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

R. Krishnamurti

dblp:46/618 · DBLP profile ↗
← Back
2ranked-venue papers
0as first author
0since 2021 · last 1984
—ORCID · none

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

Systems, architecture and hardware · 2Artificial intelligence and machine learning · 1Software engineering, systems software and programming languages · 1

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
Interconnection networks and networks-on-chip · 83% Reconfigurable computing and FPGAs · 8% Performance modeling and evaluation · 8%

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

TopicWeightPapersLastEvidence papers
Interconnection networks and networks-on-chip › switch architecture
crossbar switch
0.011984
The Architecture of REPLICA-A Special-Purpose Computer System for Active Multi-Sensory Perception of 3-Dimensional Objects · ISCA 1984
Interconnection networks and networks-on-chip › interconnect architecture
reconfigurable interconnect
0.011984
REPLICA-A reconfigurable partitionable highly parallel computer architecture for active multi-sensory perception of 3-dimensional objects · ICRA 1984
Interconnection networks and networks-on-chip
ring network
0.011984
The Architecture of REPLICA-A Special-Purpose Computer System for Active Multi-Sensory Perception of 3-Dimensional Objects · ISCA 1984
Performance modeling and evaluation › delay analysis
message delay analysis
0.011984
The Architecture of REPLICA-A Special-Purpose Computer System for Active Multi-Sensory Perception of 3-Dimensional Objects · ISCA 1984
Reconfigurable computing and FPGAs
reconfigurable architecture
0.011984
The Architecture of REPLICA-A Special-Purpose Computer System for Active Multi-Sensory Perception of 3-Dimensional Objects · ISCA 1984

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

performance attribute evaluation · 0.0crossbar switch · 0.0bidirectional ring network · 0.0
YearPublicationVenuePosition
1984 REPLICA-A reconfigurable partitionable highly parallel computer architecture for active multi-sensory perception of 3-dimensional objects
abstract
REPLICA is a special-purpose computer architecture for active multi-sensory perception of 3-D objects. Active multisensory perception of 3-D objects means the ability to manipulate and probe objects as well as to see, hear, and touch them, where the objects involved may be static, changing or in motion. It is becoming an important area in robotics research. In order to support active multi-sensory perception efficiently, the system must be highly parallel, reconfigurable, and partitionable. The ability to achieve these functional requirements depends strongly on the capabilities and flexibilities of the underlying interconnection networks. In this paper, the architecture of REPLICA is presented and a set of interconnection networks is proposed. This set consists of a capability-enhanced cross-bar switch for controller-processor communications and a set of bidirectional rings for processor-processor communications. This set of interconnection networks support the required flexibilities in partitioning and reconfiguration. The cost of these networks can also be kept low.
R. Krishnamurti
ICRA2
1984 The Architecture of REPLICA-A Special-Purpose Computer System for Active Multi-Sensory Perception of 3-Dimensional Objects
abstract
REPLICA is a special-purpose computer architecture to support active multi-sensory perception of 3-D objects. The system is partitionable, reconfigurable, and highly parallel. In this paper, we present the design of a set of interconnection networks which form the core of the REPLICA architecture. This set of interconnection networks consists of a capability-enhanced cross-bar switch and multiple fast shift-register rings. The capability-enhanced cross-bar switch supports rapid set-up of partitions, point-to-point communication and broadcasting communication between the controller and processors in a partition. The multiple rings support high bandwidth communications among the processors and also provide the partitionability and reconfigurability required by the system. The cost of these interconnection networks can also be kept low. The communication delay in the ring network is analyzed by deriving an upper bound on the maximum communication delay in a partition. A set of performance attributes for evaluating the performance of reconfigurable, partitionable, and highly parallel computer systems is proposed. Based on this set of attributes, the design of the REPLICA architecture is evaluated and compared with the other reconfigurable, partitionable, and highly parallel computer architectures in the literature.
R. Krishnamurti
ISCA2