Traditional cluster computing systems such as the supercomputers are equipped with specially designed high-performance hardware, which escalates the manufacturing cost and the energy cost of those systems. Due to such drawbacks and the diversified demand in computation, two new types of clusters are developed: the GPU clusters and the Hadoop clusters. The GPU cluster combines traditional CPU-only computing cluster with general purpose GPUs to accelerate the applications. Thanks to the massively-parallel architecture of the GPU, this type of system can deliver much higher performance-per-watt than the traditional computing clusters. The Hadoop cluster is another popular type of cluster computing system. It uses inexpensive off-the-shelf component and standard Ethernet to minimize manufacturing cost. The Hadoop systems are widely used throughout the industry. Alongside with the lowered cost, these new systems also bring their unique challenges. According to our study, the GPU clusters are prone to severe under-utilization due to the heterogeneous nature of its computation resources, and the Hadoop clusters are vulnerable to network congestion due to its limited network resources. In this research, we are trying to improve the throughput of these novel cluster computing systems by increasing the workload parallelism and network I/O parallelism.

Clusters of galaxies are particularly interesting astrophysical systems, are the largest bound structures in the Universe, and contain fair sample of cosmic ingredients. Studies of cluster abundance as a function of mass and redshift were critical in establishing the standard model of cosmology. This dissertation presents results from X-ray imaging of massive distant (M > 10^14 M; 0:3 < z < 1.2) clusters, found via X-ray emission or Sunyaev-Zeldovich eff ect. This is the world's largest sample of massive galaxy clusters. We explore the radial and azimuthal profi les of the X-ray emitting gas and show that clusters are self-similar objects: their internal structure is largely independent of the cluster's mass or redshift, and the fractions of di fferent types of clusters does not change with redshift. We also present a new statistical technique for measuring a cluster's deviations from a perfect axisymmetric shape, which is especially useful in the case of low photon count observations of distant clusters. / Physics

published_or_final_version / Physics / Master / Master of Philosophy

Transition metal clusters are a unique class of chemical substances. Not only do they have well-defined molecular structures, they also exhibit interesting and potentially useful properties that are inherent to metal-metal bonded species. They may be viewed as atom-like building blocks with adjustable properties. Detailed in this dissertation are the efforts to develop synthetic methodologies necessary to bring a cluster system out of the limited sphere of fundamental cluster chemistry and into general synthetic applicability. Specifically, the design and synthesis, structural characterization, and synthetic applications of the cluster complexes of the [Re₆(μ₃-Se)₈]²⁺ core will be discussed. Chapter 1 provides necessary background information of the [Re₆(μ₃-Se)₈]²⁺ core-containing cluster system, the rationales of cluster-supported synthetic and structural chemistry, and the impetus for these endeavors. Chapter 2 details the synthetic applications of stereospecific complexes of the [Re₆(μ₃-Se)₈]²⁺ clusters for the assembly of nanoscopic multicluster arrays using molecular and supramolecular approaches. The synthesis and structural characterization of a triangle-shaped tricluster array and a tetracluster assembly composed of two hydrogen-bonded diclusters are described. Chapters 3 and 4 describe the synthesis of the [Re₆(μ₃-Se)₈]²⁺ core-containing cluster complexes with the water-soluble 1,3,5-triaza-7-phosphaadamantane (PTA) ligand and the chemistry of these cluster-complex ligands for the coordination of a variety of secondary metal ions. In Chapter 5, the supramolecular chemistry of the cluster-PTA complexes with Ag(I) is detailed. The coordination of Ag(I) to a cluster complex with two PTA ligands disposed trans- to each other produces a 2-dimensional, porous solid with nano-sized hydrophobic pores that are potentially useful for storage of hydrocarbons and for occlusion of certain substrates for activation and possible catalysis. Chapter 6 summarizes the efforts to synthesize cluster-polymer hybrid materials by using an initiating cluster complex for the controlled radical polymerization. The synthesis, spectroscopic and structural characterizations of a novel cluster complex with an initiating ligand are described. Preliminary results of its application for the controlled polymerization of methyl methacrylate are detailed. Chapter 7 draws a set of conclusions based on the results presented in Chapters 2-6 and elaborates on some future directions aimed at moving one step forward the cluster-supported synthetic and materials chemistry.

Mass-selected cluster anions are employed as model micro-solutions to study solvent effects on the structural motifs and electronic structure of anionic solutes, including the roles of the solvent in controlling the outcomes of photochemical processes. Interaction of light with cluster anions can potentially lead to cluster photodissociation in addition to photodetachment. We investigate these competing processes by means of photoelectron imaging spectroscopy combined with tandem time-of-flight (TOF) mass spectrometry. Photoelectron images are reported for members of the [(CO2)n(H2O)m]- cluster series. For homogeneous solvation, the photodetachment bands show evidence of cluster core switching between a CO2- monomer anion and a covalent (CO2)2- dimer anionic core, confirming previous observations. The Photoelectron Angular Distributions (PADs) of the monomer- and dimer-based clusters reveal an interference effect that result in similar PADs. Stabilization of the metastable CO2- anion by water solvent molecules is highlighted because its ability to "trap" the excess electron on CO2. Most surprising is the effect of the water solvent in quenching the autodetachment channel in excited states normally embedded in the electron detachment continuum, allowing excited CO2-(H2O)m clusters to follow reaction paths that lead to cluster fragmentation. Observed O- based photoproducts are attributed to photodissociation of the CO2- cluster core and are dominant for small parent clusters, whereas a water evaporation channel dominates for larger clusters. Addition of a second CO2 to these clusters is shown to preferentially form monomer based clusters, whose photodissociation exhibit an additional CO3- based channel, characteristic of a photoinitiated intracluster ion-molecule reaction between nascent O- and the additional CO2 solvent molecule. Changes in the PADs of NO- are monitored as a function of electron kinetic energy for the NO-(N2O)n and NO-(H2O)n cluster anions. In contrast with hydration, angular distributions become progressively more isotropic for the N2O case, particularly when the photoelectron kinetic energies are in the vicinity of the 2Pi shape resonance of the N2O solvent molecules. First time observation of the CH3SOCH- anion of dimethylsulfoxide is reported along with the photoelectron images of this organic anion and of the monohydrated cluster. Observed photodissociation products are HCSO- and SO-.

Single nucleotide polymorphisms (SNPs) have been increasingly popular for a wide range of genetic studies. A high-throughput genotyping technologies usually involves a statistical genotype calling algorithm. Most calling algorithms in the literature, using methods such as k-means and mixturemodels, rely on elliptical structures of the genotyping data; they may fail when the minor allele homozygous cluster is small or absent, or when the data have extreme tails or linear patterns. We propose an automatic genotype calling algorithm by further developing a linear grouping algorithm (Van Aelst et al., 2006). The proposed algorithm clusters unnormalized data points around lines as against around centroids. In addition, we associate a quality value, silhouette width, with each DNA sample and a whole plate as well. This algorithm shows promise for genotyping data generated from TaqMan technology (Applied Biosystems). A key feature of the proposed algorithm is that it applies to unnormalized fluorescent signals when the TaqMan SNP assay is used. The algorithm could also be potentially adapted to other fluorescence-based SNP genotyping technologies such as Invader Assay. Motivated by the SNP genotyping problem, we propose a partial likelihood approach to linear clustering which explores potential linear clusters in a data set. Instead of fully modelling the data, we assume only the signed orthogonal distance from each data point to a hyperplane is normally distributed. Its relationships with several existing clustering methods are discussed. Some existing methods to determine the number of components in a data set are adapted to this linear clustering setting. Several simulated and real data sets are analyzed for comparison and illustration purpose. We also investigate some asymptotic properties of the partial likelihood approach. A Bayesian version of this methodology is helpful if some clusters are sparse but there is strong prior information about their approximate locations or properties. We propose a Bayesian hierarchical approach which is particularly appropriate for identifying sparse linear clusters. We show that the sparse cluster in SNP genotyping datasets can be successfully identified after a careful specification of the prior distributions.

Clusters and Clusterization are getting one of the most important fields to stimulate competition. In the Master study there are formulated problems of the economical development, analyzed and systemized research of the influence of practical Clusterization and clusters to the competition from various authors of Lithuania and other countries, also there are presented methods of research of the competition. The aim of this Master study is to analyze Lietuvos Gelezinkeliai SC position of its competitiveness and possibilities to improve it in cluster competition. There are comprehensively analyzed scientistic problems of sufficiency of clasters and of the management increasing. Completed analysis of the standard statement of the Republic of Lithuania that has influence to the competitiveness of Lietuvos Gelezinkeliai SC In the study presented research is based on M.Porter diamond method and on the theory of competition. There is corroborated scientific research hypothesis of the author about possibility to identify cluster of Lithuanian railway transport and its possibility to develop in a high competition range.

Food webs are good ecological macro-descriptors and their study is important in ecology in understanding nutrient cycles, tracing and quantifying energy and in describing trophic interactions within an ecosystem. The knowledge of food web finds applications in various natural sciences disciplines but also in many productive sectors. This study investigated the structure of the food web of a shallow lake in the Pantanal flood plain. The food web included two macrophytes, six aquatic insects, four crustaceans and 24 fish species. Sources of carbon for the various organisms living in the lake were identified through the values of δ13C exhibited by the organisms. The δ15N signature was used to estimate the trophic position of each organism. A cluster analysis based on the two isotopic signatures revealed six different feeding guilds and emphasized on the broad occurrence of omnivory among animals living in the lake. This study revealed that the use of food carbon was the most important factor that structured the lake community. Very low values of δ13C in zooplankton, benthic dwellers and bottom-feeder organisms as well as similarities between the gradient of δ13C and that of use of methane oxidizing bacteria informed on the possible use of biogenic methane as a source carbon and energy for the lake biota.

Galaxy clusters are powerful tools for studying various astrophysical principles. Gas accreting onto the cluster is heated to 10^7-10^8 K through adiabatic compression and shocks, making clusters highly luminous in X-ray imaging. Measurements of the gas density and temperature profiles can be used to calculate the gas mass fraction f_gas, which is expected to closely match the cosmic baryon fraction Ω_b/Ω_m. Recent observations have found entropy profiles in cluster outskirts that are shallower than predicted and values of f_gas that are higher than the Universal baryon fraction inferred from the Cosmic Microwave Background (CMB). Abell 133 was an ideal candidate for studying this phenomenon, since it had recently been observed in a wide (R≈30') Chandra mosaic with an exposure time of ∼2 Ms. The X-ray imaging was combined with existing optical imaging from the Canada-France-Hawaii Telescope (CFHT) and spectroscopy obtained from the Magellan telescope, to search for any possible gas clumps and to study their properties. The photometric analysis yielded over 3200 red sequence galaxies to a depth of r'=22.5, which were used to create a Gaussian smoothed intensity map and a significance map of the cluster (compared to CFHT Legacy Survey data). About 6 significant overdensities were detected in the significance map, although these did not fully correspond to contours obtained from the X-ray image. Spectroscopy obtained on the cluster yielded ∼700 secure redshifts, of which about 180 were cluster members. This included data from the NOAO Fundamental Plane Survey (NFPS) and the 6 Degree Field Galaxy Survey (6dFGS). We found a cluster redshift of z=0.0561±0.0002 and a velocity dispersion of σ=743±43 km/s. The dynamical analysis gave a virial radius of r_v=1.44±0.03 Mpc and a virial mass of M_v=(5.9±0.8)×10^14 M_sun. We also found values of R_500=1.21±0.07 Mpc and M_500=(5.3±0.9)×10^14 M_sun for γ=1/3 and R_500=0.99±0.05 Mpc and M_500=(2.9±0.5)×10^14 M_sun for γ=1/2, where γ is a parameter related to the assumed density profile and the velocity anisotropy. About 30 overdensities with a radius R_c≥30" were detected as gas clumps on the X-ray image. The galaxy distribution in these clumps was analyzed, both for the stacked signal as well as the individual clumps, in ten parallel colour-magnitude bands to find any significant red sequences associated with them. Most of these clumps appeared to be background systems, some consisting of 1-2 galaxies, others being small groups or clusters. Only 2-3 clumps appeared to be associated with the cluster itself. This suggests that the cluster density profile is actually quite smooth, which may not agree with recent numerical simulations. Further studies are required to determine if the cluster density distribution is consistent with what is predicted and the nature of the background systems.

No description available.