Load imbalance in parallel systems can be generated by external factors to the currently running applications like operating system noise or the underlying hardware like a heterogeneous cluster. HPC applications worki...
详细信息
Load imbalance in parallel systems can be generated by external factors to the currently running applications like operating system noise or the underlying hardware like a heterogeneous cluster. HPC applications working on irregular data structures can also have difficulties to balance their computations across the parallel tasks. In this article we extend, improve and evaluate more deeply the Task Packing mechanism proposed in a previous work. The main idea of the mechanism is to concentrate the idle cycles of unbalanced applications in such a way that one or more CPUs are freed from execution. To achieve this, CPUs are stressed with just useful work of the parallel application tasks, provided performance is not degraded. The packing is solved by an algorithm based on the knapsack problem, in a minimum number of CPUs and using oversubscription. We design and implement a more efficient version of such mechanism. To that end, we perform the Task Packing "in place", taking advantage of idle cycles generated at synchronization points of unbalanced applications. Evaluations are carried out on a heterogeneous platform using FT and miniFE benchmarks. Results showed that our proposal generates low overhead. In addition the amount of freed CPUs are related to a load imbalance metric which can be used as a prediction for it. (C) 2019 Elsevier Inc. All rights reserved.
In the last decades a massive development in data communication and networking was occurred. Media like image, audio and video have been used frequently especially with the invention of games and voice communications....
详细信息
ISBN:
(纸本)9781538691885
In the last decades a massive development in data communication and networking was occurred. Media like image, audio and video have been used frequently especially with the invention of games and voice communications. Steganography is one of the challenging fields of securing transmitted data. In this paper a new method is proposed to secure confidential messages by encrypting and hiding these messages in digital voices without any corruption on the covered media. The proposed method has a large payload, SNR,PSNR and very small BER and MSR.
The paper presents the results concerning computational burden analysis of dynamic programming and bottom-up algorithms when solving knapsack problems. It presents the efficiency of the algorithms information expresse...
详细信息
ISBN:
(纸本)9789897583049
The paper presents the results concerning computational burden analysis of dynamic programming and bottom-up algorithms when solving knapsack problems. It presents the efficiency of the algorithms information expressed both in calculation time, as well as mean number of iterations in knapsack problems up to 15,000 items and capacity of the knapsack equal to 10,000. The aim of the paper is to present the this knowledge what of practical use when solving optimization problems where estimate of execution time of the algorithm is important.
This research discusses the cutting problem encountered by a real-life window frame manufacturer. In the problem, four types of bars (upper, bottom, left and right) should be cut from raw material aluminium profiles f...
详细信息
This research discusses the cutting problem encountered by a real-life window frame manufacturer. In the problem, four types of bars (upper, bottom, left and right) should be cut from raw material aluminium profiles for each window frame order. These bars must be cut such that trim loss is minimised. Moreover, the bars should be assigned to the same raw material profile if possible to increase productivity;otherwise, they should be assigned to neighbouring raw material profiles. Furthermore, the numbers of bar types as derived from a raw material profile should not be unbalanced because this scenario induces subsequent machine load imbalance. In this study, we develop a mixed integer programming model and a knapsack-based heuristic approach that minimises the weighted sum of trim loss, bar type imbalance and the degree of order spreading. The results of computational experiments demonstrate the effectiveness of the proposed algorithm, and the proposed approach outperforms the legacy system of the company. Thus, this method is currently being used by the firm in question.
This study examines a two-stage two-dimensional cutting stock problem encountered by a paper mill company. The problem includes various machine-related and operational constraints based on real-world situations. Paper...
详细信息
This study examines a two-stage two-dimensional cutting stock problem encountered by a paper mill company. The problem includes various machine-related and operational constraints based on real-world situations. Paper products are manufactured using two major cutting processes. Each cutting machine has a specific minimum and maximum width for input and output rolls and is limited by the maximum number of rolls it can cut at the same time. A mathematical model is presented to formally address the problem and an efficient multiple-choice knapsack-based heuristic algorithm is proposed to solve the problem. To demonstrate the efficiency of the proposed heuristic algorithm, computational experiments are conducted on test data-set generated from real-world data provided by a large paper mill company in the Republic of Korea.
State of the art technologies like air-to-water heat pumps have become popular because of their high efficiencies and their supply by electricity rather than by gas or oil. As a direct result of this supply substituti...
详细信息
State of the art technologies like air-to-water heat pumps have become popular because of their high efficiencies and their supply by electricity rather than by gas or oil. As a direct result of this supply substitution and due to high load-shifting potential, electricity demand of heat pumps have to be taken into account into energy management strategies of the overall building. In this paper a new controller is discussed, which has predictive and adaptive properties and is able to optimize the running-times of non-modulating heat pumps. (C) 2015 The Authors. Published by Elsevier Ltd.
Cloud computing is a "pay per use" model, where the user or clients pay for the computational resources they use. Furthermore, in cloud failures are normal. Therefore cost is an important factor to be consid...
详细信息
ISBN:
(纸本)9783319120126;9783319120119
Cloud computing is a "pay per use" model, where the user or clients pay for the computational resources they use. Furthermore, in cloud failures are normal. Therefore cost is an important factor to be considered along with availability, performance and reliability. Also it is not necessary that the benefits accrued from the replication will be greater than the cost incurred. Thus, this paper proposes an algorithm named Dynamic Cost-aware Rereplication and Re-balancing Strategy (DCR2S). This algorithm optimizes the cost of replication using the concept of knapsack problem. The proposed algorithm is evaluated using CloudSim. Experimental results demonstrate the effectiveness of proposed algorithm.
In LTE systems, multicast services must be delivered efficiently in response to the need for strong QoS support. However, each class of quality services has its own requirements to be satisfied. These quality constrai...
详细信息
In LTE systems, multicast services must be delivered efficiently in response to the need for strong QoS support. However, each class of quality services has its own requirements to be satisfied. These quality constraints limit the scheduling flexibility, and the LTE downlink resource allocating algorithms need to assimilate these constraints while trying to maximize system performance in terms of fairness and throughput. This paper addresses this fundamental problem of LTE downlink scheduling by adopting the time-domain knapsack algorithm over the traffic overload patterns. We fine tune the knapsack algorithm, to overcome this problem and improve system performance objectives. We demonstrate that more efficient performance can be achieved in terms of fairness index and system throughput, which is evaluated using simulation results. (C) 2015 The Authors. Published by Elsevier B.V.
In LTE multi-service communication system, a trade-off between QoS assurance and fairness is a challenging issue, since the QoS provisioning at the cost of starving users in low service demand classes is not favorable...
详细信息
ISBN:
(纸本)9781479989287
In LTE multi-service communication system, a trade-off between QoS assurance and fairness is a challenging issue, since the QoS provisioning at the cost of starving users in low service demand classes is not favorable for the operator. In this paper, we adopt the time-domain knapsack algorithm and fine tune it to provide fair resource allocation while support QoS requirements in LTE downlink scheduling system when the bearers are from different classes of service, having different QoS characteristics. We demonstrate that more efficient performance can be achieved in two aspects of fairness and QoS provisioning in terms of normalized throughput, and packet loss and delay rate, which are evaluated using simulation results.
In LTE systems, multicast services must be delivered efficiently in response to the need for strong QoS support. However, each class of quality services has its own requirements to be satisfied. These quality constrai...
详细信息
In LTE systems, multicast services must be delivered efficiently in response to the need for strong QoS support. However, each class of quality services has its own requirements to be satisfied. These quality constraints limit the scheduling flexibility, and the LTE downlink resource allocating algorithms need to assimilate these con- straints while trying to maximize system performance in terms of fairness and throughput. This paper addresses this fundamental problem of LTE downlink scheduling by adopting the time-domain knapsack algorithm over the traffic overload patterns. We fine tune the knapsack algorithm, to overcome this problem and improve system performance objectives. We demonstrate that more efficient performance can be achieved in terms of fairness index and system throughput, which is evaluated using simulation results.
暂无评论