• Refine Query
  • Source
  • Publication year
  • to
  • Language
  • 2
  • Tagged with
  • 2
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • About
  • The Global ETD Search service is a free service for researchers to find electronic theses and dissertations. This service is provided by the Networked Digital Library of Theses and Dissertations.
    Our metadata is collected from universities around the world. If you manage a university/consortium/country archive and want to be added, details can be found on the NDLTD website.
1

Kompiliatorių optimizavimas IA-64 architektūroje / Compiler optimizations on ia-64 architecture

Varanavičius, Andrius 25 November 2010 (has links)
Šiame darbe buvo išnagrinėtos Intel Itanium (IA-64) architektūros savybės, įtakojančios kompiliatoriaus generuojamą kodą, ir išanalizuotos kompiliatoriaus optimizacijos, kurios buvo pritaikytos IA-64 architektūrai. Buvo prieita prie išvados, kad tokias optimizacijas galima susiskirstyti į kelis tipus. Pirmiausia nuo architektūros priklausomos optimizacijos, kurių efektyvumą galima padidinti išnaudojant predikaciją ir prognozavimo savybes ar kitas IA-64 specifines savybes. Antra, nuo architektūros nepriklausomos tradicinės optimizacijos, kurių pertvarkomo kodo efektyvumą galima padidinti parenkant kitokius šias optimizacijas valdančius kompiliavimo parametrus. Tyrime buvo išnagrinėtos ciklų optimizacijos, kurių kodą galimą būtų pakeisti valdomais parametrais. Tyrimas parodė, kad iš tiesų įmanoma sugeneruoti efektyvesnį kodą Intel Itanium architektūroje, keičiant šių parametrų reikšmes nuo numatytųjų reikšmių. / This thesis deeply explored Intel Itanium architecture features that improve a code generated by compiler. Compiler optimizations which are tuned to this architecture are also described. Accomplished research showed that there were several types of optimizations which can be improved on IA-64 architecture. Firstly, optimizations which are dependent on architecture can be optimized using predication and speculation or other unique IA-64 features. Secondly, optimizations that are undependable from traditional architecture can be improved using more aggressive compilation controllable parameters than they are by default. Loop optimizations were chosen for final research. Research proved that changing values of these parameters from default can improve program performance.
2

Optimizacinių ekonominės konkurencijos modelių tyrimas ir jų realizavimas interneto aplinkoje / Research of economical competition models and internet based realization of them

Skruodys, Nerijus 31 May 2004 (has links)
Future enterprises will be characterized by a focus on total quality, globalization, an object-oriented approach, and a business process-oriented approach. The globalization will lead to the 'virtual enterprise'. The virtual enterprise can obtain a competitive position by defining and reengineering its business processes. However, such reengineering requires an enterprise model. Also, as manufacturing is growing more difficult and competition gets more power, there are more useful to have practical realization of theoretic models. Traditional market theory is based on Walras model. There are lots of mathematic results, explaining some processes of competition of market. Programming realization would be a useful tool for simulating real economical processes. Realization on the internet would give possibility to show the algorithm on distance studies. The quality of service is defined by the average time lost by customers while waiting for services. Formal expressions of other quality measures are difficult. A customer prefers the server with lesser total service cost. The total cost includes the service price plus waiting losses. A customer goes away, if the total cost exceeds a certain critical level. A flow of customers is stochastic. Service times are stochastic, too. There is no known analytical solution for this model. The results are obtained by Monte-Carlo simulation. The first market model considers a simple case. Here the server rate depends on a single resource that... [to full text]

Page generated in 0.0265 seconds