Filters








46 Hits in 1.5 sec

NVMalloc: Exposing an Aggregate SSD Store as a Memory Partition in Extreme-Scale Machines

Chao Wang, Sudharshan S. Vazhkudai, Xiaosong Ma, Fei Meng, Youngjae Kim, Christian Engelmann
2012 2012 IEEE 26th International Parallel and Distributed Processing Symposium  
In this paper, we propose a novel approach for exploiting NVM as a secondary memory partition so that applications can explicitly allocate and manipulate memory regions therein.  ...  More specifically, we propose an NVMalloc library with a suite of services that enables applications to access a distributed NVM storage system.  ...  ACKNOWLEDGMENTS We would like to thank the reviewers for their feedback. This work was sponsored in part by ORNL, managed by UT-Battelle, LLC for the U.S. DOE (Contract No.  ... 
doi:10.1109/ipdps.2012.90 dblp:conf/ipps/WangVMMKE12 fatcat:i343iklgmnh5hd5lmwlq5sco2e

Efficient Compactions Between Storage Tiers with PrismDB [article]

Ashwini Raina, Jianan Lu, Asaf Cidon, Michael J. Freedman
2022 arXiv   pre-print
Correspondingly, data systems that employ these technologies are typically optimized either to be fast (but expensive) or cheap (but slow).  ...  Inspired by the classic cost-benefit analysis of log cleaning, we develop a new algorithm for multi-tiered storage compaction that balances the benefit of reclaiming space for hot objects in fast storage  ...  Strata [28] is a file system for multi-tiered storage that places data at a file granularity.  ... 
arXiv:2008.02352v5 fatcat:jdodwr3t7fa4namkiqu7aldga4

Exploring the Future of Out-of-Core Computing with Compute-Local Non-Volatile Memory

Myoungsoo Jung, Ellis H. Wilson, Wonil Choi, John Shalf, Hasan Metin Aktulga, Chao Yang, Erik Saule, Umit V. Catalyurek, Mahmut Kandemir
2014 Scientific Programming  
We present novel hardware and software solutions, including the new Unified File System (UFS), to enable fuller utilization of the new compute-local NVM storage.  ...  Therefore, in this work we investigate co-location of NVM and compute by varying I/O interfaces, file systems, types of NVM, and both current and future SSD architectures, uncovering numerous bottlenecks  ...  File Systems Host File Systems.  ... 
doi:10.1155/2014/303810 fatcat:xw477pufinegrjl7cwah2a3eyi

Exploring the future of out-of-core computing with compute-local non-volatile memory

Myoungsoo Jung, Ellis H. Wilson, Wonil Choi, John Shalf, Hasan Metin Aktulga, Chao Yang, Erik Saule, Umit V. Catalyurek, Mahmut Kandemir
2013 Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis on - SC '13  
We present novel hardware and software solutions, including the new Unified File System (UFS), to enable fuller utilization of the new compute-local NVM storage.  ...  Therefore, in this work we investigate co-location of NVM and compute by varying I/O interfaces, file systems, types of NVM, and both current and future SSD architectures, uncovering numerous bottlenecks  ...  File Systems Host File Systems.  ... 
doi:10.1145/2503210.2503261 dblp:conf/sc/JungWCSAYSCK13 fatcat:v6gzb7j4pzetzh7ql3bqii2xce

Nature of System Calls in CPU-centric Computing Paradigm [article]

Viacheslav Dubeyko, Om Rameshwar Gatla, Mai Zheng
2019 arXiv   pre-print
The next generation of non-volatile memory (NVM) technologies raise concerns about the efficiency of the traditional POSIX-based systems.  ...  We build histograms for the most frequent and time-consuming system calls with the goal to understand the nature of distribution on different platforms.  ...  The Ext2, Ext4 and XFS file system are selected because of their support for Direct Access (DAX). The tmpfs file system is used for the RAMdisk based platform.  ... 
arXiv:1903.04075v1 fatcat:3imx2easxzgj7b7mjzviqryll4

A Case for Asymmetric Non-Volatile Memory Architecture [article]

Teng Ma , Yongwei Wu University of Southern California
2019 arXiv   pre-print
This paper rethinks NVM deployment and makes a case for the asymmetric NVM architecture, which decouples servers from persistent data storage.  ...  We build AsymNVM framework based on AsymNVM architecture that implements: 1) high performance persistent data structure update; 2) NVM data management; 3) concurrency control; and 4) crash-consistency  ...  Symmetric Distributed NVM Symmetric architecture is widely used in distributed systems (e.g., shared memory and distributed file systems).  ... 
arXiv:1809.09395v2 fatcat:feho443vbvexno7endxmog24xa

From ARIES to MARS

Joel Coburn, Trevor Bunker, Meir Schwarz, Rajesh Gupta, Steven Swanson
2013 Proceedings of the Twenty-Fourth ACM Symposium on Operating Systems Principles - SOSP '13  
This paper presents a novel storage primitive, called editable atomic writes (EAW), that enables sophisticated, highly-optimized WAL schemes in fast NVM-based storage systems.  ...  Transaction-based systems often rely on write-ahead logging (WAL) algorithms designed to maximize performance on disk-based storage.  ...  Acknowledgements We would like to thank the anonymous reviewers, our shepherd, Michael Swift, and Geoff Voelker for their valuable feedback.  ... 
doi:10.1145/2517349.2522724 dblp:conf/sosp/CoburnBSGS13 fatcat:ub6xqs5zmvee3bxvjp24p7q33y

LB+Trees

Jihang Liu, Shimin Chen, Lujun Wang
2020 Proceedings of the VLDB Endowment  
3DXPoint memory is the first commercially available NVM solution targeting mainstream computer systems.  ...  Our micro-benchmark experiments on a real machine equipped with 3DXPoint memory shows that LB + -Tree achieves up to 1.12-2.92x speedups over state-of-the-art NVM optimized B + -Trees for insertions while  ...  File systems are installed on the 3DXPoint modules using the fsdax mode.  ... 
doi:10.14778/3384345.3384355 fatcat:pq5gbuoukjfdvpwzgy3dbhzuhu

Userland CO-PAGER

Feng Li, Daniel G. Waddington, Fengguang Song
2019 Proceedings of the 3rd International Conference on High Performance Compilation, Computing and Communications - HP3C '19  
With the emergence of low-latency non-volatile memory (NVM) storage, the software overhead, incurred by the operating system, becomes more prominent.  ...  The userspace component handles (redirected) page faults, performs memory management and I/O operations and accesses NVM storage directly.  ...  We would like to thank the Storage Systems group in IBM Research Almaden, for their assistance during the development of CO-PAGER.  ... 
doi:10.1145/3318265.3318272 fatcat:wed44x66mvajjds5pm5pmspthu

Espresso: Brewing Java For More Non-Volatility with Non-volatile Memory [article]

Mingyu Wu, Ziming Zhao, Haoyu Li, Heting Li, Haibo Chen, Binyu Zang, Haibing Guan
2017 arXiv   pre-print
Fast, byte-addressable non-volatile memory (NVM) embraces both near-DRAM latency and disk-like persistence, which has generated considerable interests to revolutionize system software stack and programming  ...  This paper proposes Espresso, a holistic extension to Java and its runtime, to enable Java programmers to exploit NVM for persistence management with high performance.  ...  Other systems combine other other emerging hardware fea-tures with NVM. Octopus [27] leverages RDMA technology together with NVM to build a distributed persistent memory file system.  ... 
arXiv:1710.09968v1 fatcat:gsiubjxvkzef5iu75ljkfmc7ba

DAOS: A Scale-Out High Performance Storage Stack for Storage Class Memory [chapter]

Zhen Liang, Johann Lombardi, Mohamad Chaarawi, Michael Hennecke
2020 Lecture Notes in Computer Science  
We also describe initial performance results of DAOS for IO500 benchmarks.  ...  For HPC workloads, DAOS provides direct MPI-IO and HDF5 support as well as POSIX access for legacy applications.  ...  Traditional storage system like Lustre [11] , Spectrum Scale [12] , or CephFS [10] can be used for disk-based storage, and it is possible to move data between DAOS and such external file systems.  ... 
doi:10.1007/978-3-030-48842-0_3 fatcat:5mnbru2fm5a63dot53xtouuo6a

Beyond block I/O: Rethinking traditional storage primitives

Xiangyong Ouyang, David Nellans, Robert Wipfel, David Flynn, Dhabaleswar K. Panda
2011 2011 IEEE 17th International Symposium on High Performance Computer Architecture  
Initial implementations of NVM storage devices have chosen to export these same disk-based seek/read/write interfaces because they provide compatibility for legacy applications.  ...  Over the last twenty years the interfaces for accessing persistent storage within a computer system have remained essentially unchanged.  ...  Log based designs work well for NAND-flash based devices because the slow erase time of physical blocks is no longer on the critical path for write operations.  ... 
doi:10.1109/hpca.2011.5749738 dblp:conf/hpca/OuyangNWFP11 fatcat:m6csjwonevbuvm45gblggiqtgi

Survey of Storage Systems for High-Performance Computing

2018 Supercomputing Frontiers and Innovations  
In current supercomputers, storage is typically provided by parallel distributed file systems for hot data and tape archives for cold data.  ...  There is a wide variety of file systems to choose from, each tuned for different use cases and implementing different optimizations.  ...  This material reflects only the authors' view and the EU commission is not responsible for any use that may be made of the information it contains.  ... 
doi:10.14529/jsfi180103 fatcat:hi3qctpl7rfvjgl53pxmqwqviy

A Survey of Non-Volatile Main Memory Technologies: State-of-the-Arts, Practices, and Future Directions [article]

Haikun Liu, Di Chen, Hai Jin, Xiaofei Liao, Bingsheng He, Kan Hu, Yu Zhang
2020 arXiv   pre-print
Non-Volatile Main Memories (NVMMs) have recently emerged as promising technologies for future memory systems.  ...  They bring many research opportunities as well as challenges on system architectural designs, memory management in operating systems (OSes), and programming models for hybrid memory systems.  ...  WADE: Writeback-aware dynamic cache management for NVM-based main memory system.  ... 
arXiv:2010.04406v1 fatcat:jna5pb7lizhvllmhfle4yikife

Data structures access model for remote shared memory

Anatoliy Nyrkov, Konstantin Ianiushkin, Andrey Nyrkov, Yulia Romanova, Vagiz Gaskarov, V. Breskich, S. Uvarova
2021 E3S Web of Conferences  
The combination of in-memory storage, remote direct memory access and remote calls requires rethinking how data organized, protected and queried in distributed systems.  ...  Recent achievements in high-performance computing significantly narrow the performance gap between single and multi-node computing, and open up opportunities for systems with remote shared memory.  ...  For this benchmark, each node launches a process to independently generate mostly reads workload. For each system, we run the benchmark few times.  ... 
doi:10.1051/e3sconf/202124407001 fatcat:q5id5kyl2jbddojeyv2w64iztm
« Previous Showing results 1 — 15 out of 46 results