Exploring Cross-layer power management for PGAS applications on the SCC platform

Technological limitations and power constraints are resulting in high-performance parallel computing architectures that are based on large numbers of high-core-count processors. Commercially available processors are now at 8 and 16 cores and experimental platforms, such as the many-core Intel Single-chip Cloud Computer (SCC) platform, provide much higher core counts. These trends are presenting new sets of challenges to HPC applications including programming complexity and the need for extreme energy efficiency.In this work, we first investigate the power behavior of scientific PGAS application kernels on the SCC platform, and explore opportunities and challenges for power management within the PGAS framework. Results obtained via empirical evaluation of Unified Parallel C (UPC) applications on the SCC platform under different constraints, show that, for specific operations, the potential for energy savings in PGAS is large; and power/performance trade-offs can be effectively managed using a cross-layerapproach. We investigate cross-layer power management using PGAS language extensions and runtime mechanisms that manipulate power/performance tradeoffs. Specifically, we present the design, implementation and evaluation of such a middleware for application-aware cross-layer power management of UPC applications on the SCC platform. Finally, based on our observations, we provide a set of recommendations and insights that can be used to support similar power management for PGAS applications on other many-core platforms.

A semantic model of selective dissemination of information for digital libraries

In this paper we present the theoretical and methodologicalfoundations for the development of a multi-agentSelective Dissemination of Information (SDI) servicemodel that applies Semantic Web technologies for specializeddigital libraries. These technologies make possibleachieving more efficient information management,improving agent–user communication processes, andfacilitating accurate access to relevant resources. Othertools used are fuzzy linguistic modelling techniques(which make possible easing the interaction betweenusers and system) and natural language processing(NLP) techniques for semiautomatic thesaurus generation.Also, RSS feeds are used as “current awareness bulletins”to generate personalized bibliographic alerts.

Cross recurrence quantification for cover song identification

There is growing evidence that nonlinear time series analysis techniques can be used to successfully characterize, classify, or process signals derived from realworld dynamics even though these are not necessarily deterministic and stationary. In the present study we proceed in this direction by addressing an important problem our modern society is facing, the automatic classification of digital information. In particular, we address the automatic identification of cover songs, i.e. alternative renditions of a previously recorded musical piece. For this purpose we here propose a recurrence quantification analysis measure that allows tracking potentially curved and disrupted traces in cross recurrence plots. We apply this measure to cross recurrence plots constructed from the state space representation of musical descriptor time series extracted from the raw audio signal. We show that our method identifies cover songs with a higher accuracy as compared to previously published techniques. Beyond the particular application proposed here, we discuss how our approach can be useful for the characterization of a variety of signals from different scientific disciplines. We study coupled Rössler dynamics with stochastically modulated mean frequencies as one concrete example to illustrate this point.

Towards the automatic merging of language resources

Language Resources are a critical component for Natural Language Processing applications. Throughout the years many resources were manually created for the same task, but with different granularity and coverage information. To create richer resources for a broad range of potential reuses, nformation from all resources has to be joined into one. The hight cost of comparing and merging different resources by hand has been a bottleneck for merging existing resources. With the objective of reducing human intervention, we present a new method for automating merging resources. We have addressed the merging of two verbs subcategorization frame (SCF) lexica for Spanish. The results achieved, a new lexicon with enriched information and conflicting information signalled, reinforce our idea that this approach can be applied for other task of NLP.

Asymmetric information and trade policy

Economists understand protectionism as a costly mechanism to redistribute from the average citizen to special-interest groups; yet political platforms that deviate from free trade have surprising popular appeal. I present an explanation based on heterogeneous information across citizens whose voting decision has an intensive margin. For each politician and each sector, the optimal trade-policy choice caters to the preferences of those voters who are more likely to be informed of that proposal. An overall protectionist bias emerges because in every industry producers are better informed than consumers. This asymmetry emerges in equilibrium because co-workers share industry-specific knwoledge, and because producers have greater incentives to engage in costly learning about their sector. My model implies that more widespread information about trade policy for an industry is associated with lower protection. Cross-sectoral evidence on U.S. non-tariff barriers and newspaper coverage is consistent with this prediction.

Interbank market integration under asymmetric information

We argue that the main barrier to an integrated international interbankmarket is the existence of asymmetric information between differentcountries, which may prevail in spite of monetary integration or successfulcurrency pegging. In order to address this issue, we study the scope forinternational interbank market integration with unsecured lending whencross-country information is noisy. We find not only that an equilibriumwith integrated markets need not always exist, but also that when it does,the integrated equilibrium may coexist with one of interbank marketsegmentation. Therefore, market deregulation, per se, does not guaranteethe emergence of an integrated interbank market. The effect of a repo marketwhich, a priori, was supposed to improve efficiency happens to be morecomplex: it reduces interest rate spreads and improves upon the segmentationequilibrium, but\ it may destroy the unsecured integrated equilibrium, sincethe repo market will attract the best borrowers. The introduction of othertransnational institutional arrangements, such as multinational banking,correspondent banking and the existence of "too-big-to-fail" banks mayreduce cross country interest spreads and provide more insurance againstcountry wide liquidity shocks. Still, multinational banking, as theintroduction of repos, may threaten the integrated interbank marketequilibrium.

Heterogeneous information and trade policy

Protectionism enjoys surprising popular support, in spite of deadweight losses. At thesame time, trade barriers appear to decline with public information about protection.This paper develops an electoral model with heterogeneously informed voters whichexplains both facts and predicts the pattern of trade policy across industries. In themodel, each agent endogenously acquires more information about his sector of employment. As a result, voters support protectionism, because they learn more about thetrade barriers that help them as producers than those that hurt them as consumers.In equilibrium, asymmetric information induces a universal protectionist bias. Thestructure of protection is Pareto inefficient, in contrast to existing models. The modelpredicts a Dracula effect: trade policy for a sector is less protectionist when there ismore public information about it. Using a measure of newspaper coverage across industries, I find that cross-sector evidence from the United States bears out my theoreticalpredictions.

Specificity revisited: The role of cross-investments

Previous analysis has shown that traders may opt for specific technologies with nojoint productivity advantage as a way to commit themselves to trading jointly, butonly when long-term contracting is infeasible. This paper proves that speciÞcity canalso be optimal (by relaxing the budget-balance constraint) in settings with long-termcontracting. Traders will opt for specificity when one trader makes a cross-investmentand either (1) this cross-investment has a direct externality on the other trader, (2) bothparties invest, or (3) private information is present. The specificity (e.g. from non-salvageable investments, specific assets and technologies, narrow business strategies,and exclusivity restrictions) is equally effective regardless of which trader's alternativetrade payoff is reduced. Specificity supports long-term contracts in a broad rangeof settings - both with and without renegotiation. The theory also offers a novelperspective on franchising and vertical integration.

Parental altruism under imperfect information: Theory and evidence

Can we reconcile the predictions of the altruism model of the familywith the evidence on intervivos transfers in the US? This paper expandsthe altruism model by introducing e ?ort of the child and by relaxingthe assumption of perfect information of the parent about the labormarket opportunities of the child. First, I solve and simulate a modelof altruism under imperfect information. Second, I use cross-sectionaldata to test a prediction of the model: Are parental transfers especiallyresponsive to the income variations of children who are very attached tothe labor market? The results suggest that imperfect information accountsfor several patterns of intergenerational transfers in the US.

Parental altruism under imperfect information: Theory and evidence

Can we reconcile the predictions of the altruism model of the family withthe evidence on parental monetary transfers in the US? This paper providesa new assessment of this question. I expand the altruism model by introducingeffort of the child and by relaxing the assumption of perfect informationof the parent about the labor market opportunities of the child. First,I solve and simulate a model of altruism and labor supply under imperfectinformation. Second, I use cross-sectional data to test the following prediction of the model: Are parental transfers especially responsive tothe income variations of children who are very attached to the labor market? The results of the analysis suggest that imperfect informationaccounts for many of the patterns of intergenerational transfers in theUS.

Research into the information referents of deaf signers

La sociedad de la información está ampliando los referentes informativos de las personas sordas signantes (usuarias de una lengua de signos). Con el objetivo de explorar las vías más utilizadas para acceder a la actualidad periodística por parte del colectivo mencionado, este artículo resume una investigación cuantitativa realizada con una encuesta a 138 personas Sordas: 69 en España y 69 en EE.UU. El trabajo se complementa cualitativamente con un focus group en la Gallaudet University de Washington, en American Sign Language (ASL), y con la observación de tres conferencias-magazines en Barcelona, para participantes sordos, en Lengua de Signos Catalana (LSC).

Browsing a component library using Non-functional Information

This paper highlights the role of non-functional information when reusing from a component library. We describe a method for selecting appropriate implementations of Ada packages taking non-functional constraints into account; these constraints model the context of reuse. Constraints take the form of queries using an interface description language called NoFun, which is also used to state non-functional information in Ada packages; query results are trees of implementations, following the import relationships between components. We define two different situations when reusing components, depending whether we take the library being searched as closed or extendible. The resulting tree of implementations can be manipulated by the user to solve ambiguities, to state default behaviours, and by the like. As part of the proposal, we face the problem of computing from code the non-functional information that determines the selection process.