Power and Latency Optimized Deadlock-Free Routing Algorithm on Irregular 2D Mesh NoC using LBDRe

Author(s):  
Renu Verma ◽  
Mohammad Ayoub Khan ◽  
Amit Zinzuwadiya

Efficient routing is challenging and crucial problem in the irregular mesh NoC topologies because of increasing hardware cost and routing tables. In this paper, the authors propose an efficient deadlock-free routing algorithm for irregular mesh NoCs which reduces the latency and power consumption significantly. The problem with degree priority based routing algorithm is that it cannot remove deadlocks in irregular mesh topologies. Therefore, the authors use the extended Logic Based Distributed Routing (LBDRe) to remove deadlock situations without using any virtual channel in the degree priority based routing algorithm. The proposed LBDRe based technique also removes the dependency on routing tables. The authors further apply odd-Even routing algorithm to LBDRe to ensure that some turns are prohibited to remove deadlocks. Experimental results show that the proposed routing algorithm reduces power consumption by 9–22% and overall average latency by 8–12% with the minimum hardware cost for the irregular mesh NoC topologies.

2011 ◽  
Vol 474-476 ◽  
pp. 413-416
Author(s):  
Jia Jia ◽  
Duan Zhou ◽  
Jian Xian Zhang

In this paper, we propose a novel adaptive routing algorithm to solve the communication congestion problem for Network-on-Chip (NoC). The strategy competing for output ports in both X and Y directions is employed to utilize the output ports of the router sufficiently, and to reduce the transmission latency and improve the throughput. Experimental results show that the proposed algorithm is very effective in relieving the communication congestion, and a reduction in average latency by 45.7% and an improvement in throughput by 44.4% are achieved compared with the deterministic XY routing algorithm and the simple XY adaptive routing algorithm.


1999 ◽  
Vol 09 (01n02) ◽  
pp. 99-112
Author(s):  
NOZOMU TOGAWA ◽  
KAORU UKAI ◽  
MASAO YANAGISAWA ◽  
TATSUO OHTSUKI

This paper proposes a simultaneous placement and global routing algorithm for FPGAs with power optimization. The algorithm is based on hierarchical bipartitioning of layout regions and sets of logic-blocks. When bipartitioning a layout region, pseudo-blocks are introduced to preserve connections if there exist connections between bipartitioned logic-block sets. A global route is represented by a sequence of pseudo-blocks. Since pseudo-blocks and logic-blocks can be dealt with equally, placement and global routing are processed simultaneously. The algorithm gives weights to nets with high switching probabilities and attempts to assign the blocks connected by weighted nets to the same region. Thus their length is shortened and the power consumption of a whole circuit can be reduced. The experimental results demonstrate the effectiveness and efficiency of the algorithm.


Author(s):  
Mehdi Mohtashamzadeh ◽  
Ladan Momeni ◽  
Arshin Rezazadeh ◽  
Shaghayegh Baghbani ◽  
Asma Mahdipoor

Author(s):  
V. Samadi Bokharaei ◽  
A. Shamaei ◽  
H. Sarbaziazad ◽  
M. Abbaspour

1970 ◽  
Vol 5 (1) ◽  
Author(s):  
Shariq Haseeb Khairul Azami Sidek Ahmad Faris Ismail, Lai W.K. ◽  
Aw Yit Mei

Successful implementation and operation of a network largely depends on the routing algorithm in use. To date, several routing algorithms are in use but the problem with these algorithms is that they are either not adaptive or not robust enough, thus limiting the proper use of bandwidth.  AntNet is an innovative algorithm that may be used for data networks. It is a combination of both static and dynamic routing algorithms. In this algorithm, a group of mobile agents (compared to real ants) form paths between source and destination nodes. They explore the network continuously and exchange obtained information indirectly, in order to update the routing tables at different nodes. Our version of AntNet (hereinafter referred to as AntNet2.0) has been improved to overcome the problems with other algorithms. This paper compares the performance of AntNet2.0 against two other commercially popular algorithms, viz. link state routing algorithm and distant vector routing algorithm. The performance matrix used to compare the algorithms is based on average throughput, packet loss, packet drop and end-to-end delay. Convergence time for this algorithm on a nation-wide telecommunications network will also be discussed. Conclusions and areas of further work will also be presented in lucid manner, so that it may be transformed into real practice in the future.Key Words: mobile agents, swarm intelligence, networks and constant bit rate


2012 ◽  
Vol 182-183 ◽  
pp. 427-430
Author(s):  
Li Feng Wei ◽  
Liang Cheng ◽  
Xing Man Yang

A adaptive control method of the pulse demagnetizer was presented, Can adjust the strength of the charge current automatically according to the changes of the magnetic content to ensure the constant of the magnetic field.The experimental results have shown that it has the advantages of low power consumption, strong anti-interference capability, stable and reliable operation, long life and good demagnetizing effect, when compared to the conventional demagnetizers.


Sign in / Sign up

Export Citation Format

Share Document