In order to quickly take advantage of petascale machines, which are in the near future, we must improve the performance of current codes. In the past, general-purpose adaptive codes, such as Uintah, have been shown to scale to thousands of processors. However, in order to take advantage of petascale machines these codes will have to scale to hundreds of thousands of processors. One important as...

We analyze the equilibrium behavior of deterministic haploid mutation-selection models. To this end, both the forward and the time-reversed evolution processes are considered. The stationary state of the latter is called the ancestral distribution, which turns out as a key for the study of mutation-selection balance. We find that the ancestral genotype frequencies determine the sensitivity of t...

In block-partitioned parallel matrix factorization algorithms, where the matrix is distributed over a logical torus processor grid with an rs block-cyclic matrix distribution, the greatest scope for optimization exists in the formation of (block) panels. Let ! be the panel width, with ! m being an optimal value based on the characteristics a single processor's memory hierarchy. To date, two wel...

Preservation of locality of reference is the most critical issue for performance in high performance archi-tectures, especially scalable architectures with electronic interconnection networks. We brieey discuss some of the issues in this paper and give results from the use of spectral bisection on irregular grid computations on the Connection Machine systems.

The paper presents a novel mapping and routing technique for the mesh based NoC design problem with an objective of minimizing the energy consumption and normalized worst link-load. The proposed algorithm PLBMR is a particle swarm optimization (PSO) based two phases process, one is mapping core onto NoC to minimize the NoC communication energy consumption, and another is the allocation of routi...

CISUC , Department of Informatic Engineering, University of Coimbra Portugal {lorleans, pnf}@dei.uc.pt Abstract: Many of the load-balancing algorithms used in parallel systems do not have a concern about response times: tasks (or requests) are simply dispatched to a server, which provides no guarantees about their execution times. When there is a maximum acceptable response time (i.e. deadline)...

Distributed Systems are useful for computation and storage of large scale data at dispersed location. Distributed File System (DFS) is a subsystem of Distributed System. DFS is a means of sharing of storage space and data. Servers, Storage devices and Clients are on dispersed location in DFS. Fault tolerance and Scalability are two main features of distributed file system. Performance of DFS is...

Distributed Dynamic load balancing (DDLB) is an important system function destined to distribute workload among available processors to improve throughput and/or execution times of parallel computer in Cluster Computing. Instead of balancing the load in cluster by process migration, or by moving an entire process to a less loaded computer, we make an attempt to balance load by splitting process...

From a user’s perspective, VCR functions of variable-rate fast-forward and rewind are desirable features in many video applications. To implement VCR-like operations, the video server should be able to access non-contiguous segments of a video object without additional overhead. In this paper, we propose prime round-robin (PRR) placement scheme, which achieves uniform load-balance of disks for ...

thermal modeling of an automotive cabin was performed in this paper to predict the inside cabin air temperature. to implement this task, thermal and ventilation loads were estimated and the mass and energy balance conservation equations for dry air and water vapor with considering a new parameter (air circulation ratio) as well as the balance equations of internal components of a cabin were der...