-
[show abstract]
[hide abstract]
ABSTRACT: Microarchitectural redundancy has been proposed as a means of improving chip lifetime reliability. It is typically used in a reactive way, allowing chips to maintain operability in the presence of failures by detecting and isolating, correcting, and/or replacing components on a first-come, first-served basis only after they become faulty. In this paper, we explore an alternative, more preferred method of exploiting microarchitectural redundancy to enhance chip lifetime reliability. In our proposed approach, redundancy is used proactively to allow non-faulty microarchitecture components to be temporarily deactivated, on a rotating basis, to suspend and/or recover from certain wearout effects. This approach improves chip lifetime reliability by warding off the onset of wearout failures as opposed to reacting to them posteriorly. Applied to on-chip cache SRAM for combating NBTI-induced wearout failure, our proactive wearout recovery approach increases lifetime reliability (measured in mean-time-to-failure) of the cache by about a factor of seven relative to no use of microarchitectural redundancy and a factor of five relative to conventional reactive use of redundancy having similar area overhead.
Computer Architecture, 2008. ISCA '08. 35th International Symposium on; 07/2008
-
[show abstract]
[hide abstract]
ABSTRACT: Component failures and planned component replacements cause changes in the topology and routing paths supplied by the interconnection network of a parallel processor system over time. Such changes may require the network to be reconfigured such that the existing routing function is replaced by one that enables packets to reach their intended destinations amid the changes. Efficient reconfiguration methods are desired which allow the network to function uninterruptedly over the course of the reconfiguration process while remaining free from deadlocking behavior. In this paper, we propose, evaluate, and prove the deadlock freedom of a new network reconfiguration protocol that overlaps various phases of "static" reconfiguration processes traditionally used in commercial and research systems to provide performance efficiency on par with that of recently proposed "dynamic" reconfiguration processes but without their complexity. Simulation results show that the proposed Overlapping Static Reconfiguration protocol can reduce reconfiguration time by up to 50 percent, reduce packet latency by several orders of magnitude, reduce packet dropping by an order of magnitude, and provide unhalted packet injection as compared to traditional static reconfiguration while allowing network throughput similar to dynamic reconfiguration.
IEEE Transactions on Computers 07/2008; 57(6):762-779. · 1.10 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: As the level of chip-integration continues to advance at a fast pace, the desire for efficient interconnects - whether on-chip or off-chip - is rapidly increasing. Traditional interconnects like buses, point-to-point wires, and regular topologies may suffer from poor resource sharing in the time and space domains, leading to high contention or low resource utilization. In this paper, we propose a design methodology for constructing networks for special-purpose computer systems with well-behaved (known) communication characteristics. A temporal and spatial model is proposed to define the sufficient condition for contention-free communication. Based upon this model, a design methodology using a recursive bisection technique is applied to systematically partition a parallel system such that the required number of links and switches is minimized while achieving low contention. Results show that the design methodology can generate more optimized on-chip networks with up to 60 percent fewer resources than meshes or tori while providing blocking performance closer to that of a fully connected crossbar.
IEEE Transactions on Parallel and Distributed Systems 03/2006; · 1.40 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: Efficient and reliable communication is essential for achieving high performance in a networked computing environment. Finite network resources bring about unavoidable competition among in-flight network packets, resulting in network congestion and, possibly, deadlock. Many techniques have been proposed to improve network performance by efficiently handling network congestion and potential deadlock. However, none of them provide an efficient way of accelerating the movement of network packets in congestion toward their destinations. In this paper, we propose a new mechanism for detecting and resolving network congestion and potential deadlocks. The proposed mechanism is based on efficiently tracking paths of congestion and increasing the scheduling priority of packets along those paths. This acts to throttle other packets trying to enter those congested regions - in effect, locking out packets from congested regions until congestion has had the opportunity to disperse. Simulation results show that the proposed technique effectively disperses network congestion and is also applicable in helping to resolve potential deadlock.
IEEE Transactions on Parallel and Distributed Systems 09/2005; 16(8):686- 701. · 1.40 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: For pt.I see ibid., vol.16, no.5, p.412-427 (2005). Dynamic network reconfiguration is defined as the process of changing from one routing function to another while the network remains up and running. The main challenge is in avoiding deadlock anomalies while keeping restrictions on packet injection and forwarding minimal. Current approaches either require virtual channels in the network or they work only for a limited set of routing algorithms and/or fault patterns. In this paper, we present a methodology for devising deadlock free and dynamic transitions between old and new routing functions that is consistent with newly proposed theory [J. Duato et al., (2005)]. The methodology is independent of topology, can be applied to any deadlock-free routing function, and puts no restrictions on the routing function changes that can be supported. Furthermore, it does not require any virtual channels to guarantee deadlock freedom. This research is motivated by current trends toward using increasingly larger Internet and transaction processing servers based on clusters of PCs that have very high availability and dependability requirements, as well as other local, system, and storage area network-based computing systems.
IEEE Transactions on Parallel and Distributed Systems 06/2005; · 1.40 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: This paper develops theoretical support useful for determining deadlock properties of dynamic network reconfiguration techniques and also serves as a basis for the development of design methodologies useful for deriving deadlock-free reconfiguration techniques. It is applicable to interconnection networks typically used in multiprocessor servers, network-based computing clusters, and distributed storage systems, and also has potential application to system-on-chip networks. This theory builds on basic principles established by previous theories while pioneering new concepts fundamental to the case of dynamic network reconfiguration.
IEEE Transactions on Parallel and Distributed Systems 06/2005; · 1.40 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: Support for arbitrary topologies has become more popular for system-area networks but very little has been done in trying to characterize their behavior and performance. Traditional parameters like diameter and bisection width are not sufficient for characterizing the irregularities that abound in such networks and fail to give much insight into throughput performance. A clustering approach for partitioning a network into clusters of richly-connected regions is proposed as a means of defining two performance-correlated characterization metrics: intercluster bandwidth index and intercluster link-cost index. The two characterization metrics are shown to have a strong correlation to saturation throughput when link and load distribution of a network is imbalanced. Simulation results also show that the clustering algorithm can be applied to a variety of network configurations and traffic scenarios, particularly irregular ones. With the proposed characterization metrics that correlate more strongly with performance, it is possible to classify networks into categories having similar performance.
IEEE Transactions on Parallel and Distributed Systems 01/2004; 14(12):1222- 1239. · 1.40 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: InfiniBand is becoming an industry standard both for communication between processing nodes and I/O devices, and for interprocessor communication. Instead of using a shared bus, InfiniBand employs an arbitrary (possibly irregular) switched point-to-point network. InfiniBand specification defines a basic management infrastructure that is responsible for subnet configuration, activation, and fault tolerance. After the detection of a topology change, management entities collect the current subnet topology. The topology discovery algorithm is one of the management issues that are outside the scope of the current specification. Preliminary implementations obtain the entire topological information each time a change is detected. In this work, we present and analyze an optimized implementation, based on exploring only the region that has been affected by the change.
Cluster Computing, 2003. Proceedings. 2003 IEEE International Conference on; 01/2004
-
[show abstract]
[hide abstract]
ABSTRACT: The InfiniBand architecture is a high-performance network technology for the interconnection of processor nodes and I/O devices using a point-to-point switch-based fabric. The InfiniBand specification defines a basic management infrastructure that is responsible for subnet configuration, activation, and fault tolerance. Subnet management entities and functions are described, but the specifications do not impose any particular implementation. We present and analyze a complete subnet management mechanism for this architecture. We allow to anticipate future directions to obtain efficient management protocols
Parallel Processing, 2003. Proceedings. 2003 International Conference on; 11/2003
-
[show abstract]
[hide abstract]
ABSTRACT: Dynamic network reconfiguration is defined as the change from one routing function to another while the network is up and running. The main challenge is avoidance of deadlocks, while keeping restrictions on packet injection and forwarding minimal. Current approaches either require virtual channels in the network, or they work only for a limited set of routing algorithms. We present a methodology for devising deadlock free and dynamic transitions between an old and a new routing function. The methodology is independent of topology and puts no restrictions on either routing function. Furthermore, it does not require any virtual channels to guarantee deadlock freedom. This research is motivated by the current trend toward using increasingly larger Internet servers based on clusters of PCs and the very high availability requirements of those as well as other local, system, and storage area network-based systems
Parallel Processing, 2003. Proceedings. 2003 International Conference on; 11/2003
-
[show abstract]
[hide abstract]
ABSTRACT: Network-based parallel computing systems often require the ability to reconfigure the routing algorithm to reflect changes in network topology if and when voluntary or involuntary changes occur. The process of reconfiguring a network's routing capabilities may be very inefficient and/or deadlock-prone if not handled properly. We propose efficient and deadlock-free dynamic reconfiguration schemes that are applicable to routing algorithms and networks which use wormhole, virtual cut-through, or store-and-forward switching, combined with hard link-level flow control. One requirement is that the network architecture use virtual channels or duplicate physical channels for deadlock-handling as well as performance purposes. The proposed schemes do not impede the injection, transmission, or delivery of user packets during the reconfiguration process. Instead, they provide uninterrupted service, increased availability/reliability, and improved overall quality-of-service support as compared to traditional techniques based on static reconfiguration.
IEEE Transactions on Parallel and Distributed Systems 09/2003; 14(8):780- 794. · 1.40 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: Handling deadlocks is essential for providing reliable communication paths between processing nodes in parallel computer systems. The existence of multiple message types and associated inter-message dependencies may cause message-dependent deadlocks in networks that are designed to be free of routing deadlock. Most methods currently used for dealing with message-dependent deadlocks require more system resources than are necessary and/or do not use system resources efficiently. This may have an adverse effect on system performance if resources are scarce. In this paper, we characterize the frequency of message-dependent deadlocks in multiprocessor/multicomputer systems. We also propose a handling technique for message-dependent deadlocks based on progressive deadlock recovery and evaluate its performance with other approaches. Results show that message-dependent deadlocks occur very infrequently under typical circumstances thus, rendering approaches based on avoiding them overly restrictive in the common case. The proposed technique relaxes restrictions considerably, allowing the routing of packets and the handling of message-dependent deadlocks to be much more efficient-particularly when network resources are scarce.
IEEE Transactions on Parallel and Distributed Systems 04/2003; 14(3):259- 275. · 1.40 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: As the level of chip integration continues to advance at a fast pace, the desire for efficient interconnects - whether on-chip or off-chip - is rapidly increasing. Traditional interconnects like buses, point-to-point wires and regular topologies may suffer from poor resource sharing in the time and space domains, leading to high contention or low resource utilization. In this paper, we propose a design methodology for constructing networks for special-purpose computer systems with well-behaved (known) communication characteristics. A temporal and spatial model is proposed to define the sufficient condition for contention-free communication. Based upon this model, a design methodology using a recursive bisection technique is applied to systematically partition a parallel system such that the required number of links and switches is minimized while achieving low contention. Results show that the design methodology can generate more optimized on-chip networks with up to 60% fewer resources than meshes or tori while providing blocking performance closer to that of a fully connected crossbar.
High-Performance Computer Architecture, 2003. HPCA-9 2003. Proceedings. The Ninth International Symposium on; 03/2003
-
[show abstract]
[hide abstract]
ABSTRACT: Efficient and reliable communication is essential for achieving high performance in a networked computing environment. Limited network resources bring about unavoidable competition among in-flight packets, resulting in network congestion and possibly deadlock. Many techniques have been proposed to improve performance by efficiently handling network congestion and deadlock. However, none of them provide an efficient way of accelerating the movement of packets involved in congestion onward to their destinations. In this paper, we propose a new mechanism for the detection and resolution of network congestion and deadlocks. The proposed mechanism is based on increasing the scheduling priority of packets involved in congestion and providing necessary resources for those packets to make forward progress. Simulation results show that the proposed technique outperforms previously proposed techniques by effectively dispersing network congestion.
Parallel Processing, 2002. Proceedings. International Conference on; 02/2002
-
[show abstract]
[hide abstract]
ABSTRACT: This paper presents a theoretical framework for the design of deadlock-free fully adaptive routing algorithms for a general class of network topologies and switching techniques in a single, unified theory. A general theory is proposed that allows the design of deadlock avoidance-based as well as deadlock recovery-based wormhole and virtual cut-through adaptive routing algorithms that use a homogeneous or a heterogeneous (mixed) set of resources. The theory also allows channel queues to be allocated nonatomically, utilizing resources efficiently. A general methodology for the design of fully adaptive routing algorithms applicable to arbitrary network topologies is also proposed. The proposed theory and methodology allow the design of efficient network routers that require minimal resources for handling infrequent deadlocks
IEEE Transactions on Parallel and Distributed Systems 01/2002; 12(12):1219-1235. · 1.40 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: The existence of multiple message types and associated
inter-message dependencies may cause message-dependent deadlock in
networks that are designed to be free of routing deadlock. Most methods
currently used for dealing with message-dependent deadlocks require more
system resources than are necessary and/or do not use system resources
efficiently. This may have an adverse effect on system performance if
resources are scarce. In this paper, we evaluate different approaches
for handling message-dependent deadlocks, and we propose an alternative
technique based on progressive deadlock recovery. Results show that the
proposed technique relaxes restrictions considerably, allowing the
routing of packets and handling of message-dependent deadlocks to be
much more efficient-particularly when network resources are scarce
Parallel and Distributed Processing Symposium., Proceedings 15th International; 05/2001
-
[show abstract]
[hide abstract]
ABSTRACT: This paper presents a theoretical model of resource allocations
and dependencies in wormhole and virtual cut-through interconnection
networks. This model allows various types of message blocking to be
described precisely, including deadlock. The model distinguishes between
messages involved in deadlock and those simply dependent upon deadlock,
thus establishing a framework for evaluating the accuracy and
correctness of deadlock detection mechanisms. The paper also identifies
the necessary and sufficient conditions for the occurrence and
resolution of deadlock in interconnection networks, thus providing
efficiency and correctness criteria for deadlock resolution mechanisms.
Theorems derived from the model are related to various routing
algorithms which are based on deadlock recovery
IEEE Transactions on Parallel and Distributed Systems 04/2000; · 1.40 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: Network-based computing systems often require the ability to reconfigure the routing algorithm to reflect changes in network topology if and when those changes occur. The process of reconfiguring a network's routing capabilities may lead to deadlock if not handled properly. In this paper we propose efficient and deadlock-free dynamic reconfiguration techniques that are generically applicable to distributed routing algorithms and networks, including those which use wormhole switching. The proposed techniques do not impede the transmission of packets during the reconfiguration process, thus providing increased network availability and quality-of-service (QoS) support as compared to traditional techniques based on static reconfiguration
Parallel Processing, 2000. Proceedings. 2000 International Conference on; 02/2000
-
[show abstract]
[hide abstract]
ABSTRACT: A spate of deadlock avoidance-based and deadlock recovery-based
routing algorithms have been proposed in recent years without full
understanding of the likelihood and characteristics of actual deadlocks
in interconnection networks. This work models the interrelationships
between routing freedom, message blocking, correlated resource
dependencies, and deadlock formation. It is empirically shown that
increasing routing freedom, as achieved by allowing unrestricted routing
over multiple physical and virtual channels, reduces the probability of
deadlocks and the likelihood of other types of correlated message
blocking that can degrade performance. Moreover, when true fully
adaptive routing is used in k-ary n-cube networks with two or more
virtual channels (wormhole OF virtual cut-through switched), it is
empirically shown that deadlocks are virtually eliminated in networks
with n⩾2. These results indicate that deadlocks are very infrequent
when the network and routing algorithm inherently provide sufficient
routing freedom, thus increasing the viability of deadlock recovery
routing strategies
IEEE Transactions on Parallel and Distributed Systems 10/1999; · 1.40 Impact Factor
-
T.M. Pinkston
[show abstract]
[hide abstract]
ABSTRACT: The development of fully adaptive, cut-through (wormhole) networks
is important for achieving high performance in communication-critical
parallel processor systems. Increased flexibility in routing allows
network bandwidth to be used efficiently, but also creates more
opportunity for cyclic resource dependencies to form which can cause
deadlock. If not guarded against, deadlocks in routing make packets
block in the network indefinitely and, eventually, could result in the
entire network coming to a complete standstill. The paper presents a
simple, flexible, and efficient routing approach for multicomputer
interconnection networks which is based on progressive deadlock recovery
as proposed to deadlock avoidance or regressive deadlock recovery.
Performance is optimized by allowing the maximum routing freedom
provided by network resources to be exploited. True fully adaptive
routing is supported in which all physical and virtual channels at each
node in the network are available to packets without regard for
deadlocks. Deadlock cycles, upon forming, are efficiently broken in
finite time by progressively routing one of the blocked packets through
a connected, deadlock-free recovery path. This routing approach enables
the design of high-throughput networks that provide excellent
performance. Simulations indicate that progressive deadlock recovery
routing can improve throughput by as much as 45 percent and 25 percent
over leading deadlock avoidance-based and regressive recovery-based
routing schemes, respectively
IEEE Transactions on Computers 08/1999; · 1.10 Impact Factor
