparison is made of the special requirements for single-atom detection in vacuum and in a buffer gas. Finally, the probability distribution of the bursts from many atoms is measured. It further proves that the bursts observed on resonance are due to single atoms and not simply to noise fluctuations.

Pan, C. L.↗

Buffered freepointer management memory system

A system and method of buffered freepointer management to handle burst traffic to fixed size structures in an external memory system. A circular queue stores implicitly linked free memory locations, along with an explicitly linked list in memory. The queue is updated at the head with newly released locations, and new locations from memory are added at the tail. When a freed location in the queue is reused, external memory need not be updated. When the queue is full, the system attempts to release some of the freepointers such as by dropping them if they are already linked, updating the linked list in memory only if those dropped are not already linked. Latency can be further reduced by loading new locations from memory when the queue is nearly empty, rather than waiting for empty condition, and by writing unlinked locations to memory when the queue is nearly full.

Jacob, Philip↗

A Synthesis Methodology for Intelligent Memory Interfaces in Accelerator Systems

Domain-specific systems improve the performance of a specific set of applications compared to general-purpose processing systems by deploying custom hardware accelerators. These hardware accelerators are generated using high-level synthesis (HLS) tools. The HLS tools enable a comprehensive design space exploration to optimize the compute performance of the generated accelerators. However, they often ignore the challenges of implementing the accelerators in a system-on-chip, particularly how the accelerators access memory. Our work introduces a buffering system design that improves accelerators' memory accesses by intelligently employing burst transactions to prefetch useful data from external memory to on-chip local buffers. Our design is dynamic, parametric, and transparent to the accelerators generated by HLS tools. We derive the buffering system parameters using appropriate compiler-based analysis passes and memory channel latency constraints. The proposed buffering system design results in, on average, 8.8x performance improvements while lowering memory channel utilization on average by 53.2% for a set of PolyBench kernels.

Limaye, Ankur M. (ORCID:0000000194062584)↗

DUNE DAQ R&D integration in ProtoDUNE Single-Phase at CERN

The DAQ system of ProtoDUNE-SP successfully proved its design principles and met the requirements of the beam run of 2018. The technical design of the DAQ system for the DUNE experiment has major differences compared to the prototype due to different requirements placed on the detector, as well as a radically different location of operation. The single-phase prototype at CERN is a major integration facility for R&D aspects of the DUNE DAQ system. The facility allows for the exploration of additional data processing capabilities and optimization of the FELIX system, which is the chosen TPC readout solution for the DUNE single-phase detectors. One of the fundamental differences from the prototype is that the DUNE DAQ relies on self-triggering. Therefore, real-time processing of the data stream for hit and trigger primitive finding is essential for the requirement of continuous readout. The supernova burst trigger requires a large and fast buffering technique, where 3D XPoint persistent memory solutions are evaluated and integrated. In order to maximize resource utilization of the FELIX hosting servers, the elimination of the 100 Gb network communication stack is desired. This implies the design and development of a single-host application layer, which is a fundamental element of the self-triggering chain. This paper discusses the evaluation and integration of these developments for the DUNE DAQ, in the ProtoDUNE environment.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Balloon Data Readout via Free-Space Optical Communications

Free-space optical communications is a mature technology, using eye-safe beams and small-terminals, that has the potential to increase balloon data readout rates by orders of magnitude. Large commercial memory buffers paired with incredibly high-rate bursted communications – 100 Gbps has already been demonstrated with higher rates possible with existing commercial technologies – could revolutionize how science missions are designed and run. We will give an overview of the technology, show some examples of how it could already be useful to missions, and propose next steps to make affordable and robust systems be available to the balloon community.

Don Boroson↗

Distributed data transfers in CMS

The multi-tiered computing infrastructure of the CMS experiment at the LHC depends on the reliable and fast transfer of data between the different CMS computing sites. Data have to be transferred from the Tier-0 to the Tier-l sites for archival in a timely manner to avoid overflowing disk buffers at CERN. Data have to be transferred in bursts to all Tier-2 level sites for analysis as well as synchronized between the different Tier-l sites. The data transfer system is the key ingredient which enables the optimal usage of all distributed resources. The operation of the transfer system consists of monitoring and debugging of transfer issues to guarantee a timely delivery of data to all corners of the CMS computing infrastructure. Further task of transfer operation is to guarantee the consistency of the data at all sites, both on disk and on tape. Procedures to verify the consistency and to debug and repair problems will be discussed.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Characterization of throughput on the AXI DMA bus for burst data transfer over Ethernet

cThe Xilinx AXI Direct Memory Access (AXI DMA) module is an efficient solution for medium-speed data transfer in Xilinx SoC FPGAs, supporting data rates greater than 1000 Gbps even in very suboptimal operating modes. It facilitates direct transfer of AXI stream data into processor memory without constant software intervention, which reduces overhead and ensures consistent data logging. By utilizing the FPGA's available memory, large circular buffers (1-5 GiB) are used to buffer data and accommodate network limitations, enabling high-rate data bursts. In this study, we measured the performance of AXI DMA under conditions simulating its lowest practical data transfer speeds. The Arbitrary Length Data Sender was used to transmit AXI stream packets at 32-bit width and 100 MHz frequency, a narrow width and slow speed. Results show that the AXI DMA can transfer up to 3192.76 Mbps with large packet sizes but experiences reduced performance for smaller packets, as low as 2.6 Mbps for 4-byte packets. For Ethernet-limited applications, packet sizes between 8,000 and 16,000 bytes provided optimal transfer speeds of 874 to 1600 Mbps. These findings suggest that the AXI DMA is not the limiting factor in systems where packet sizes exceed 8,000 bytes.

43 PARTICLE ACCELERATORS↗

An effective write policy for software coherence schemes

The authors study the write behavior and evaluate the performance of various write strategies and buffering techniques for a MIN-based multiprocessor system using the simple software coherence scheme. Hit ratios, memory latencies, total execution time, and total write traffic are used as the performance indices. The write-through write-allocate no-fetch cache using a write-back write buffer is shown to have a better performance than both write-through and write-back caches. This type of write buffer is effective in reducing the volume as well as bursts of write traffic. On average, the use of a write-back cache reduces by 60 percent the total write traffic generated by a write-through cache.

Chen, Yung-Chin↗

In vitro release of organophosphorus acid anhydrolase from functionalized mesoporous silica against nerve agents.

We report here that under different physiological conditions, biomolecular drugs can be stockpiled in a nanoporous support and afterward can be instantly released when needed for acute responses, and the biomolecular drug molecules can also be gradually released from the nanoporous support over a long time for a complete recovery. Organophosphorus acid anhydrolase (OPAA) was spontaneously and largely entrapped in functionalized mesoporous silica (FMS) due to the dominant electrostatic interaction. The OPAA-FMS composite exhibited a burst release in a pH 9.0 NaHCO(3)-Na(2)CO(3) buffer system and a gradual release in pH 7.4 simulated body fluid. The binding of OPAA to NH(2)-FMS can result in less tyrosinyl and tryptophanyl exposure OPAA molecules to aqueous environment. The bound OPAA in FMS displayed lower activity than the free OPAA in solution prior to the enzyme entrapment. However, the released enzyme maintained the native conformational structure and the same high enzymatic activity as that prior to the enzyme entrapment. The in vitro results in the rabbit serum demonstrate that both OPAA-FMS and the released OPAA may be used as a medical countermeasure against the organophosphorus nerve agents.

59 BASIC BIOLOGICAL SCIENCES↗