Filters








6,824 Hits in 7.2 sec

Micro-benchmark level performance comparison of high-speed cluster interconnects

Jiuxing Liu, B. Chandrasekaran, Weikuan Yu, Jiesheng Wu, D. Buntinas, S. Kini, P. Wyckoff, D.K. Panda
11th Symposium on High Performance Interconnects, 2003. Proceedings.  
In this paper, we present a comprehensive performance evaluation of three high speed cluster interconnects: Infini-Band, Myrinet and Quadrics.  ...  In order to explore the full communication capability of the interconnects, we have implemented the micro-benchmark suite at the low level messaging layer provided by each interconnect.  ...  Conclusions In this paper, we have used a set of micro-benchmarks to evaluate three high performance cluster interconnects: In-finiBand, Myrinet and Quadrics.  ... 
doi:10.1109/conect.2003.1231479 dblp:conf/hoti/LiuCYWBKWP03 fatcat:7rhphwl6hvagjjess2axo7fyke

Comparing Ethernet and Soft RoCE for MPI Communication

Gurkirat Kaur, Manoj Kumar, Manju Bala
2014 IOSR Journal of Computer Engineering  
A comparison between the IMB and OSU Micro benchmark is done and our results shows that IMB Benchmark is performing better in case of collective Benchmark class as OSU Micro benchmark is performing better  ...  In this paper, we evaluate the Linux cluster, having multi nodes with fast interconnects i.e. Gigabit Ethernet & Soft RoCE and evaluates its performance using IMB and OSU Micro Benchmarks.  ...  We used Intel's MPI Benchmark and Ohio State University (OSU) Micro Benchmark to measure and compare the performance of the two interconnect i.e. Ethernet and Soft RoCE of the Linux cluster.  ... 
doi:10.9790/0661-16415258 fatcat:b3eql5x5z5fcthiwisnquebxyq

A Micro-benchmark Suite for Evaluating Hadoop MapReduce on High-Performance Networks [chapter]

Dipti Shankar, Xiaoyi Lu, Md. Wasi-ur-Rahman, Nusrat Islam, Dhabaleswar K. (DK) Panda
2014 Lecture Notes in Computer Science  
Since a majority of the existing clusters today are equipped with modern, high-speed interconnects such as InfiniBand and 10 GigE, that offer high bandwidth and low communication latency, it is essential  ...  We also show how this micro-benchmark suite can be used to evaluate the performance of Hadoop MapReduce over different networks/protocols and parameter configurations on modern clusters.  ...  For MR-SKEW micro-benchmark, Fig. 3(c) shows that the performance of the MapReduce job improves by about 10-12 % with the use of high-speed interconnects.  ... 
doi:10.1007/978-3-319-13021-7_2 fatcat:aubbemw6knffdg5o5nqwjdpsrq

Performance Comparison of MPI Implementations over InfiniBand, Myrinet and Quadrics

Jiuxing Liu, Balasubramanian Chandrasekaran, Jiesheng Wu, Weihang Jiang, Sushmitha Kini, Weikuan Yu, Darius Buntinas, Peter Wyckoff, D K. Panda
2003 Proceedings of the 2003 ACM/IEEE conference on Supercomputing - SC '03  
The first part consists of a set of MPI level micro-benchmarks that characterize different aspects of MPI implementations.  ...  The second part of the performance evaluation consists of application level benchmarks. We have used the NAS Parallel Benchmarks and the sweep3D benchmark.  ...  We would also like to thank Ben Eiref, Robert Starmer, and Lorne Boden from Topspin Communications for all their efforts in providing us access to their 16 node InfiniBand cluster.  ... 
doi:10.1145/1048935.1050208 dblp:conf/sc/LiuCWJKYBWP03 fatcat:jvcinutju5drdlt2b7bwshbyna

High Performance Message-passing InfiniBand Communication Device for Java HPC

Omar Khan, Mohsan Jameel, Aamir Shafi
2014 Procedia Computer Science  
Apart from increasing instruction execution speed of software, a lot of research is also under way to develop high speed proprietary communication interconnects.  ...  In the context of cluster mode, the current release of MPJ Express provides two transport level devices.  ...  NPB are one of the most well-known benchmarks for evaluating performance of high performance parallel systems.  ... 
doi:10.1016/j.procs.2014.05.133 fatcat:zqcht2axs5dazlhhwcutpzn3vq

Benefits of high speed interconnects to cluster file systems: a case study with Lustre

Weikuan Yu, R. Noronha, Shuang Liang, D.K. Panda
2006 Proceedings 20th IEEE International Parallel & Distributed Processing Symposium  
In this paper, we perform an evaluation of a popular cluster file system, Lustre, over two of the leading high speed cluster interconnects: InfiniBand and Quadrics.  ...  Furthermore, our experimental results indicate that Lustre meta-data operations do not scale with an increasing number of OSSs, in spite of using high performance interconnects. *  ...  These benchmarks test the speed of parallel IO capability of high performance computing applications.  ... 
doi:10.1109/ipdps.2006.1639564 dblp:conf/ipps/YuNLP06 fatcat:cttwtymv7fgwfm66gmok4m2or4

A Performance-Based Methodology to Improve Grid Exploitation [chapter]

A. Clematis, A. Corana, D. D'Agostino, A. Galizia, A. Quarati
2009 Lecture Notes in Computer Science  
Leveraging on a two-level benchmarking methodology, GREEN allows users to express performance preference through an appropriate extension to Grid submission and description languages such as JSDL and Glue  ...  We introduce GREEN a management tool primarily devoted to the matchmaking process, based on a performance characterization of both resources and job requirements.  ...  Micro-Benchmarks In order to supply a basic resource characterization, mainly based on low-level performance capacity, we considered the use of traditional micro-benchmarks.  ... 
doi:10.1007/978-3-642-04989-7_4 fatcat:niuayufdifhkhhcqpm2d7hnc7a

A preliminary analysis of the InfiniPath and XD1 network interfaces

R. Brightwell, D. Doerfler, K.D. Underwood
2006 Proceedings 20th IEEE International Parallel & Distributed Processing Symposium  
Our evaluation includes several micro-benchmark results as well as some initial application performance data.  ...  Two recently delivered systems have begun a new trend in cluster interconnects.  ...  The XD1 machine is a resource of the National Center for Computational Sciences at Oak Ridge National Laboratory, which is supported by the Office of Science of the U.S.  ... 
doi:10.1109/ipdps.2006.1639568 dblp:conf/ipps/BrightwellDU06 fatcat:bewvypammnehfdlinqtwnn2gfu

A Software Based Approach for Providing Network Fault Tolerance in Clusters with uDAPL interface: MPI Level Design and Performance Evaluation

Abhinav Vishnu, Prachi Gupta, Amith Mamidala, Dhabaleswar Panda
2006 ACM/IEEE SC 2006 Conference (SC'06)  
At the same time, introduction of high speed RDMA-enabled interconnects like InfiniBand, Myrinet, Quadrics, RDMAenabled Ethernet has escalated the trends in cluster computing.  ...  We implement our design and evaluate it with micro-benchmarks and applications.  ...  At the same time, introduction of high speed RDMA-enabled interconnects like InfiniBand, Myrinet, Quadrics, RDMAenabled Ethernet has escalated the trends in cluster computing.  ... 
doi:10.1109/sc.2006.5 fatcat:okgzbioj5bgsbdy76e2g3o5tw4

Scalable systems software---A software based approach for providing network fault tolerance in clusters with uDAPL interface

Abhinav Vishnu, Prachi Gupta, Amith R. Mamidala, Dhabaleswar K. Panda
2006 Proceedings of the 2006 ACM/IEEE conference on Supercomputing - SC '06  
At the same time, introduction of high speed RDMA-enabled interconnects like InfiniBand, Myrinet, Quadrics, RDMAenabled Ethernet has escalated the trends in cluster computing.  ...  We implement our design and evaluate it with micro-benchmarks and applications.  ...  At the same time, introduction of high speed RDMA-enabled interconnects like InfiniBand, Myrinet, Quadrics, RDMAenabled Ethernet has escalated the trends in cluster computing.  ... 
doi:10.1145/1188455.1188545 dblp:conf/sc/VishnuGMP06 fatcat:gpnejxsgibdivkz5wsrxgbl7ka

Breaking HPC Barriers with the 56GbE Cloud

Muhammad Atif, Rika Kobayashi, Benjamin J. Menadue, Ching Yeh Lin, Matthew Sanderson, Allan Williams
2016 Procedia Computer Science  
A detailed analysis of these metrics in the context of micro-benchmarks and scientific applications is presented, including the affects of using TCP and RDMA on scientific applications.  ...  This cloud is targeted at high-performance scientific applications, and enables scientists to build their own clusters when their demands and software stacks conflict with traditional bare-metal HPC environments  ...  Conclusion and Future Work We used the OSU MPI micro-benchmarks, the NPB MPI macro-benchmarks to characterise performance of three different platforms including a peak supercomputing HPC cluster, a private  ... 
doi:10.1016/j.procs.2016.07.174 fatcat:7nym2pkzrndkpagvtx2ujkscje

High Speed Cycle-Approximate Simulation of Embedded Cache-Incoherent and Coherent Chip-Multiprocessors

Christopher Thompson, Miles Gould, Nigel Topham
2018 International journal of parallel programming  
We present a new high-speed, near cycle-accurate simulator, addressing an important but neglected category of multicore systems: deeply-embedded cacheincoherent MPSoCs.  ...  The tool flow takes a high level MPSoC description and generates Verilog for FPGA and silicon implementation, and equivalent configuration files for the simulator.  ...  Acknowledgements This work has made use of the resources provided by the Edinburgh Compute and Data Facility (ECDF) [25].  ... 
doi:10.1007/s10766-018-0566-x fatcat:pmvu23xlafaejhva5o2vzldr4e

S4BXI: the MPI-ready Portals 4 Simulator

Julien Emmanuel, Matthieu Moy, Ludovic Henrio, Gregoire Pichon
2021 2021 29th International Symposium on Modeling, Analysis, and Simulation of Computer and Telecommunication Systems (MASCOTS)  
We present a simulator for High Performance Computing (HPC) interconnection networks.  ...  It models Portals 4, a standard low-level API for communication, and it allows running unmodified applications that use higher-level network APIs such as the Message Passing Interface (MPI).  ...  the development of our simulator, especially Tom Cornebize for his help on the tuning of SMPI, and Arnaud Legrand for his advices.  ... 
doi:10.1109/mascots53633.2021.9614285 fatcat:tenquktuznh2xhktfnktecrqxm

Scalability and communication performance of HPC on Azure Cloud

Hanan A. Hassan, Shimaa A. Mohamed, Walaa M. Sheta
2016 Egyptian Informatics Journal  
Our objective was to evaluate the scalability and performance of High Performance Cloud Computing on Microsoft Azure Cloud infrastructure by using well known Benchmarks, namely, IMB point-to-point communication  ...  In our experiments, performance of the HPC applications on the cloud is assessed in terms of MOPS and speedup, and is tested under different configurations of cluster sizes.  ...  For example, authors in [4, 6, 15, 16] proposed high speed cluster interconnects InfiniBand to improve communication performance between cluster nodes, while authors in [15, [18] [19] [20] [21] analyzed  ... 
doi:10.1016/j.eij.2015.11.001 fatcat:sa5n3j2tuncmtgwa2baxdfzule

Design and Evaluation of Network Topology-/Speed- Aware Broadcast Algorithms for InfiniBand Clusters

H. Subramoni, K. Kandalla, J. Vienne, S. Sur, B. Barth, K. Tomko, R. Mclay, K. Schulz, D.K. Panda
2011 2011 IEEE International Conference on Cluster Computing  
scheme over the default scheme at the micro-benchmark level.  ...  The proposed network speed-aware algorithms are able to attain micro-benchmark performance on the heterogeneous SDR-DDR InfiniBand cluster to perform on par with runs on the DDR only portion of the cluster  ...  ACKNOWLEDGMENTS We would like to express our sincere appreciation to Doug Johnson of Ohio Supercomputer Center for his timely help in enabling us to evaluate our algorithms.  ... 
doi:10.1109/cluster.2011.43 dblp:conf/cluster/SubramoniKVSBTMSP11 fatcat:4lksqieff5hnvkslsflsbe5bue
« Previous Showing results 1 — 15 out of 6,824 results