Given the explosive growth of biological sequence databases and the computational complexity of aligning large sequences over extremely large databases most researchers have opted for utilizing the BLAST algorithm. Wh...
详细信息
Large-scale parallel and distributedcomputing environments have several problems concerning power consumption, thermal ascent, and setting space. In particular, power consumption and thermal ascent are the two seriou...
详细信息
ISBN:
(纸本)9780889867741
Large-scale parallel and distributedcomputing environments have several problems concerning power consumption, thermal ascent, and setting space. In particular, power consumption and thermal ascent are the two serious problems that have to be taken care of when a system is used for a long time. We had proposed a CPU power control and scheduling technique (PCST) for realtime parallel and distributedcomputingsystems in order to solve the abovementioned problems;in these systems, the processors can change the frequencies and voltages. In this study, we evaluate the PCST by using the tracking program. the evaluation results show that when the PCST is employed, the power consumption is 60%-80% when compared to that of the commercial power control;moreover, when compared to that of the no power control, the power consumption is only 20%-40%.
Mutual exclusion is a fundamental problem in distributed processing systems. A generalization of mutual exclusion called k-exclusion for shared memory systems was introduced by Fischer et al. in [8] and subsequently s...
详细信息
Mutual exclusion is a fundamental problem in distributed processing systems. A generalization of mutual exclusion called k-exclusion for shared memory systems was introduced by Fischer et al. in [8] and subsequently studied in [1, 2, 3, 5]. In this paper, we present a simple solution to this problem and prove its correctness. Our solution is efficient both in time and space.
In linear algebra, Cholesky factorization is useful in solving a system of equations with a symmetric positive definite coefficient matrix. Cholesky factorization is roughly twice as fast relative to LU factorization ...
详细信息
ISBN:
(纸本)9781467345651;9780769549033
In linear algebra, Cholesky factorization is useful in solving a system of equations with a symmetric positive definite coefficient matrix. Cholesky factorization is roughly twice as fast relative to LU factorization which applies to general matrices. In recent years, with advances in technology, a Fermi GPU card can accommodate hundreds of cores compared to the small number of 8 or 16 cores on CPU. therefore a trend is seen to use the graphics card as a general purpose graphics processing unit (GPGPU) for parallel computation. In this work, Volkov's hybrid implementation of Cholesky factorization is evaluated on the new Fermi GPU with others and then some improvement strategies were proposed. After experiments, compared to the CPU version using Intel Math Kernel Library (MKL), our proposed GPU improvement strategy can achieve a speedup of 3.85x on Cholesky factorization of a square matrix of dimension 10,000.
We study the performance of an adaptive space sharing policy for parallelsystems when the application involves I/O operations. We consider four types of I/O models in this paper: NO, NOP, O1 and O2 models. In the NO ...
详细信息
the cascade of two baseline networks in tandem is a rearrangeable network. the cascade of two omega networks through a certain interconnection pattern is also rearrangeable. these belong to the general problem: for wh...
详细信息
Path planning is one of the most computation expensive tasks in the field of mobile robotics especially in dynamically changing environments. It is difficult to meet realtime requirements with serial path planning alg...
详细信息
ISBN:
(纸本)9780889867741
Path planning is one of the most computation expensive tasks in the field of mobile robotics especially in dynamically changing environments. It is difficult to meet realtime requirements with serial path planning algorithms. this would require a high speed processor. Particularly in small autonomous robot systems, this is inefficient due to the energy consumption and space requirement. In contrast, we propose a parallel path planning approach based on Marching Pixels which is a new innovative Organic computing principle. It can be used as coarse global path planner in dynamically changing environments because the algorithm is very fast and only requires few resources.
作者：
Derbel, B.Mosbah, M.LaBRI
Université Bordeaux I ENSEIRB 351 Cours de la libration 33405 Talence France
We present a linear time distributed algorithm for decomposing a graph into a disjoint set of clusters. this algorithm is truly parallel since many clusters can be constructed in parallel, which gives an answer to a q...
详细信息
We present a linear time distributed algorithm for decomposing a graph into a disjoint set of clusters. this algorithm is truly parallel since many clusters can be constructed in parallel, which gives an answer to a question asked by S. Moran and S. Snir in [1]. Moreover, no precomputed spanning tree is required for the computation of clusters. We apply the designed algorithm to construct covers for synchronizers γ1 and γ2.
In Machine learning (ML) the model we use is increasingly important, and the model's parameters, the key point of the ML, are adjusted through iteratively processing a training dataset until convergence. Although ...
详细信息
ISBN:
(纸本)9781509042975
In Machine learning (ML) the model we use is increasingly important, and the model's parameters, the key point of the ML, are adjusted through iteratively processing a training dataset until convergence. Although data-parallel ML systems often engage a perfect error tolerance when synchronizing the model parameters for maximizing parallelism, the synchronization of model parameters may delay in completion, a problem that generally gets worse at a large scale. this paper presents a Bounded Asynchronous parallel (BAP) model of computation that allows computations using stale model parameters in order to reduce synchronization overheads. In the meanwhile, our BAP model ensures theoretical convergence guarantees for large scale data-parallel ML applications. this model permits distributed workers to use the stale parameters storing in the local cache, instead of waiting until the Parameter Server (PS) produces a new version. this expressively reduces the time workers spend on waiting. Furthermore, the BAP model guarantees the convergence of ML algorithm by bounding the maximum distance of the stale parameters. Experiments conducted on 4 cluster nodes with up to 32 GPUs showed that our model significantly improved the proportion of computing time relative to the waiting time and led to 1.2-2 X speedup. Besides, we elaborated how to choose the staleness threshold when considering the tradeoff between Efficiency and Speed.
this paper proposes a method to convert a factor score matrix for the cooperation between multiple video retrieval systems based on impression. A matrix transforming the factor loading matrix is used. As a result, we ...
详细信息
ISBN:
(纸本)9781509055043
this paper proposes a method to convert a factor score matrix for the cooperation between multiple video retrieval systems based on impression. A matrix transforming the factor loading matrix is used. As a result, we showed that the results of two applications of the factor analysis can satisfactorily be transformed by using the transformation matrix.
暂无评论