On cluster resource allocation for multiple parallel task graphs

Many scientific applications can be structured as parallel task graphs (PTGs), that is, graphs of data-parallel tasks. Adding data parallelism to a task-parallel application provides opportunities for higher performance and scalability, but poses additional scheduling challenges. In this paper, we study the off-line scheduling of multiple PTGs on a single, homogeneous cluster. The objective is to optimize performance without compromising fairness among the PTGs. We consider the range of previously proposed scheduling algorithms applicable to this problem, from both the applied and the theoretical literature, and we propose minor improvements when possible. Our main contribution is an extensive evaluation of these algorithms in simulation, using both synthetic and real-world application configurations, using two different metrics for performance and one metric for fairness. We identify a handful of algorithms that provide good trade-offs when considering all these metrics. The best algorithm overall is one that structures the schedule as a sequence of phases of increasing duration based on a makespan guarantee produced by an approximation algorithm.     

Handling disruptions in manufacturing systems: An immune perspective

One of the major issues in the monitoring and control of manufacturing systems is to determine how to effectively deal with unexpected disruptions (e.g. material unavailability, resource failures, unavailability of operators, rush orders, etc.). Existing approaches and tools offer few concepts that are specific enough and sufficiently generic to help in handling a broad variety of such unexpected events. The biological immune system potentially offers interesting features to face the threats (bacteria, viruses, cancers, etc.) that may harm an organism. This research aims to investigate this potential for the monitoring and control of manufacturing systems at the occurrence of disruptions. Based on analogies that we point out, we suggest a framework to help with the design of software tools that are more able to assist decision makers in dealing with various types of disruptions occurring in a manufacturing system. A first prototype implementation, developed using a multi agent approach, contributes to show the feasibility and the interest of this immune based framework.     

User transparent data and task parallel multimedia computing with Pyxis-DT

The research area of Multimedia Content Analysis (MMCA) considers all aspects of the automated extraction of knowledge from multimedia archives and data streams. To satisfy the increasing computational demands of emerging MMCA problems, there is an urgent need to apply High Performance Computing (HPC) techniques. As most MMCA researchers are not also HPC experts, however, there is a demand for programming models and tools that are both efficient and easy to use. Existing user transparent parallelization tools generally use a data parallel approach in which data structures (e.g. video frames) are scattered among the available nodes in a compute cluster. For certain MMCA applications a data parallel approach induces intensive communication, however, which significantly decreases performance. In these situations, we can benefit from applying alternative approaches.We present Pyxis-DT, a user transparent parallel programming model for MMCA applications that employs both data and task parallelism. Hybrid parallel execution is obtained by run-time construction and execution of a task graph consisting of strictly defined building block operations. Results show that for realistic MMCA applications the concurrent use of data and task parallelism can significantly improve performance compared to using either approach in isolation. Extensions for GPU clusters are also presented.     

Database preference queries��a possibilistic logic approach with symbolic priorities

The paper presents a new approach to database preference queries, where preferences are represented in a possibilistic logic manner, using symbolic weights. The symbolic weights may be processed without assessing their precise value, which leaves the freedom for the user to not specify any priority among the preferences. The user may also enforce a (partial) ordering between them, if necessary. The approach can be related to the processing of fuzzy queries whose components are conditionally weighted in terms of importance. In this paper, importance levels are symbolically processed, and refinements of both Pareto ordering and minimum ordering are used. The representational power of the proposed setting is stressed, while the approach is compared with database Best operator-like methods and with the CP-net approach developed in artificial intelligence. The paper also provides a structured and rather broad overview of the different lines of research in the literature dealing with the handling of preferences in database queries.     

Automated optical method for ultrasonic bond pull force estimation

An automated method to non-destructively estimate ultrasonic bond pull force is presented and validated. Scanning white light interferometry (SWLI) measures bond geometry whereas the singular value decomposition of a SWLI image extracts the characteristics of the imaged bond geometry in terms of eigenvectors. Soft modeling selects those parts of the eigenvectors that are important for predicting the highest sustainable pull force. We show that SWLI measurement, statistical feature selection and bond pull force prediction can be combined to automatically perform non-destructive bond quality monitoring. Such automation removes subjectivity as well as operator limitations and errors present in earlier approaches (Schäfer et al., 2007), since no pre-selection of bond geometry or shape is needed. The proposed method was verified by experimental measurements on 132 single-point tape automated bonds. The results show that the method predicts maximum sustainable bond pull force with a prediction accuracy comparable to that of the operator based method. The three most important features in the image of the bond predicted the maximum sustainable bond pull force with an error of 11.4%. Not having to rely on the input of an experienced operator is the major advantage of this contribution.     

Multifunctional covalent and ionic coupling agents of maleic anhydride modified polyethylene

Amine and alcohol functional dendritic polymers were used as covalent coupling agents for maleic anhydride grafted polyethylene and their efficiency compared with metallic salt and zinc acetate, acting as an ionic coupling agent. The crosslinking reactions in the molten state were carried out at 150°C in an internal mixer and the resulting modified polyethylenes were characterized by dynamic mechanical analysis and Small angle X‐ray Scattering. These agents led to efficient coupling and crosslinking. The covalent coupling reactions were conducted in a heterogeneous multiphase system: The kinetics of crosslinking depends on the nature of the coupling agent: at 150°C, the reactions were completed within 5 min. with the amine hyperbranched polymer and 20 min. with the alcohol functionalized dendritic polymer. The zinc acetate forms ionic interactions with the grafted polyethylene in less than 1 min. Ionic interactions are thermo mechanically reversible; they correspond to electronic density fluctuations evidenced by SAXS. © 2007 Wiley Periodicals, Inc. J Appl Polym Sci, 2007     

Scheduling Task Parallel Applications for Rapid Turnaround on Enterprise Desktop Grids

Desktop Grids are popular platforms for high throughput applications, but due to their inherent resource volatility it is difficult to exploit them for applications that require rapid turnaround. Efficient desktop Grid execution of short-lived applications is an attractive proposition and we claim that it is achievable via intelligent resource selection. We propose three general techniques for resource selection: resource prioritization, resource exclusion, and task duplication. We use these techniques to instantiate several scheduling heuristics. We evaluate these heuristics through trace-driven simulations of four representative desktop Grid configurations. We find that ranking desktop resources according to their clock rates, without taking into account their availability history, is surprisingly effective in practice. Our main result is that a heuristic that uses the appropriate combination of resource prioritization, resource exclusion, and task replication can achieve performance within a factor of 1.7 of optimal in practice.     

Performance measurements on an ATM-based metropolitan area network: OASICE case study

The city of Colmar is developing a Metropolitan Area Network (MAN) called OASICE. Our MAN uses ATM technology to provide a range of business, academic and community tele-services such as tele-teaching, tele-information, tele-documentation, tele-management and tele-surveillance of property, high definition tele-conference, VOD (Video On Demand), etc. This revised version was published online in August 2006 with corrections to the Cover Date.     

Synthesis and Antimalarial Properties of Streptocyanine Dyes

Several streptocyanine dyes were synthesized that contain polymethine chains of varying length. Their in vitro antimalarial activities were evaluated against the virulent P. falciparum parasite. In addition to the influence of polymethine chain length, the effects of structural modifications at nitrogen end groups, para substitution of the phenyl groups, and counter‐anions were studied. The most potent antimalarial activities were found for heptacarbon chain streptocyanines, with an IC₅₀ value of 60 nM . Interestingly, most of the compounds were less cytotoxic toward the mammalian cells tested. The best selective toxicity profiles were found for pentacarbon chain streptocyanines, which have a good in vitro specificity index.