Demonstration venue · read-only. Every page can be browsed; the buttons that would change it are switched off. Create an account to run TaxoReview on your own data. A simulation from public publication records: the people listed were not invited and did not agree to review.
? Tutorial for this page

Honda Shing

dblp:62/4459 · DBLP profile ↗
← Back
5ranked-venue papers
3as first author
0since 2021 · last 1993
—ORCID · none

Domains — the database's venue-derived domains; a paper can count in several

Systems, architecture and hardware · 5 · 3 first-author

Expertise — from the expertise taxonomy: the topics of the expert's papers under the CCF categories. A weight counts papers with recency: 1 for a paper about the topic, 0.3 when the topic is its context, halved every five years.

Computer architecture, parallel and distributed computing, and storage systems
2 papers
Parallel and multicore computing · 37% Memory systems · 21% Interconnection networks and networks-on-chip · 21%

Topics — the 4 heaviest of 5, each with the papers that count most for it

TopicWeightPapersLastEvidence papers
Interconnection networks and networks-on-chip
multiprocessor interconnection
0.011991
A conflict-free memory design for multiprocessors · SC 1991
Distributed systems › concurrency control
process synchronization
0.011991
A conflict-free memory design for multiprocessors · SC 1991
Parallel and multicore computing
parallel programming models
0.011990
Resource binding - a universal approach to parallel programming · SC 1990
Parallel and multicore computing › synchronization
synchronization mechanisms
0.011990
Resource binding - a universal approach to parallel programming · SC 1990

Methods — techniques the papers use, named apart from their topics

conflict-free memory design · 0.0resource binding · 0.0
YearPublicationVenuePosition
1993 A Novel Approach to the Design of Scalable Shared-Memory Multiprocessors
abstract
This paper presents the Conflict-Free Mem ory (CFM) architecture for designing scalable sharedmemory multiprocessors. For a single-level CFM archi tecture, the architecture improves multiprocessor perfor mance by eliminating memory and interconnection net work contention and reducing network latency. A hierar chical CFM architecture is introduced in this paper, which presents a new approach in implementing scalable sharedmemory multiprocessors. An invalidation-based write-back cache protocol is proposed for the CFM architecture. This cache coherence protocol preserves the low storage overhead of snoopy cache protocols, while it offers the high scalability of directory-based protocols. Furthermore, with the CFM cache protocol, efficient synchronization operations can be implemented.
Honda Shing, Lionel M. Ni
ICPP (1)1
1991 A conflict-free memory design for multiprocessors
abstract
The authors introduce a novel scheme of shared memory organization and interconnection network structure which supports conflict-free accesses to the shared memory in multiprocessors. This scheme also provides a low latency and low cost mechanism for process synchronization. The hot spot problem observed in many multiprocessors can easily be eliminated with this memory organization and interconnection network structure. Extensions to the architecture for supporting large-scale multiprocessors are discussed. It is demonstrated how data consistency can be maintained among concurrent accesses to the same location in a shared memory. In addition, higher level process synchronization supports based on the conflict-free memory architecture are explained.
Honda Shing, Lionel M. Ni
SC1
1991 Resource Contention in Shared-Memory Multiprocessors: A Parameterized Performance Degradation Model
Arun K. Nanda, Honda Shing, Ten H. Tzen, Lionel M. Ni
J. Parallel Distributed Comput.2
1990 A Replicate Workload Framework to Study Performance Degradation in Shared-Memory Multiprocessors
Arun K. Nanda, Honda Shing, Ten H. Tzen, Lionel M. Ni
ICPP (1)2
1990 Resource binding - a universal approach to parallel programming
abstract
The authors present a parallel programming paradigm, resource binding, which offers an architecture-independent environment for various parallel computation models. The resource binding technique, with a simple set of primitives, enables a flexible and consistent way of handling process synchronization, shared resource management, and other problems in parallel programming. High portability can be maintained without losing performance. For a clear introduction to the technique, comparisons are made among resource binding and other well-defined schemes such as semaphore, monitor, message passing, and Linda.>
Honda Shing, Lionel M. Ni
SC1