A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2017; you can also visit the original URL.
The file type is application/pdf.
Filters
EMP
2001
Proceedings of the 2001 ACM/IEEE conference on Supercomputing (CDROM) - Supercomputing '01
Preserved Fulltext
In this paper we propose Ethernet Message Passing (EMP), a completely new zero-copy, OS-bypass messaging layer for Gigabit Ethernet on Alteon NICs where the entire protocol processing is done at the NIC ...
To the best of our knowledge, this is the first NIC-level implementation of a zero-copy message passing layer for Gigabit Ethernet. ...
We plan to explore the design space of EMP in terms of NIC v s. host. Currently we have aimed to implement the entire protocol processing at the NIC. ...
doi:10.1145/582034.582091
dblp:conf/sc/ShivamWP01
fatcat:teqapyhl3rcnrh6fzz37q6ctkm
High-performance message-passing over generic Ethernet hardware with Open-MX
2011
Parallel Computing
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2013; you can also visit the original URL.
The file type is application/pdf.
We present Open-MX, an open-source message passing stack over generic Ethernet. ...
OS-bypass operations allowing the application to submit communication requests to the NIC without going across all the expensive operating system layers. ...
For this purpose, OS-bypass communication is used, as along with zero-copy for large messages. The operating system is not involved in the communication. ...
doi:10.1016/j.parco.2010.11.001
fatcat:fcwsyzdkinelnjw4qrevjb3bi4
Design and implementation of Open-MX: High-performance message passing over generic Ethernet hardware
2008
Proceedings, International Parallel and Distributed Processing Symposium (IPDPS)
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2011; you can also visit the original URL.
The file type is application/pdf.
Open-MX is a new message passing layer implemented on top of the generic Ethernet stack of the Linux kernel. ...
The first performance evaluation shows interesting latency and bandwidth results on 1 and 10 gigabit hardware. ...
To do so, OS-bypass communication is used for small messages and zero-copy for large messages. ...
doi:10.1109/ipdps.2008.4536140
dblp:conf/ipps/Goglin08
fatcat:oc4qm75m7vh6roe2ns5forp2na
Comparing Ethernet and Myrinet for MPI communication
2004
Proceedings of the 7th workshop on Workshop on languages, compilers, and run-time support for scalable systems - LCR '04
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2012; you can also visit the original URL.
The file type is application/pdf.
This shows that commodity networking, if used efficiently, can be a viable alternative to specialized networking for high-performance message passing. ...
However, on the NAS benchmarks, GM messaging over Myrinet only achieves 5% higher application performance than TCP messaging over Ethernet. ...
GM allows memory-protected user-level OS-bypass network interface access to applications, thereby achieving zero-copy message sends and receives at the application level. ...
doi:10.1145/1066650.1066659
fatcat:ltio2h4fr5abjicw5nak2opawa
An overview of security issues in cluster interconnects
2006
Sixth IEEE International Symposium on Cluster Computing and the Grid (CCGRID'06)
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2017; you can also visit the original URL.
The file type is application/pdf.
interconnects used in Top500 supercomputers and explores possible schemes to ensure secure cluster intra-communication encompassing the host processor, secure coprocessor and the Network Interface Card (NIC ...
EMP [24] is a use-level communication protocol facilitating zero copy, OS bypass, and NIC-driven communication on Gigabit Ethernet. ...
In [26] , a NIC design was proposed for fast Ethernet network ture of this specification is that it provides zero copy and OS bypass while it uses TCP/IP as its underlying communication protocol which ...
doi:10.1109/ccgrid.2006.1630920
fatcat:iyybq4xx6vhrbnpnyit5w3lgmu
NIC-Assisted Cache-Efficient Receive Stack for Message Passing over Ethernet
[chapter]
2009
Lecture Notes in Computer Science
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2011; you can also visit the original URL.
The file type is application/pdf.
High-speed networking in clusters usually relies on advanced hardware features in the NICs, such as zero-copy capability. ...
However, like QMP [3] and PM [23] (or any other software-based message passing), being compatible with any legacy Ethernet NICs also means that Open-MX suffers from limited hardware features. ...
Such a strategy has been studied for a long time to offer zero-copy socket implementations [4] and more recently for Ethernetbased message passing [19] . ...
doi:10.1007/978-3-642-03869-3_98
fatcat:3mnxv6hn65ht7coqw3ioqtlq3i
NIC-assisted cache-efficient receive stack for message passing over Ethernet
2011
Concurrency and Computation
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2019; you can also visit the original URL.
The file type is application/pdf.
High-speed networking in clusters usually relies on advanced hardware features in the NICs, such as zero-copy capability. ...
However, like QMP [3] and PM [23] (or any other software-based message passing), being compatible with any legacy Ethernet NICs also means that Open-MX suffers from limited hardware features. ...
Such a strategy has been studied for a long time to offer zero-copy socket implementations [4] and more recently for Ethernetbased message passing [19] . ...
doi:10.1002/cpe.1632
fatcat:icewpjgdffg3xbmpipsnuvopsy
On a NIC's Operating System, Schedulers and High-Performance Networking Applications
[chapter]
2006
Lecture Notes in Computer Science
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2007; you can also visit the original URL.
The file type is application/pdf.
In most cases, these CPU's are only used by the vendor and are operated by a proprietary OS, which makes them inaccessible to the HPC application developer. ...
Today's modern high-end Network Interface Cards (NICs) are equipped with an onboard CPU. ...
The Ethernet Message Passing (EMP) [4] system, of the Ohio Supercomputer Center (OSC), is a zero-copy and OS-bypass messaging layer for Gigabit Ethernet. ...
doi:10.1007/11847366_39
fatcat:acr2psaau5h7jpnvaearw7hz3m
iWarp protocol kernel space software implementation
2006
Proceedings 20th IEEE International Parallel & Distributed Processing Symposium
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2006; you can also visit the original URL.
The file type is application/pdf.
Zero-copy, RDMA, and protocol offload are three very important characteristics of high performance interconnects. ...
With the introduction of iWarp, it is now possible to achieve all three over existing low-cost TCP/IP networks. iWarp is a step in the right direction, but currently requires an expensive RNIC to enable zero-copy ...
Though there have been attempts to enable zero-copy without OS bypass, such as [11] , these attempts have made certain assumptions that will not always hold true. ...
doi:10.1109/ipdps.2006.1639565
dblp:conf/ipps/DalessandroDW06
fatcat:dfxrokgpdzgyzn3archot2fprq
Accelerating Distributed Computing Applications Using a Network Offloading Framework
2007
2007 IEEE International Parallel and Distributed Processing Symposium
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2007; you can also visit the original URL.
The file type is application/pdf.
We demonstrate this claim by evaluating several offloaded applications: a distributed total message ordering algorithm and a packet generator. ...
This enables applications to directly access the NIC, thus bypassing the OS. The Ethernet Message Passing (EMP) [20] system is a zero-copy and OS-bypass messaging layer for Gigabit Ethernet. ...
Arsenic [17] is a Gigabit Ethernet NIC program that exports an extended interface to the host operating system. ...
doi:10.1109/ipdps.2007.370319
dblp:conf/ipps/WeinsbergDWA07
fatcat:iuldebnz7rb6vad3bjhg4s4kii
Performance Evaluation of a QoS-Aware Mechanism for Polling-Based High-Speed Network Interfaces
2005
Journal of Computer Science
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2017; you can also visit the original URL.
The file type is application/pdf.
The authors of [13] proposed Ethernet Message Passing (EMP), a completely new zero-copy, OS-bypass messaging layer for Gigabit Ethernet on Alteon NICs where the entire protocol processing is done at ...
OS bypass means that the OS is removed from the critical path of the message. ...
doi:10.3844/jcssp.2005.461.470
fatcat:ticrwogamngjbis4ps3eesxfxa
BUCS - a bottom-up cache structure for networked storage servers
2004
International Conference on Parallel Processing, 2004. ICPP 2004.
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2006; you can also visit the original URL.
The file type is application/pdf.
The bottom level cache is located on an embedded controller that is a combination of a network interface card (NIC) and a storage host bus adapter (HBA). ...
Storage data coming from or going to a network are cached at this bottom level cache and meta-data related to these data are passed to the host for processing. ...
EMP [15] also provides a zero-copy OS-bypass message passing interface by offloading the protocol to programmable NICs. ...
doi:10.1109/icpp.2004.1327937
dblp:conf/icpp/ZhangY04a
fatcat:2gfrhlaqfzh3tlgpvn4vtsozk4
High performance user level sockets over Gigabit Ethernet
Proceedings. IEEE International Conference on Cluster Computing
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2004; you can also visit the original URL.
The file type is application/pdf.
To the best of our knowledge, this is the first such design and implementation for Gigabit Ethernet. ...
Both the latency and the throughput numbers are close to those achievable by EMP. ...
To the best of our knowledge EMP is the only complete OS-bypass, zero-copy and NIC-driven protocol over Gigabit Ethernet. Thus, we focus our research on the EMP protocol. ...
doi:10.1109/clustr.2002.1137745
dblp:conf/cluster/BalajiSWP02
fatcat:6aaozkj4dfgv5lvb5e3ngrwacy
Experience in offloading protocol processing to a programmable NIC
Proceedings. IEEE International Conference on Cluster Computing
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2006; you can also visit the original URL.
The file type is application/pdf.
In this paper we describe our experience in offloading protocol processing to a programmable gigabit Ethernet network interface card. ...
For our experiments, we selected a simple RTS/CTS (request to send/clear to send) protocol called RMPP (Reliable Message Passing Protocol). ...
The Alteon Acenic NICs were borrowed from the Albuquerque High Performance Computing Center and were acquired as part of an IBM SUR (Sponsored University Research) grant. ...
doi:10.1109/clustr.2002.1137730
dblp:conf/cluster/MaccabeZOR02
fatcat:yd6ikja7nff2fgsisaocveujcm
Protocol-dependent message-passing performance on Linux clusters
Proceedings. IEEE International Conference on Cluster Computing
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2008; you can also visit the original URL.
The file type is application/pdf.
The NetPIPE utility is being used to compare the latency and throughput of all current message-passing libraries and the native software layers they run upon for a variety of hardware configurations. ...
The authors would like to thank Ricky Kendall, Shoba Selvarajan, and Weiyi Chen for their contributions to this project, and to MPI Software Technology for providing evaluation copies of their MPI/Pro ...
A few omissions include CHIMP/MPI [11] , which apparently has not been supported since 1997, EMP [12] , a NIC-driven message-passing system, and GPSHMEM [13] , a general-purpose implementation of the ...
doi:10.1109/clustr.2002.1137746
dblp:conf/cluster/TurnerC02
fatcat:6bil5oirwvgbngc325plrk24ay
« Previous
Showing results 1 — 15 out of 19 results