IFIP/ACM International Conference on Distributed systems platforms
Efficient wire formats for high performance computing
Proceedings of the 2000 ACM/IEEE conference on Supercomputing
A middleware toolkit for client-initiated service specialization
ACM SIGOPS Operating Systems Review
ENSEMBLE: A Communication Layer for Embedded Multi-Processor Systems
OM '01 Proceedings of the 2001 ACM SIGPLAN workshop on Optimization of middleware and distributed systems
Communication overlap in multi-tier parallel algorithms
SC '98 Proceedings of the 1998 ACM/IEEE conference on Supercomputing
Applying patterns to develop a pluggable protocols framework for ORB middleware
Design patterns in communications software
Supporting high-performance I/O in QoS-enabled ORB middleware
Cluster Computing
Event Services in High Performance Systems
Cluster Computing
High Performance Network of PC Cluster Maestro
Cluster Computing
Native Data Representation: An Efficient Wire Format for High-Performance Distributed Computing
IEEE Transactions on Parallel and Distributed Systems
Portals 3.0: Protocol Building Blocks for Low Overhead Communication
IPDPS '02 Proceedings of the 16th International Parallel and Distributed Processing Symposium
High Performance Implementation of MPI for Myrinet
ParNum '99 Proceedings of the 4th International ACPC Conference Including Special Tracks on Parallel Numerics and Parallel Computing in Image Processing, Video Processing, and Multimedia: Parallel Computation
High-speed I/O: the operating system as a signalling mechanism
NICELI '03 Proceedings of the ACM SIGCOMM workshop on Network-I/O convergence: experience, lessons, implications
Transport protocols for high performance
Communications of the ACM - Blueprint for the future of high-performance networking
Direct connect device core: design and applications
Integration, the VLSI Journal
Design and Evaluation of an HPVM-Based Windows NT Supercomputer
International Journal of High Performance Computing Applications
Publish-Subscribe for High-Performance Computing
IEEE Internet Computing
PDCN'06 Proceedings of the 24th IASTED international conference on Parallel and distributed computing and networks
RISC: A resilient interconnection network for scalable cluster storage systems
Journal of Systems Architecture: the EUROMICRO Journal
On chip novel video streaming system for bi-network multicasting protocols
Integration, the VLSI Journal
Parallel and distributed computing on multidomain non-routable networks
International Journal of High Performance Computing and Networking
A multi-protocol communication architecture for metacomputing
ICCOM'06 Proceedings of the 10th WSEAS international conference on Communications
Exploiting multidomain non routable networks
ISPA'06 Proceedings of the 4th international conference on Parallel and Distributed Processing and Applications
| Hi-index | 0.00 |
We describe our experience designing, implementing, and evaluating two generations of our high performance communication library, Fast Messages (FM) for Myrinet. In FM 1.x, we designed a simple interface and provided guarantees of reliable and in-order delivery, and flow control. While this was a significant improvement over previous systems, it was not enough. Layering MPI atop FM 1.x showed that only about 20\% of the FM 1.x bandwidth could be delivered to higher level communication APIs. Our second generation communication layer, FM 2.0, addresses the identified problems, providing gather-scatter, interlayer scheduling, receiver flow control, as well as some convenient API features which simplify programming. FM 2.x can deliver 70-90\% to higher level APIs such as MPI. This is especially impressive as the absolute bandwidths delivered have increased nearly fourfold to 70 MB/s. We describe general issues encountered in matching two communication layers, and our solutions as embodied in FM 2.x.