parallel converters have been widely applied in distributed generations. The use of LCL filter in parallel converters may induce several stability challenges due to the filter resonance. To eliminate the filter resona...
详细信息
ISBN:
(纸本)9781728124551
parallel converters have been widely applied in distributed generations. The use of LCL filter in parallel converters may induce several stability challenges due to the filter resonance. To eliminate the filter resonance related issues, this paper presents a system level design that allows the use of small L filter without sacrificing system current quality. The key of the design is to identify the required number of converters in parallel. The L-filter based design is applicable in systems with either common or separate DC links. In this paper, parallel two-level converters with separate DC links are considered in the design, showing that the required number of converters is irrelevant to system parameters like DC voltage or power rating. Multilevel converters are also included in the study. Furthermore, the behaviour of dominant harmonics in different load condition is also investigated, using common DC link type parallel converters as an example. Simulation results and preliminary experimental results are obtained and presented.
Given a set of functional dependencies Σ and a single dependency σ, we show that the algorithm to test whether Σ implies σ is log-space complete in P. The functional dependencies Σ are represented as a directed h...
详细信息
ISBN:
(纸本)0818620528
Given a set of functional dependencies Σ and a single dependency σ, we show that the algorithm to test whether Σ implies σ is log-space complete in P. The functional dependencies Σ are represented as a directed hypergraph HΣ [1]. We first present a parallel algorithm which solves the above implication problem using P processors on a EREW-PRAM in Ο(e/P + ***) time and on a CRCW-PRAM in Ο(e/P + n) time, where e and n are the number of arcs and nodes of the graph HΣ. For graphs HΣ with fixed degree and diameter, we show that the closure HΣ+ can be computed in NC. We present NC algorithms to obtain a non-redundant and a LR-Minimum cover for the set of functional dependencies Σ. All our algorithms on a n-node directed hypergraph with fixed degree and diameter can be implemented to run in Ο(log2n) time with M(n) processors on a CREW-PRAM model, where M(n) is the cost of multiplying two binary matrices. The algorithms are efficient based on the transitive closure bottleneck phenomenon [7] that is, any improvement in the time and processor complexity of the transitive closure algorithm will result in an improvement by the same amount for the algorithms presented here.
Index is an important element in databases, and the existence of index is unavoidable. When an index has been built on a particular attribute, database operations (e.g. selection, join) on this attribute will become m...
详细信息
ISBN:
(纸本)0769515797
Index is an important element in databases, and the existence of index is unavoidable. When an index has been built on a particular attribute, database operations (e.g. selection, join) on this attribute will become more efficient by utilizing the index. In this paper we focus on parallel algorithms for selection queries involving index that is data searching on indexed attributes. In this paper, we propose two categories of parallel selection queries using index: parallel exact match and range selections;depending on the type of selection conditions. As parallel algorithms for these selection queries are very much influenced by indexing schemes, we will also describe various index partitioning methods for paralleldatabases, and discusses their efficiency in supporting parallel selection query processing.
As the number of cores grow in HPC systems, so does the effect of system noise on applications running on these systems. With the knowledge that future large-scale parallel computer systems, including exascale systems...
详细信息
ISBN:
(纸本)9781509036820
As the number of cores grow in HPC systems, so does the effect of system noise on applications running on these systems. With the knowledge that future large-scale parallel computer systems, including exascale systems, will operate under an overall power bound, we claim to have found a solution that can counter the effects of noise. We present two methods that estimate the effects of noise on an application and then optimally redistributes power among nodes, such that the effects of noise are "hidden".
System logs are a valuable source of information for the analysis and understanding of systems behavior for the purpose of improving their performance. Such logs contain various types of information, including sensiti...
详细信息
ISBN:
(纸本)9781538653302
System logs are a valuable source of information for the analysis and understanding of systems behavior for the purpose of improving their performance. Such logs contain various types of information, including sensitive information. Information deemed sensitive can either directly be extracted from system log entries by correlation of several log entries, or can be inferred from the combination of the (non-sensitive) information contained within system logs with other logs and/or additional datasets. The analysis of system logs containing sensitive information compromises data privacy. Therefore, various anonymization techniques, such as generalization and suppression have been employed, over the years, by data and computing centers to protect the privacy of their users, their data, and the system as a whole. Privacy-preserving data resulting from anonymization via generalization and suppression may lead to significantly decreased data usefulness, thus, hindering the intended analysis for understanding the system behavior. Maintaining a balance between data usefulness and privacy preservation, therefore, remains an open and important challenge. Irreversible encoding of system logs using collision-resistant hashing algorithms, such as SHAKE-128, is a novel approach previously introduced by the authors to mitigate data privacy concerns. The present work describes a study of the applicability of the encoding approach from earlier work on the system logs of a production high performance computing system. Moreover, a metric is introduced to assess the data usefulness of the anonymized system logs to detect and identify the failures encountered in the system.
Computer systems tend to be heterogeneous parallel and distributed computing systems, which are characterized by having various types of computational units interconnected via networks for executing multiple parallel ...
详细信息
ISBN:
(纸本)9781467371483
Computer systems tend to be heterogeneous parallel and distributed computing systems, which are characterized by having various types of computational units interconnected via networks for executing multiple parallel jobs precedence-constrained tasks. Scheduling multiple jobs, which arrive at different instants, on such systems for fastest execution is a well-known NP-hard optimization problem. In order to achieve high-performance of systems, two important factors can be improved. One factor is the heterogeneity. Most algorithms use the upward rank value for ordering tasks and the earliest finish time for assigning processors. These two criteria can be improved to permit creating accurate and efficient schedules in heterogeneous distributed computing systems. Another factor is the fairness, existing algorithms are for static scheduling, and failed to make full use of the fairness in dynamic environments, such that obvious unfairness to longer-makespan jobs or shorter-makespan jobs can be caused. A dynamic multiple parallel jobs scheduling algorithm called F DMHSV (Fairness of Dynamic Multiple Heterogeneous Selection Value) is proposed to address the above problems to achieve high-performance of systems in this paper. Both example and extensive experimental evaluation demonstrate significant improvement of the F_DMHSV algorithm.
With the reduced cost of hardware due to technological advancements and efficient elastic networks, distributedsystems size is increasing to meet better availability requirements. Existing consensus-based transaction...
详细信息
ISBN:
(纸本)9781509036820
With the reduced cost of hardware due to technological advancements and efficient elastic networks, distributedsystems size is increasing to meet better availability requirements. Existing consensus-based transactional systems show limited scal-ability due to either their reliance on a single leader or increased complexity in ordering mechanism. In this position paper we outline state-of-the-art solutions, highlight their weaknesses, and propose a vision to enhance leaderless consensus protocols by using reliable fast deliveries.
In this paper we present an efficient algorithm for compile-time scheduling and clustering of parallel programs onto parallel processing systems with distributed memory, which is called The Dynamic Critical Path Sched...
详细信息
In this paper we present an efficient algorithm for compile-time scheduling and clustering of parallel programs onto parallel processing systems with distributed memory, which is called The Dynamic Critical Path Scheduling DCPS. The DCPS is superior to several other algorithms from the literature in terms of computational complexity, processors consumption and solution quality. DCPS has a time complexity of O(e + v logv), as opposed to DSC algorithm O((e+v) logv) which is the best known algorithm. Experimental results demonstrate the superiority of DCPS over the DSC algorithm.
Processor scheduling in distributed-memory systems has received considerable attention in recent years. Several commercial distributed-memory systems use space-sharing processor scheduling. in space-sharing, the set o...
详细信息
ISBN:
(纸本)0818684038
Processor scheduling in distributed-memory systems has received considerable attention in recent years. Several commercial distributed-memory systems use space-sharing processor scheduling. in space-sharing, the set of processors in a system is partitioned and each partition is assigned for the exclusive use of a job. Space-sharing policies can be divided into fixed, static, or dynamic categories. For distributed-memory systems, dynamic policies incur high overhead. Thus, static policies are considered as these policies provide a better performance than the fixed policies. Several static policies have been proposed in the literature. In a previously proposed adaptive static policy, the partition size is a function of the number of queued jobs. This policy, however, tends to underutilize the system resources. To improve the performance of this policy, we propose a new policy in which the partition size is a function of the total number of jobs in the system, as opposed to only the queued jobs. The results presented here demonstrate that the new policy performs substantially better than the original policy for the various workload and system parameters. Another major contribution is the evaluation of the performance sensitivity to job structure, variances in inter-arrival times and job service times, and network topology.
Vector prefix and reduction are collective communication primitives in which all processors must cooperate. We present two parallel algorithms, the direct algorithm and the split algorithm, for vector prefix and reduc...
详细信息
ISBN:
(纸本)0818684038
Vector prefix and reduction are collective communication primitives in which all processors must cooperate. We present two parallel algorithms, the direct algorithm and the split algorithm, for vector prefix and reduction computation on coarse-grained, distributed-memory parallel machines. Our algorithms are relatively architecture independent and can be used effectively in many applications such as Pack/Unpack, Array Prefix/Reduction Functions, and Array Combining Scatter Functions, which are defined in Fortran 90 and in High Performance Fortran. Experimental results on the CM-5 are presented.
暂无评论