Hardware Implementation of Parallel Algorithm for Setting Up Benes Networks

نویسندگان

  • Yikun Jiang
  • Mei Yang
چکیده

Both Benes and Clos networks are rearrangeably non-blocking multi-stage interconnection networks. Benes network is a special case of Clos network which has inputs and outputs. The Benes network is constructed with switching nodes recursively. Due to their non-blocking property and relative smaller number of crosspoints, Benes/Clos networks have received much attention in both academia and industry. Benes/Clos networks have been used in many areas, such as interconnection network in parallel computers, multiprocessors system [1], and networks-on-chip [2][3][4][12][13]. In packet switching systems, the switch fabric must be able to provide internally conflict-free paths for the requesting packets in each time slot [5]. This is implemented by setting the states of all switches in the network. It is clear that the routing assignment (i.e., switch setting) scheme in Benes/Clos networks has a strong impact to the efficiency of the Bene/Clos networks. A number of switch setting algorithms have been developed in the past few decades, including sequential algorithms and parallel algorithms. Sequential algorithms such as looping algorithms [7] are designed for circuit switching systems where the switching configuration can be rearranged at relatively low speed. In [7], a switch setting algorithm with time complexity is proposed based on Waksman proof. As a matter of fact, using sequential algorithm, the N×N Benes network cannot be set up in less than time, because there are switches. The set-up time is much longer than the latency in Benes networks, which is for N×N network. In order to obtain a switch setting algorithm of complexity comparable to the network latency, parallel algorithms are needed. In [9], Nassimi and Sahni developed a parallel set-up algorithm which runs significantly faster than the sequential

برای دانلود رایگان متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Parallel Routing Algorithms in Benes and Clos Networks: A Survey

Benes and Clos networks are an important class of multistage interconnection network due to their expandability and modularity. In this article we present a survey of parallel routing algorithms in Benes and Clos Networks. These parallel routing algorithms are compared in terms of their interconnection topologies, implementation and time complexity. Index Terms : Multistage Interconnection netw...

متن کامل

A Fast Parallel Routing Algorithm for Benes Group Switches

A parallel routing algorithm for controlling the class of interconnection networks called group connectors is presented. Given any legal mapping from input to output groups with busy inputs, this algorithm can determine the switch setting of a Benes group connector with inputs and output groups in time on a completely connected computer or the EREW PRAM model with processing elements. The imple...

متن کامل

Arbitrary Size Benes Networks

The Benes network is a rearrangeable nonblocking network which can realize any arbitrary permutation. Overall, the r-dimensional Benes network connects 2 r inputs to 2 r outputs through 2r ? 1 levels of 2 2 switches. Each level of switches consists of 2 r?1 switches, and hence the size of the network has to be a power of two. In this paper, we extend Benes networks to arbitrary sizes. We also s...

متن کامل

High Performance Implementation of Fuzzy C-Means and Watershed Algorithms for MRI Segmentation

Image segmentation is one of the most common steps in digital image processing. The area many image segmentation algorithms (e.g., thresholding, edge detection, and region growing) employed for classifying a digital image into different segments. In this connection, finding a suitable algorithm for medical image segmentation is a challenging task due to mainly the noise, low contrast, and steep...

متن کامل

High Performance Implementation of Fuzzy C-Means and Watershed Algorithms for MRI Segmentation

Image segmentation is one of the most common steps in digital image processing. The area many image segmentation algorithms (e.g., thresholding, edge detection, and region growing) employed for classifying a digital image into different segments. In this connection, finding a suitable algorithm for medical image segmentation is a challenging task due to mainly the noise, low contrast, and steep...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2016