Parallel Routing Algorithm for Extra Level Omega Networks on Reconfigurable Systems

A new parallel algorithm for route assignment in Benes-Clos network is studied. In packet switching systems, switch fabrics must be able to provide internally conflict-free paths simultaneously and to accommodate packets requesting for connections in real-time as they arrive at the inputs. Most known sequential route assignment algorithms, such as the looping algorithm for Benes (1962) networks or Clos (1953) networks, are designed for circuit switching systems where the switching configuration can be rearranged at a relatively low speed. Most existing parallel routing algorithms are not practical for packet switching because they either assume the set of connection requests is a full permutation or fail to deal with output contentions among the set of input packets. We develop a parallel routing algorithm by solving a set of Boolean equations which are derived from the connection requests and the symmetric structure of the Benes network. Our approach can handle both the partial permutations and the output contention problem easily. The time complexity of our algorithm is O(log/sup 2/N), where N is the network size. Furthermore, we extend the algorithm and show that it can be applied to the Clos network if the number of central modules is a power of two.

In this paper, we propose three different parallel algorithms based on a state-of-the-art global router called TimberWolfSC. The parallel algorithms have been implemented by using the Message Passing Interface (MPI), and have been evaluated on a wide range of parallel platforms such as the Sun SparcCenter 1000 and the Intel Paragon. Our experimental results show good speedups and qualities from two of these parallel algorithms. We have been able to reduce runtimes of some circuits from half an hour to 5 minutes, obtained speedups of about 4.0 to 5.0 on 8 processors, with less than 2-3% degradation of quality of the solutions.

Fast placement and routing by extending coarse-grained reconfigurable arrays with Omega Networks

Chimpanzee females of East and West African populations differ in the average number of cycles per interbirth interval. Whereas females in Gombe, Mahale, and Kibale (eastern chimpanzees, Pan troglodytes schweinfurthii) average <9 cycles before they conceive, females at Tai (western chimpanzees; Pan troglodytes verus) average 29 cycles. We examined data from 2 different study groups (North and South) at Tai. By showing that Tai females interrupt cyclic activity for, on average, 7.4 mo between the end of the postpartum amenorrhea period (PPA) and the subsequent conception, and by calculating the number of cycles to conception based on a probability distribution of cycles over the interbirth interval, we show that Tai females average 19.4 cycles (North Group) and 11.7 cycles (South Group) to conception; therefore, the earlier calculation of 29 cycles was an overestimate. Further, at Tai young parous females have a significantly shorter PPA than those of older females, but the number of cycles to conception does not differ significantly between primiparous and nulliparous females. Some of our results are therefore not in line with the predictions of the cost-of-sexual-attraction hypothesis, which proposes that the factor ultimately responsible for the intensity of female sexual attractiveness is female intragroup scramble competition. We discuss an alternative hypothesis—the social passport hypothesis—which is compatible with the results of the study.

We construct nonblocking networks that are efficient not only as regards their cost and delay, but also as regards the time and space required to control them. In this paper we present the first simultaneous “weakly optimal” solutions for the explicit construction of nonblocking networks, the design of algorithms and data-structures. “Weakly optimal” is in the sense that all measures of complexity (size and depth of the network, time for the algorithm, space for the data-structure, and number of processor-time product) are within one or more logarithmic factors of their smallest possible values. In fact, we construct a scheme in which networks withn inputs andn outputs have sizeO(n(logn)2) and depthO(logn), and we present deterministic and randomized on-line parallel algorithms to establish and abolish routes dynamically in these networks. In particular, the deterministic algorithm usesO((logn)5) steps to process any number of transactions in parallel (with one processor per transaction), maintaining a data structure that useO(n(logn)2) words.

In the last decade, we have witnessed the dramatic development of the smart home industry. Smart home systems are currently facing an explosive growth of data. Making good use of this vast amount of data has become an attractive research topic in recent years. In order to develop smart home systems’ abilities for learning users’ behaviors autonomously and offering services spontaneously, a user behavior prediction model based on parallel back propagation neural network (BPNN) and k-nearest neighbor (KNN) algorithms is introduced in this paper. Based on MapReduce, a parallel BPNN algorithm is proposed to improve the prediction accuracy and speed, and a parallel KNN algorithm is developed for user decision-making rule selection. The experimental results indicate that the proposed model is significantly better than traditional user behavior prediction models in term of prediction accuracy and speed. A case study on smart home also illustrates the effectiveness of the proposed model.

Two quasi-static minimax open shortest path first (OSPF) routing algorithms in networks supporting the Switched Multi-megabit Data Service (SMDS) are presented and compared. In OSPF routing, the network is modeled as a graph and each link is associated with a nonnegative arc weight. A shortest path spanning tree is calculated for each origin to carry both the individually addressed and the group addressed (multicast) traffic. The OSPF routing protocol is adopted as a major part of the default inter-switching system interface (ISSI) routing algorithm for SMDS networks where arc weights are inversely proportional to the aggregate link set capacities. The problem of choosing a set of link set metrics is considered so that the maximum link utilization factor is minimized in an SMDS network. The problem is formulated as a nonlinear mixed integer programming problem.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

In hardware generated using CAOS, concurrent execution of maximal set of actions in each clock cycle reduces the latency (number of clock cycles) of the design at the cost of increase in its peak power, which is defined as the maximum instantaneous power (due to switching activity) during one clock cycle. In Chapter 5, various heuristics targeting the minimization of peak power in designs generated from CAOS are proposed. The peak power heuristics presented in that chapter postpone some actions (among all the actions that can be executed in a clock cycle) to the future clock cycles for peak power reduction. Also, actions which were postponed in the previous clock cycles are considered for execution in the present clock cycle. For this, those heuristics propose the use of extra state elements in order to remember which actions were postponed in the previous clock. The main drawback of those approaches is that such a use of extra state elements is associated with the corresponding area and power overheads.

Biomedical Engineering, usually known as Bioengineering, is among the fastest developing and one of most important interdisciplinary fields today. It connects natural and technical sciences, for all of which biological and medical phenomena, computation, and data management play important roles in science and industry. Distributed computing and parallel algorithms have proved to be effective in solving the problem with high computational complexity in a wide range of domains, including areas of computational bioengineering. This special edition collects high-quality research papers in the field of application of distributed computing systems in bioengineering applications. This issue is the achievement of the developments of modern computer and biomedical technology. Most bioengineering or bioinformatics methods need to access and analyses large amounts of data. For that it is necessary to achieve an effective way that from Big data creates knowledge and useful applications. The nature of such scientific data/information/knowledge demands the use of a powerful computing and intelligent database management systems. Areas of current and future research include advanced topics in this special issue. The paper from Juhasz et al. present a novel, GPU-based streaming architecture that has the potential for drastically reducing execution times and at the same time providing simultaneous 2D and 3D visualization facilities. The system uses a highly-optimised and re-configurable pipeline of CPU and GPU cores that attempts to exploit the tremendous computing power whenever possible. The system can process live data arriving from an EEG device or data stored in EEG data files. The computer drives a large display wall system of four 46 monitors that provides a 4K-resolution drawing surface for visualizing raw EEG data, potential maps and various 3D views of the patient's head. Authors from the group of M. Gusev presents the scalable balancer in Cloud computing extended to Dew Computing level. Given a successful implementation of scalable low level load balancer, implemented on the network layer. The scalability is proved with series of experiments. The experiments showed that it adds small latency of several milliseconds and thus it slightly reduces the performance when the distributed system is underutilized. Rybicki et al. demonstrate how to realize a Software-as-a-Service solution for a variety of science software using container technologies. The presented solution utilized DARIAH-DE research infrastructure based on OpenStack and UNICORE grid to deliver an extensible solution for the digital humanities domain. The paper from authors Forer et al. describe the extension of Cloudflow to support ApacheSpark without any adaptions to already implemented pipelines. The described performance evaluation demonstrates that Spark can bring an additional boost for analyzing next generation sequencing(NGS) data to the eld of genetics. The Cloudflow framework is open source and freely available at https://github.com/genepi/cloudflow . The paper by Memon et al. describes an approach for executing computational jobs across HPC and HTC resources while operating on geographically dispersed data via a global federated file system. The solution is realized as a new framework that integrates UNICORE and GFFS to provide a standards-based environment to support large scale data intensive computations frequent in today's biomedical analyses. The paper Parallel algorithm for the state space exploration, authored by L. Allal et al., proposes a new automatic verification technique based on model checking that determines whether a given system satisfies its specification. Such a technique suffers from the state explosion problem when traversing all possible states of systems. A new synchronized parallel algorithm (SPA) of exploration is proposed based on a fixed number of threads. Exhaustive comparative studies between the standard parallel exploration algorithm in SPIN and the new SPA show that the SPA performs slightly better regarding the execution time and memory complexity. Smart systems in telemedicine frequently use intelligent sensor devices at large scale. Practitioners can monitor non-stop the vital parameters of hundreds of patients in real-time. The most important pillars of remote patient monitoring services are communication and data processing. Large scale data processing is done mainly using workflows. In the paper Eszter et al. give a brief overview of the different check pointing techniques and propose two new provenance based check pointing algorithms which uses the information stored in the workflow structure to dynamically change the frequency of check pointing and can be efficiently used for dynamic health care smart systems.

Sorting appears the most attention among all computational tasks over the past years because sorted data is at the heart of many computations. Sorting is of additional importance to parallel computing because of its close relation to the task of routing data among processes, which is an essential part of many parallel algorithms. Many parallel sorting algorithms have been investigated for a variety of parallel computer architectures. In this paper, three parallel sorting algorithms have been implemented and compared in terms of their overall execution time. The algorithms implemented are the odd-even transposition sort, parallel merge sort and parallel shell sort. Cluster of Workstations or Windows Compute Cluster has been used to compare the algorithms implemented. The C# programming language is used to develop the sorting algorithms. The MPI library has been selected to establish the communication and synchronization between processors. The time complexity for each parallel sorting algorithm will also be mentioned and analyzed.

&#x0D; &#x0D; &#x0D; &#x0D; Apriori algorithm is a classical algorithm of association rule mining and widely used for generating frequent item sets. However, the original Apriori algorithm has some limitation such as it needs to scan the dataset many times to discover all frequent itemset and generate huge number of candidate itemset. To overcome these limitations, researchers have made a lot of improvements to the Apriori such as candidate generation, without candidate generation, transaction reduction, partitioning, and sampling. When it comes to mine massive data, these algorithms failed to prove efficiency because limitation of the processing capacity, storage capacity, and main memory constraints. Therefore, parallel and distributed algorithms are developed to perform large-scale computing in ARM on multiple processors. However, the problems with most of the parallel and distributed framework are overheads of managing distributed system, lack of high level parallel programming language, and node failures. Hadoop-MapReduce is an efficient, scalable, and simplified programming model for massive data processing and it also available on cloud environment. Cloud computing offers huge computing resources, and capacities to solve big data challenges. Recently many parallel algorithms have been proposed on Hadoop-MapReduce to enhance the performance of Apriori algorithm but there are some drawbacks: since multiple scan over the dataset is needed to generate candidate itemset, it consume more execution time. The aim of this study is to propose a parallel Transaction Reduction MapReduce Apriori algorithm (TRMR-Apriori) which is reduce unnecessary transaction values and transactions from the dataset in parallel manner to overcome above problems. The experiments show that TRMR-Apriori is able to achieve better execution time to discover frequent itemset those of previous sequential ARM algorithms such as Apriori, AprioriTid, Eclat, and FP-Growth and the previous parallel algorithms such as PApriori, MRApriori, and Modified Apriori with different condition on homogeneous computing environment using Hadoop-MapReduce platform in cloud. Overall, the TRMR-Apriori shows the strength to extract the frequent itemset from massive dataset in cloud.&#x0D; &#x0D; &#x0D; &#x0D; &#x0D;

Data partitioning algorithms aiming to minimize the execution time and the energy of computations in self-adaptable data-parallel applications on modern extreme-scale multicore platforms must address two critical challenges. First, they must take into account the new complexities inherent in these platforms such as severe resource contention and non-uniform memory access. Second, they must have low practical runtime and memory costs. The sequential data partitioning algorithms addressing the first challenge have a theoretical time complexity of O( $m * m * p * p$ ) where $m$ is the number of points in the discrete speed/energy function and $p$ is the number of available processors. They, however, exhibit high practical runtime cost and excessive memory footprint, therefore, rendering them impracticable for employment in self-adaptable applications executing on extreme-scale multicore platforms. We present, in this paper, the parallel data partitioning algorithms that address both the challenges. They take as input the functional models of performance and energy consumption against problem size and output workload distributions, which are globally optimal solutions. They have a low time complexity of O( $m * m * p$ ) thereby providing a linear speedup of O( $p$ ) and low memory complexity of O( $n$ ) where $n$ is the workload size expressed as a multiple of granularity. They employ dynamic programming approach, which also facilitates the easier integration of performance and energy models of communications. We experimentally study the practical cost of application of our algorithms in two data-parallel applications, matrix multiplication and fast Fourier transform, on a cluster in Grid’5000 platform. We demonstrate that their practical runtime and memory costs are low making them ideal for employment in self-adaptable applications. We also show that the parallel algorithms exhibit tremendous speedups over the sequential algorithms. Finally, using theoretical analysis for a forecast exascale platform, we demonstrate that the parallel algorithms have negligible execution times compared to the matrix multiplication application executing on the platform.

A generalized recursive circulant network(GR) is widely used in the design and implementation of local area networks and parallel processing architectures. In this paper, we investigate the routing of a message on this network, that is a key to the performance of this network. We would like to transmit maximum number of packets from a source node to a destination node simultaneously along paths on this network, where the i th packet traverses along the i th path. In order for all packets to arrive at the destination node securely, the i th path must be node-disjoint from all other paths. For construction of these paths, employing the Hamiltonian Circuit Latin Square(HCLS), a special class of (n x n) matrices, we present O(n 2 ) parallel routing algorithm on generalized recursive circulant networks.

Most reconfigurable computing architectures suffer from computationally demanding Placement and Routing (P&R) steps which might hamper their use in contexts requiring dynamic compilation (e.g., to guarantee application portability in embedded systems). Bearing in mind the simplification of P&R steps, this paper presents and analyzes a coarse-grained reconfigurable array extended with global Omega Networks. We show that integrating one or two Omega Networks in a coarse-grained array simplifies the P&R stage with both low hardware resource overhead and low performance degradation (18% for an 8×8 array). The experimental results included permit to compare the coarse-grained array with one or two Omega Networks with a coarse-grained array based on a grid of processing elements with neighbor connections. When comparing the execution time to perform the P&R stage needed for the two arrays, we show that the array using two Omega Networks needs a far simple P&R which for the benchmarks used completed on average in about 20× less time.

Energy balance is an important performance index in the design of routing algorithm for wireless sensor networks. A swarm intelligence optimization based routing algorithm for wireless sensor networks is proposed, whose kernel idea is taking less hop numbers into consideration and choosing the nodes with less pheromone as next hop to avoid some nodespsila prematurely exhausting their energy because of too concentrated routes through the nodes. The experiments show that the algorithm proposed in this paper is better than the directed diffusion routing protocol both in end-to-end delay and global energy balance and can effectively balance the global energy consumption and prolong the network lifetime.