357 Hits in 8.6 sec

Performance Evaluation of Distributed Computing over Heterogeneous Networks [chapter]

Ouissem Ben Fredj, Éric Renault
2007 Lecture Notes in Computer Science  
RWAPI supports Ethernet, InfiniBand and Myrinet network interconnects. This paper introduces RWAPI and the design of its network component on top of both InfiniBand and Myrinet interconnects.  ...  It has been developed to provide performance to higher applications on a wide variety of architectures. We implemented RWAPI on top of the modular software architecture called GRWA.  ...  RWAPI over Myrinet Interconnect: There are many ways to implement the Remote-Write protocol on top of Myrinet.  ... 
doi:10.1007/978-3-540-75444-2_11 fatcat:jiijoody2bhezoz34r6x5uovgq

A feasibility analysis of power-awareness and energy minimization in modern interconnects for high-performance computing

Reza Zamani, Ahmad Afsahi, Ying Qian, Carl Hamacher
2007 2007 IEEE International Conference on Cluster Computing  
Operations finish without the remote side being involved. The RDMA operations include RDMA Write, gm_put(), and RDMA Read, gm_get().  ...  In the following, we provide an overview of each interconnect. 1) Myrinet-2000: For the Myrinet tests, we experimented with a 16-port Myricom Myrinet-2000 [29] network, consisting of a 16-port switch  ... 
doi:10.1109/clustr.2007.4629224 dblp:conf/cluster/ZamaniAQH07 fatcat:bqfxcvt3tjdxvkx3w34c6eenum

Protocols and strategies for optimizing performance of remote memory operations on clusters

J. Nieplocha, V. Tipparaju, A. Saify, D. Panda
2002 Proceedings 16th International Parallel and Distributed Processing Symposium  
The paper describes software architecture for supporting remote memory operations on clusters equipped with high-performance networks such as Myrinet and Giganet/Emulex cLAN.  ...  In particular, the issues of memory registration, management of network resources and memory consumption on the host, are discussed and solved to achieve an efficient implementation.  ...  We are interested in exploiting the high-performance networks and protocols for optimizing remote memory copy on commodity clusters. The predominant high-performance network for clusters is Myrinet.  ... 
doi:10.1109/ipdps.2002.1016563 dblp:conf/ipps/NieplochaTSP02 fatcat:kcv3abvcejgdrmxkexc5jzidhu

Efficient replicated method invocation in Java

Jason Maassen, Thilo Kielmann, Henri E. Bal
2000 Proceedings of the ACM 2000 conference on Java Grande - JAVA '00  
Performance • Benchmark (replicated stack, in microseconds) -Pentium Pro 200, Myrinet network + efficient multicast cpus push top sequential 3.1 0.1 RMI, local 6.1 2.7 RMI, remote 55.3  ...  June 2000  ... 
doi:10.1145/337449.337486 dblp:conf/java/MaassenKB00 fatcat:qdu7rxdfsbeknirsh5kyhude74

10-Gigabit iWARP Ethernet: Comparative Performance Analysis with InfiniBand and Myrinet-10G

Mohammad J. Rashti, Ahmad Afsahi
2007 2007 IEEE International Parallel and Distributed Processing Symposium  
iWARP is a set of standards enabling Remote Direct Memory Access (RDMA) over Ethernet. iWARP supporting RDMA and OS bypass, coupled with TCP/IP Offload Engines, can fully eliminate the host CPU involvement  ...  In this paper we assess the potential of such an interconnect for high-performance computing by comparing its performance with two leading cluster interconnects, InfiniBand and Myrinet-10G.  ...  We plan to put these networks to the test in a larger testbed to have a better evaluation of the extent to which the multiple-connection performance of the NetEffect device will affect real world applications  ... 
doi:10.1109/ipdps.2007.370480 dblp:conf/ipps/RashtiA07 fatcat:fznac5yc3bchvn5on7tppbzofa

ATOLL, a new switched, high speed Interconnect in Comparison to Myrinet and SCI [chapter]

Markus Fischer, Ulrich Brüning, Jörg Kluge, Lars Rzymianowicz, Patrick Schulz, Mathias Waack
2000 Lecture Notes in Computer Science  
While standard processors achieve supercomputer performance, a performance gap exists between the interconnect of MPP's and COTS.  ...  Hence, high speed interconnects have an important impact on a cluster's performance.  ...  to achieve highest performance 74MB s remote write vs. 7.5 MB s remote read using a Pentium II450.  ... 
doi:10.1007/3-540-45591-4_3 fatcat:mvftv4ud3rh6rj7y5krw57o7n4

Efficient Block Device Sharing over Myrinet with Memory Bypass

Evangelos Koukis, Nectarios Koziris
2007 2007 IEEE International Parallel and Distributed Processing Symposium  
In this paper we present gmbock, a client/server system for network sharing of storage devices over Myrinet, which uses an optimized data path in order to transfer data directly from the storage medium  ...  Its design enhances existing programming abstractions, combining the user level networking characteristics of Myrinet with Linux's virtual memory infrastructure, in order to construct the data path in  ...  The Myrinet software stack is displayed in Fig. 2(a) . Myrinet NICs reside on the peripheral bus of a cluster node (PCI/PCI-X in the case of Myrinet-2000 used on our testbed).  ... 
doi:10.1109/ipdps.2007.370219 dblp:conf/ipps/KoukisK07 fatcat:i2puge6ekzdyjjv6qcgiarsxyu

Performance Evaluation of Fast Ethernet, Giganet and Myrinet on a Cluster [chapter]

Marcelo Lobosco, Vítor Santos Costa, Claudio L. de Amorim
2002 Lecture Notes in Computer Science  
The impacts caused by the use of Remote DMA Write are also evaluated.  ...  Two RDMA operations, RDMA Write and RDMA Read, respectively write and read remote data. Myrinet is the most popular high-speed interconnect used to build clusters.  ...  These results show the influence of the MPI implementation over performance.  ... 
doi:10.1007/3-540-46043-8_29 fatcat:voho6crkqjeu7drkfkvn4jhhva

Architectural and Performance Evaluation of GigaNet and Myrinet Interconnects on Clusters of Small-Scale SMP Servers

Jenwei Hsieh, Tau Leng, V. Mashayekhi, R. Rooholamini
2000 ACM/IEEE SC 2000 Conference (SC'00)  
The performance measurements reported here suggest that the implementation of Message Passing Interface (MPI) significantly affects the cluster performance.  ...  Although MPICH-GM over Myrinet demonstrates lower latency with small messages, the polling-driven implementation of MPICH-GM often leads to tight synchronization between communication processes and higher  ...  Most of the NPB programs show slightly lower performance with MPICH-GM over Myrinet than MPI/ Pro over GigaNet on the 6350 cluster.  ... 
doi:10.1109/sc.2000.10027 dblp:conf/sc/HsiehLMR00 fatcat:y62eu4l7srcpthqmq2fesarkgi

Microbenchmark performance comparison of high-speed cluster interconnects

Jiuxing Liu, B. Chandrasekaran, Weikuan Yu, Jiesheng Wu, D. Buntinas, S. Kini, D.K. Panda, P. Wyckoff
2004 IEEE Micro  
Our Myrinet network consists of M3F-PCIXD-2 network interface cards connected by a Myrinet 2000 switch. The link bandwidth of the Myrinet network is 2 Gbps for each direction (2 Gbps plus 2 Gbps).  ...  All interconnects we study here support access to the memory space of remote nodes via RDMA. We therefore measured the latency to finish a remote write operation.  ... 
doi:10.1109/mm.2004.1268994 fatcat:2tmrx5boybgtxkvbhlmvweey3y

Head-to-TOE Evaluation of High-Performance Sockets over Protocol Offload Engines

P. Balaji, W. Feng, Q. Gao, R. Noronha, W. Yu, D. K. Panda
2005 Proceedings IEEE International Conference on Cluster Computing  
Although there exists significant research on individual network technologies such as 10GigE, InfiniBand (IBA), and Myrinet; to the best of our knowledge, there has been no work that compares the capabilities  ...  Therefore, we present performance evaluations across 10GigE, IBA, and Myrinet (with identical cluster-compute nodes) in order to enable a coherent comparison with respect to the sockets interface.  ...  Myrinet: The Myrinet network was based on Myrinet-2000 'E' (dual-port) adapters connected by a Myrinet-2000 wormhole router crossbar switch.  ... 
doi:10.1109/clustr.2005.347068 dblp:conf/cluster/BalajiFGNYP05 fatcat:jxxjo3t42rds3nvtktzrow6fei

Bridging the Ethernet-Ethernot Performance Gap

P. Balaji, Wu-chun Feng, D.K. Panda
2006 IEEE Micro  
Finally, the Myrinet network in Cluster 3 is based on Myrinet-2000 E (dual-port) adapters connected by a Myrinet-2000 wormhole-routed crossbar switch.  ...  The Myrinet network in our evaluation consists of a Myrinet-2000 switch-a 16-port crossbarthat connects Myrinet-2000 E cards, each of which has two ports, and each port has a 2-Gbps bandwidth.  ... 
doi:10.1109/mm.2006.48 fatcat:fepv7mvg4vfepdfgcu5tahjskm

PadicoTM: an open integration framework for communication middleware and runtimes

Alexandre Denis, Christian Pérez, Thierry Priol
2003 Future generations computer systems  
The second obstacle is the design of low-level communication layers for System Area Networks (SAN) in parallel systems or PC clusters (Myrinet, SCI, ...) in a grid system.  ...  within high-performance applications.  ...  Acknowledgments This work was supported by the Incentive Concerted Action "GRID" (ACI GRID) of the French Ministry of Research.  ... 
doi:10.1016/s0167-739x(03)00034-7 fatcat:bjgpb7mcgzevbk2n2yl7f75qhe

High Performance Remote Memory Access Communication: The Armci Approach

J. Nieplocha, V. Tipparaju, M. Krishnan, D. K. Panda
2006 The international journal of high performance computing applications  
This paper describes the Aggregate Remote Memory Copy Interface (ARMCI), a portable high performance remote memory access communication interface, developed originally under the U.S.  ...  The paper discusses the model, addresses challenges of portable implementations, and demonstrates that ARMCI delivers high performance on a variety of platforms.  ...  The RMA model is closely aligned with RDMA capabilities of modern networks (Infiniband, Myrinet, VIA, Elan), which provide hardware support to read from or write to remote memory locations.  ... 
doi:10.1177/1094342006064504 fatcat:374gflhodrgxrpaznnc3zs5fna

Efficient Java RMI for parallel programming

Jason Maassen, Rob Van Nieuwpoort, Ronald Veldema, Henri Bal, Thilo Kielmann, Ceriel Jacobs, Rutger Hofman
2001 ACM Transactions on Programming Languages and Systems  
We evaluate the performance of Manta using benchmarks and applications that run on a 32-node Myrinet cluster.  ...  This high performance is accomplished by pushing almost all of the runtime overhead of RMI to compile time. We study the performance differences between the Manta and the Sun RMI protocols in detail.  ...  We thank the reviewers for their valuable comments on an earlier version of the paper. We thank Ronald Blankendaal, Monique Dewanchand, and Martijn Thieme for developing the Java applications.  ... 
doi:10.1145/506315.506317 fatcat:jmixxxvjzjcyddd7lyw6dhpfbm
« Previous Showing results 1 — 15 out of 357 results