Exploiting Significance of Computations for Energy-Constrained Approximate Computing

Approximate execution is a viable technique for environments with energy constraints, provided that applications are given the mechanisms to produce outputs of the highest possible quality within the available energy budget. This paper introduces a framework for energy-constrained execution with controlled and graceful quality loss. A simple programming model allows developers to structure the computation in different tasks, and to express the relative importance of these tasks for the quality of the end result. For non-significant tasks, the developer can also supply less costly, approximate versions. The target energy consumption for a given execution is specified when the application is launched. A significance-aware runtime system employs an application-specific analytical energy model to decide how many cores to use for the execution, the operating frequency for these cores, as well as the degree of task approximation, so as to maximize the quality of the output while meeting the user-specified energy constraints. Evaluation on a dual-socket 16-core Intel platform using 9 kernels and applications shows that the proposed framework performs very close to an oracle always selecting the optimal configuration, both in terms of energy efficiency and quality of results. Also, a comparison with loop perforation (a well-known compile-time approximation technique), shows that the proposed framework results in significantly higher quality for the same energy budget.     

POMDP and MOMDP solutions for structural life-cycle cost minimization under partial and mixed observability

Scheduling of inspection and maintenance policies during the life-cycle of operating infrastructure necessitates optimization of long-term objectives in stochastic environments. Modern answers to the problem should focus on quantitative decision-making techniques, taking advantage of informative but uncertain data that become available in time. As such, the problem is efficiently addressed within the framework of stochastic dynamic programming by means of Partially Observable Markov Decision Processes (POMDPs) and Mixed Observability Markov Decision Processes (MOMDPs). Although these methodologies can provide very sophisticated solutions with optimality guarantees, important computational challenges often emerge, mainly due to the continuity of the multidimensional belief space on the probability simplex. In response, recent value iteration algorithms based on point-based approaches have been suggested, focusing on reachable belief points that can support an accurate value function. In this work, several POMDP and MOMDP point-based algorithms, with various characteristics regarding the exploration of the belief space and the value function update procedures, are rigorously analyzed. The algorithms are compared and evaluated in terms of accuracy and performance in stationary and nonstationary problems of structural inspection and maintenance for life-cycle cost minimization. Results are thoroughly discussed and several insights along with practical suggestions for similar problems are provided.     

High accuracy block-matching sub-pixel motion estimation through detection of error surface minima

Multimedia Tools and Applications - The present paper focuses on high-accuracy block-based sub-pixel motion estimation utilizing a straightforward error minimization approach. In particular, the...     

Determination of biogenic amines in wines by HPLC with precolumn dansylation and fluorimetric detection

Thirty-two samples of commercially available Greek wines were analysed in order to determine the content of biogenic amines. The method involves pre-column dansylation of the amines and subsequent solid phase extraction (SPE) of the derivatives through C₁₈ cartridges. For the analysis, RP-HPLC (reversed phase-high performance liquid chromatography) coupled with fluorimetric detection at excitation and emission wavelengths of 320 and 523 nm, respectively was used. All amines measured had recoveries over 85%. The highest detection limit was for agmatine (0.18 mg l⁻¹). Putrescine, cadaverine, histamine and isoamylamine are the most abundant amines in the samples analysed. The relative concentrations of biogenic amines expressed in mg l⁻¹ had as follows: putrescine > histamine > isoamylamine > ethylamine > methylamine > cadaverine = tyramine = agmatine = tryptamine. Higher amounts of biogenic amines were generally detected in wines, aged for long periods in barrels or in bottles. However, young wines contained lower amounts of these compounds as they were directly bottled after winemaking and have not undergone any further maturation processing. Moreover, less acid wines gave rise to higher histamine contents.     

Mineralogical and chemical properties of FGD gypsum from Florina,Greece

BACKGROUND: The aim of this work is to define the chemical and mineralogical composition of the fuel gas desulphurization (FGD) gypsum produced from the Meliti thermal power plant in the region of Florina in North West Greece, in order to investigate potential uses in the cement industry. Mineralogical and microprobe analyses were carried out on FGD gypsum samples collected from the Meliti 330 MW lignite‐fired power plant. RESULTS: Results show that the main component of the FGD gypsum is pure mineral gypsum (CaSO₄·2H₂O). The particle size of the gypsum ranges from 5 to 50 µm and the crystals are mainly of rhomboid shape. Microprobe analysis shows that the concentration of CaO and SO₃, which are the main components, range from 31.9%–32.5% and from 45.90–46.40%, respectively. CONCLUSION: This FGD gypsum can easily substitute the natural gypsum used in the production of cement. Copyright © 2007 Society of Chemical Industry     

Performing integrated system tests using malicious component insertion

In this paper, a testing method suitable for strengthening fault tolerance in the event of unexpected situations within a software system is presented. It is based on the idea of testing an integrated system, by substituting system components with other, similar in design and functionality that operate in an erroneous and even malicious manner. The approach adopted, is similar to the concept of inserting a virus within an organization so that the defense mechanisms of the latter can be tested and the necessary lines of defense are formed, so that the virus cannot affect any of the organization critical parts. The focal point is to ensure that in case of a module malfunction, the integrated system will continue to operate, isolating the malfunctioning software at the greatest possible extend, preventing the erroneous behavior from affecting other (and sometimes critical) modules. The testing method proposed is based first on isolated components testing adopting and enhancing the Component Off The Self method, and second on integrated system testing using malicious components that emulate erroneous operation.     

Numerical studies on the effect of plan irregularities in the progressive collapse of steel structures

This research examines the effect of plan irregularities on the progressive collapse of four steel structures located in regions with different seismic activity. The plans of the first and second structure are irregular, whilst those of the third and fourth structures are regular. The collapse patterns of the four buildings are examined and compared under seven loading scenarios using non-linear dynamic and static analyses. In the non-linear dynamic analyses, node displacements above the removed columns and the additional force on the columns adjacent to them are discussed. Furthermore, the strength and capacities of the columns are compared to determine their susceptibility to collapse. In the non-linear static analyses, the pushdown curve and yield load factor of the structures are obtained after column removal. The results indicate that an irregular structure designed in site class C seismic zone, collapses in most of the column-removal scenarios. Moreover, when comparing regular and irregular structures designed in site class E seismic zone, the demand force to capacity ratio (D/C) of the columns in the irregular structures is on average between 1.5 and 2 times that of the regular ones.     

A Small Area Charge Sensitive Readout Chain with a Dual Mode of Operation

A charge sensitive readout chain has been designed and fabricated in acommercially available 0.8 m CMOS technology. The readout chain is optimizedfor pixel detectors measuring soft X-ray energies up to 20 KeV. In the first modean analog signal proportional to input charge is generated and processed in realtime. In the second mode a peak-and-hold operation is enabled and therelevant signal is processed in later time. This dual mode of operation iscontrolled by an external digital signal. The readout chain consists of a chargeamplifier, a shaper, an operational amplifier which can either operate as avoltage amplifier or a peak detector and an output buffer. Its area is . The gain at the shaper output is 378 mv/fC, theENC is 16 rms at 160 nsec shaping time. The overall gainis 557 mV/fC, the ENC is rms with 240 nsec peaking timeand 1.4 sec recovery time. The overall power dissipation is 1.5 mWatt with aload capacitance of 25 pF.     

Automatic derivation of BPEL4WS from IDEF0 process models

Integration definition for function modelling (IDEF0) is one of the most popular notations for modelling business processes. It employs a rather simple and intuitive modelling construct, consisting of boxes representing functions and arrows connecting them signifying flow of information and materials. Web services on the other hand are an emerging technology for implementing distributed systems. Web service orchestration languages, such as Business Process Execution Language for Web Services (BPEL4WS), are the emerging approach for describing processes as networks of coordinated web services. Business processes as captured in IDEF0 models, however, may contain both web services as well as other types of activities which need to be coordinated. By automatically analysing the Extensible Markup Language (XML) definition of an IDEF0 model, we can identify how web services interact with other activities and at runtime generate code to support the orchestration of web services with the overall business process. The approach proposed is independent of the orchestration language and ensures an implementation independent model for specifying web service orchestrations. This approach also enables the top-down analysis of a business process to its constituent web services and avoids any misalignment problems during design time between the two.