Design choices in the SHRIMP system: an empirical study
- 27 November 2002
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 330-341
- https://doi.org/10.1109/isca.1998.694792
Abstract
The SHRIMP cluster-computing system has progressed to a point of relative maturity; a variety of applications are running on a 16-node system. We have enough experience to understand what we did right and wrong in designing and building the system. In this paper we discuss some of the lessons we learned about computer architecture, and about the challenges involved in building a significant working system in an academic research environment. We evaluate significant design choices by modifying the network interface firmware and the system software in order to empirically compare our design to other approaches.Keywords
This publication has 34 references indexed in Scilit:
- A high-performance, memory-based interconnection system for multicomputer environmentsPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2002
- Fast RPC on the SHRIMP Virtual Memory Mapped Network InterfaceJournal of Parallel and Distributed Computing, 1997
- An implementation of the Hamlyn sender-managed interface architecturePublished by Association for Computing Machinery (ACM) ,1996
- Performance evaluation of two home-based lazy release consistency protocols for shared virtual memory systemsPublished by Association for Computing Machinery (ACM) ,1996
- Early experience with message-passing on the SHRIMP multicomputerACM SIGARCH Computer Architecture News, 1996
- Myrinet: a gigabit-per-second local area networkIEEE Micro, 1995
- The NX message passing interfaceParallel Computing, 1994
- The DASH prototype: Logic overhead and performanceIEEE Transactions on Parallel and Distributed Systems, 1993
- Memory coherence in shared virtual memory systemsACM Transactions on Computer Systems, 1989
- Performing remote operations efficiently on a local computer networkCommunications of the ACM, 1982