作者：
Uno, AJAMSTEC
Earth Simulator Ctr Kanazawa Ku Yokohama Kanagawa 2360001 Japan
As many large-scale distributed parallel programs are run efficiently in the Earth Simulator (ES), I/O throughput between the tape library system and the work disk of the PNs became the bottleneck for simulations. So,...
详细信息
ISBN:
(纸本)9812563547
As many large-scale distributed parallel programs are run efficiently in the Earth Simulator (ES), I/O throughput between the tape library system and the work disk of the PNs became the bottleneck for simulations. So, in 2003, we introduced NQSII and Mass Data Processing System (MDPS) to address this problem and to improve system utilization and maintainability. In this paper, the new system (a scheduling algorithm and MDPS) and the recent program execution status for the ES are described.
Crosstalk effect is crucial for timing analysis in very deep submicron design. In this paper, we present and compare multiple scheduling algorithms to compute switching windows for static timing analysis in presence o...
详细信息
ISBN:
(纸本)0780364457;0780364465
Crosstalk effect is crucial for timing analysis in very deep submicron design. In this paper, we present and compare multiple scheduling algorithms to compute switching windows for static timing analysis in presence of crosstalk noise. We also introduce an efficient technique to evaluate the worst case alignment of multiple aggressors.
This paper describes a distributed optimization algorithm for scheduling the operation of electro-thermal heating units, such as heat pumps (HP) or electric heaters (EH) within a day-ahead (DA) operation for covering ...
详细信息
In this paper, multiuser waterfilling theory applied to subcarrier and power allocation in orthogonal frequency division multiplexing (OFDM) system is researched. Note that the traditional scheduling algorithm making ...
详细信息
In this paper we consider linear wireless networks with variable number of nodes. We first derive a basic cut set bound for transport capacity of this class of networks highlighting the functional dependence on both n...
详细信息
ISBN:
(纸本)9781424419968
In this paper we consider linear wireless networks with variable number of nodes. We first derive a basic cut set bound for transport capacity of this class of networks highlighting the functional dependence on both node density and network length. After, we investigate how transport throughput scales with respect to node density 6 in a wireless network with multihop communication. While in linear multihop networks without fading transport throughput does not increase by increasing node density, i.e. it scales as 60 with a gap with respect to capacity bounds, we show that a capacity increase up to log 6 is achieved in the presence of fading by using, instead of a simple nearest neighbor communication, suitable algorithms that adaptively schedule source destination links in the networks. Finally, we compare multihop schemes with schemes based on distributed MIMO communication by discussing the relevant differences and showing that, despite the latter strategy is able to asymptotically achieve scaling law bound approaching 6, multihop schemes appear better in terms of capacity in a significant range of values of 6 and when the network becomes extended.
Among many DAG scheduling algorithms targeting at maximizing DAG ready tasks to cope with temporal uncertainties that occurs during task execution, the PB heuristic seems promising. This paper presents the work of ada...
详细信息
ISBN:
(纸本)9781509036776
Among many DAG scheduling algorithms targeting at maximizing DAG ready tasks to cope with temporal uncertainties that occurs during task execution, the PB heuristic seems promising. This paper presents the work of adapting PB to a real distributed multi-core computing system with the PaRSEC scheduling runtime. The application performance results before and after adapting the PB heuristic to PaRSEC are evaluated and compared with different settings on application size and compute node configuration. We analyze the results and conclude with an assessment of the usability of the DAG ready task maximization algorithm on multi-core computing platforms.
In this paper, we propose a QoS-aware resource allocation scheme for IEEE 802.16d fixed WiMAX system operating in TDD mode. It focuses on the performance enhancement of the resource allocation. The proposed scheduling...
详细信息
ISBN:
(纸本)9780769534978
In this paper, we propose a QoS-aware resource allocation scheme for IEEE 802.16d fixed WiMAX system operating in TDD mode. It focuses on the performance enhancement of the resource allocation. The proposed scheduling algorithm flexibly adjusts bandwidth among the different traffics to try to get the best efficiency while considering the fairness among the different traffics in a way. The efficiency of the proposed QoS-aware resource allocation scheme is revealed through simulation.
In large-scale distributed systems, such as data centers resource failures are the norm rather than an exception. In this paper, we propose a holistic approach to study resource failures from resource failure modellin...
详细信息
ISBN:
(纸本)9781728125220
In large-scale distributed systems, such as data centers resource failures are the norm rather than an exception. In this paper, we propose a holistic approach to study resource failures from resource failure modelling to distributed system simulation to failure-aware scheduling algorithm design. In particular, we present (1) a simple and yet practical way to model resource failures using real-world failure traces, (2) a new distributed systems simulator and (3) two failure-aware scheduling algorithms. These scheduling algorithms are designed primarily to validate (1) and (2). Our evaluation results demonstrate the feasibility and effectiveness of our holistic approach.
This paper addresses the problem of scheduling parallel real-time tasks of Directed Acyclic Graph (DAG) model on multiprocessor systems. We propose a new scheduling method based on a subtask-level, which means that th...
详细信息
ISBN:
(纸本)9783319083117;9783319083100
This paper addresses the problem of scheduling parallel real-time tasks of Directed Acyclic Graph (DAG) model on multiprocessor systems. We propose a new scheduling method based on a subtask-level, which means that the schedulability decisions are taken based on the local temporal parameters of subtasks. This method requires modifying the subtasks to add more parameters which are necessary for the analysis, such as local offsets, deadlines and release jitters. Then we provide interference and workload analyses of DAG tasks, and we provide a schedulability test for any work conserving scheduling algorithm.
The work proposes a design for a new external scheduler for SLURM (Simple Linux Utility for Resource Management). Schedulers, included in SLURM by default are good enough for many sites, but big supercomputers serving...
详细信息
The work proposes a design for a new external scheduler for SLURM (Simple Linux Utility for Resource Management). Schedulers, included in SLURM by default are good enough for many sites, but big supercomputers serving many users meet limitations of standard SLURM schedulers. In this work we discover methods to break these limitations via implementing new portable SLURM scheduler. We address the problem of maximizing the number of users, whose requests are processed in each given moment of time, and decrease start time of user's first task. Our approach is based on standard backfill algorithm and includes additional features, such as simplification of SLURM priority system, replacing slow SQL-based accounting checks by faster ACL checks and upgrading cluster administration convenience.
暂无评论