A strategy for scheduling tightly coupled parallel applications on clusters

Although various strategies have been developed for scheduling parallel applications with independent tasks, very little work exists for scheduling tightly coupled parallel applications on cluster environments. In this paper, we compare four different strategies based on performance models of tightly coupled parallel applications for scheduling the applications on clusters. In addition to algorithms based on existing popular optimization techniques, we also propose a new algorithm called Box Elimination that searches the space of performance model parameters to determine the best schedule of machines. By means of real and simulation experiments, we evaluated the algorithms on single cluster and multi-cluster setups. We show that our Box Elimination algorithm generates up to 80% more efficient schedule than other algorithms. We also show that the execution times of the schedules produced by our algorithm are more robust against the performance modeling errors.

Inverse surface melting in confined clusters: Ar-13 in zeolite L

Monte Carlo and molecular dynamics simulations on an Ar-13 cluster in zeolite L have been carried out at a series of temperatures to understand the rigid-nonrigid transition corresponding to the solid-liquid transition exhibited by the free Ar-13 cluster. The icosahedral geometry of the free cluster is no longer preferred when the cluster is confined in the zeolite. The root-mean-squared pair distance fluctuation, delta, exhibits a sharp, well-defined rigid-nonrigid transition at 17 K as compared to 27 K for the free cluster. Multiple peaks in the distribution of short-time averages of the guest-host interaction energy indicate coexistence of two phases.; It is shown that this transition is associated with the inner atoms becoming mobile at 17 K even while the outer layer atoms, which are in close proximity to the zeolitic wall, continue to be comparatively immobile. This may be contrasted with the melting of large free clusters of 40 or more atoms which exhibit surface melting. Guest-host interactions seem to play a predominant role in determining the properties of confined clusters. We demonstrate that the volume of the cluster increases rather sharply at 17 and 27 K respectively for the confined and the free cluster. Power spectra suggest that the motion of the inner atoms is generally parallel to the atoms which form the cage wall.

Monte Carlo and molecular dynamics simulation of argon clusters andn-alkanes in the confined regions of zeolites

Geometry and energy of argon clusters confined in zeolite NaCaA are compared with those of free clusters. Results indicate the possible existence of magic numbers among the confined clusters. Spectra obtained from instantaneous normal mode analysis of free and confined clusters give a larger percentage of imaginary frequencies for the latter indicating that the confined cluster atoms populate the saddle points of the potential energy surface significantly. The variation of the percentage of imaginary frequencies with temperature during melting is akin to the variation of other properties. It is shown that confined clusters might exhibit inverse surface melting, unlike medium-to-large-sized free clusters that exhibit surface melting. Configurational-bias Monte Carte (CBMC) simulations of n-alkanes in zeolites Y and A are reported. CBMC method gives reliable estimates of the properties relating to the conformation of molecules. Changes in the conformational properties of n-butane and other longer n-alkanes such as n-hexane and n-heptane when they are confined in different zeolites are presented. The changes in the conformational properties of n-butane and n-hexane with temperature and concentration is discussed. In general, in zeolite Y as well as A, there is significant enhancement of the gauche population as compared to the pure unconfined fluid.

Exploiting Programmable Network Interfaces for Parallel Query Execution in Workstation Clusters

Workstation clusters equipped with high performance interconnect having programmable network processors facilitate interesting opportunities to enhance the performance of parallel application run on them. In this paper, we propose schemes where certain application level processing in parallel database query execution is performed on the network processor. We evaluate the performance of TPC-H queries executing on a high end cluster where all tuple processing is done on the host processor, using a timed Petri net model, and find that tuple processing costs on the host processor dominate the execution time. These results are validated using a small cluster. We therefore propose 4 schemes where certain tuple processing activity is offloaded to the network processor. The first 2 schemes offload the tuple splitting activity - computation to identify the node on which to process the tuples, resulting in an execution time speedup of 1.09 relative to the base scheme, but with I/O bus becoming the bottleneck resource. In the 3rd scheme in addition to offloading tuple processing activity, the disk and network interface are combined to avoid the I/O bus bottleneck, which results in speedups up to 1.16, but with high host processor utilization. Our 4th scheme where the network processor also performs apart of join operation along with the host processor, gives a speedup of 1.47 along with balanced system resource utilizations. Further we observe that the proposed schemes perform equally well even in a scaled architecture i.e., when the number of processors is increased from 2 to 64