Download:
|
by Yvonne Coady, Joon Suan Ong, Michael J. Feeley
In Proceedings of The Eighth IEEE International Symposium on High Performance Distributed Computing
http://www.cs.ubc.ca/~feeley/papers/gmsnp_hpdc.ps
Add To MetaCart
Abstract:
Advances in network technology continue to improve the communication performance of workstation and PC clusters, making high-performance workstation-cluster computing increasingly viable. These hardware advances, however, are taxing traditional host-software network protocols to the breaking point. A modern gigabit network can swamp a host's IO bus and processor, limiting communication performance and slowing computation unacceptably. Fortunately, host-programmable network processors used by these networks present a potential solution. Offloading selected host processing to these embedded network processors lowers host overhead and improves latency. This paper examines the use of embedded network processors to improve the performance of workstation-cluster global memory management. We have implemented a revised version of the GMS global memory system that eliminates host overhead by as much as 29 % on active nodes and improves page fault latency by as much as 39%.
Citations
|
784
|
Myrinet: A Gigabit-per-second Local Area Network
– Boden, Cohen, et al.
- 1995
|
|
318
|
The Stanford FLASH Multiprocessor
– Kuskin, Ofelt, et al.
- 1994
|
|
153
|
BIP: a new protocol designed for high performance networking on myrinet
– Prylli, Tourancheau
- 1998
|
|
98
|
The DASH Prototype: Logic Overhead and Performance
– Lenoski, Laudon, et al.
- 1993
|
|
52
|
Thekkath. Implementing Global Memory Management in a Workstation Cluster
– Feeley, Morgan, et al.
- 1995
|
|
52
|
Implementing cooperative prefetching and caching in a global memory system
– Voelker, Andreson, et al.
- 1998
|
|
48
|
Cheating the I/O Bottleneck: Network Storage with Trapeze/Myrinet
– Anderson, Chase, et al.
- 1998
|
|
45
|
Eicken, “Incorporating memory management into user-level network interfaces,” http://www2.cs.cornell.edu/UNet/papers/unetmm.pdf
– Basu, Welsh, et al.
- 1996
|
|
35
|
Reducing network latency using subpages in a global memory environment
– Jamrozik, Voelker, et al.
- 1996
|
|
31
|
Cut-through delivery in trapeze: an exercise in low-latency messaging
– Yocum, Chase, et al.
- 1997
|
|
27
|
Early Experience with Message-Passing on the SHRIMP Multicomputer
– Felten, Alpert, et al.
- 1996
|
|
26
|
Spine: a safe programmable and integrated network environment
– Fiuczynski, Martin, et al.
- 1998
|
|
19
|
SPINE: An Operating System for Intelligent Network Adapters
– Fiuczynski, Bershad, et al.
- 1998
|
|
12
|
Active Messages: Organization and Applications Programming Interface
– Mainwaring, Culler
- 1995
|
|
10
|
Modeling and optimizing communication pipelines
– Wang, Krishnamurthy, et al.
- 1998
|
|
5
|
A front-end telnet/rlogin server implementation
– Powers
- 1986
|
|
4
|
Trapeze messaging API
– Chase, Gallatin, et al.
- 1997
|
