synchronous and asynchronous algorithms are presented for distributed minimax optimization. The objective here is to realize the minimization of the maximum of component functions over the standard multi-agent network...
详细信息
synchronous and asynchronous algorithms are presented for distributed minimax optimization. The objective here is to realize the minimization of the maximum of component functions over the standard multi-agent network, where each node of the network knows its own function and it exchanges its decision variable with its neighbors. In fact, the proposed algorithms are standard consensus and gossip based subgradient methods, while the original minimax optimization is recast as minimization of the sum of component functions by using a p-norm approximation. A scalable step size depending on the approximation ratio p is also presented in order to avoid slow convergence. Numerical examples illustrate that the algorithms with this step size work well even in the high approximation ratios.
In this paper, we focus on applying parallel processing techniques to HEVC encoder in order to significantly reduce the computational power requirements without disturbing its coding efficiency. So, we propose several...
详细信息
In this paper, we focus on applying parallel processing techniques to HEVC encoder in order to significantly reduce the computational power requirements without disturbing its coding efficiency. So, we propose several, synchronous and asynchronous, parallelization approaches working at a coarse grain parallelization level, based on the Group Of Pictures (GOP), which we call GOP-based level. GOP -based approaches encode simultaneously several groups of consecutive frames. Depending on how these GOPs are conformed and distributed it is critical to obtain good parallel performance. The results show that near ideal efficiencies are obtained using up to 10 cores. Furthermore, when the computational load is unbalanced, the asynchronous versions outperform the synchronous ones. The parallel algorithms developed in this work support all standard coding modes proposed by the reference software. (C) 2016 Civil-Comp Ltd. and Elseider Ltd. All rights reserved.
We develop an efficient parallel distributed algorithm for matrix completion, named NOMAD (Non-locking, stOchastic Multi-machine algorithm for Asynchronous and Decentralized matrix completion). NOMAD is a decentralize...
详细信息
We develop an efficient parallel distributed algorithm for matrix completion, named NOMAD (Non-locking, stOchastic Multi-machine algorithm for Asynchronous and Decentralized matrix completion). NOMAD is a decentralized algorithm with non-blocking communication between processors. One of the key features of NOMAD is that the ownership of a variable is asynchronously transferred between processors in a decentralized fashion. As a consequence it is a lock-free parallel algorithm. In spite of being asynchronous, the variable updates of NOMAD are serializable, that is, there is an equivalent update ordering in a serial implementation. NOMAD outperforms synchronous algorithms which require explicit bulk synchronization after every iteration: our extensive empirical evaluation shows that not only does our algorithm perform well in distributed setting on commodity hardware, but also outperforms stateof-the-art algorithms on a HPC cluster both in multi-core and distributed memory settings.
A group membership protocol is presented and proven correct for a synchronous time-triggered model of computation with processors in a ring that broadcast in turn. The protocol, derived from one used for critical cont...
详细信息
ISBN:
(纸本)3540635750
A group membership protocol is presented and proven correct for a synchronous time-triggered model of computation with processors in a ring that broadcast in turn. The protocol, derived from one used for critical control functions in automobiles, accepts a very restrictive fault model to achieve low overhead and requires only one bit of membership information piggybacked on regular broadcasts. Given its strong fault model, the protocol guarantees that a faulty processor will be promptly diagnosed and removed from the agreed group of processors, and will also diagnose itself as faulty. The protocol is correct under a fault-arrival assumption that new faults arrive at least n + 1 time units apart, when there are n processors. Exploiting this assumption leads to unusual real-time reasoning in the correctness proof.
Hybrid algorithms incorporated with parallel processing techniques are very powerful tools for efficiently solving very complex optimization problems. We present asynchronous parallel computer architecture adaptation ...
详细信息
ISBN:
(纸本)9781479941735
Hybrid algorithms incorporated with parallel processing techniques are very powerful tools for efficiently solving very complex optimization problems. We present asynchronous parallel computer architecture adaptation based on hybridization of Genetic algorithms (GAs) and Estimation of Distribution algorithms (EDAs). In this master-slave formulation, slaves perform evolutionary computation independently using GAs, while master supervises and controls the searching process. Master's role is to probabilistically study the characteristics of solution space and directs the slaves on good searching spots. This study reports some few findings on the ability of our hybrid algorithm to solve some instances of BQP problem as well as AODV routing optimization in VANETs. For both problems our hybrid algorithm has obtained best results in terms of quality of solutions as well as computational speed.
暂无评论