|
|---|
|
|
|
|
|
|
d**y
发帖数: 32 | 1 两篇文章,请方向相符及有兴趣的发站内信，并请附上简要个人介绍（学历、研究方向
）及
email联系方式，谢谢。
1: Stochastic Control Mechanism on Routing Multiclass Switching Network
Through Game theory
ABSTRACT
Over the last few decades stream control mechanism has more attention in the
early stages of research in communication network. The motivation of this
work is deep understanding on the stream control mechanism problem in multi-
class networks. In this paper, game theoretic perspectives are presented and
the appropriate frame work for the study of stream control mechanism
problem is analysed. Consider a min-max routing problem, where the control
mechanism has to decide to which of N queues that the arriving customer
should be sent. The service rate in each queue is dependent on the state of
the system, may change in time and is unknown to the control mechanism. The
goal of the control mechanism is to design an efficient policy which
guarantees the best performance under the worst case service conditions. ie.
, Arriving customers are routed by a control mechanism, with the purpose of
minimizing the total discounted holding cost under worst-case service
conditions. Then the problem is viewed as a zero-sum (Stochastic)Markov game
between the routing control mechanism and a service control mechanism. In
zero-sum (Stochastic)Markov game, where the server acts as player 1 and the
stream control mechanism acts as a player 2. Each player is assumed to have
the information of all the previous actions of both players as well as the
current and the past states of the system. The main results obtained are to
identify the optimal strategy for both players. A value iteration technique
is used to establish properties of the value of the game, which are related
to super modularity and convexity.
2: Recalling Instructions from Idling Threads to Maximize Resource
Utilization for Simultaneous Multi-Threading (SMT) Processors
Abstract
Simultaneous multi-threading (SMT) has been a very popular design in
improving resource utilization by sharing key datapath components among
multiple independent threads. When critical resources are shared by multiple
threads, effective use of these resources proves to be the most
important factor in fully exploiting the system potential. Allowing any of
the threads to overwhelm these shared resources not only leads to unfair
thread processing but may also result in severely degraded overall
performance. How to prevent idling threads from clogging the critical
resources in the pipeline becomes a must in sustaining desired system
performance. In this paper, we show that, if one can manage to recall
instructions of idling threads from Issue Queue (IQ), a very important
shared resource in the SMT pipeline, the system performance is easily
enhanced by a significant margin, with up to 20% for some benchmark mixes.
An even more noteworthy feature about this technique is that the ensuing
hardware overhead is very insignificant and its effect on clock timing is
also negligible. In addition, this proposed feature can also be easily
coupled with other advanced techniques employed
in other stages of the SMT pipeline for potentially additive benefits. |
|
|
|
|
|
|
|
