Modern architectures allow multiple instructions to be issued at once and have other complex features. To account for this, compilers perform instruction scheduling after generating the output code. The instruction scheduling problem is to find an optimal schedule given the limitations and capabilities of the architecture. While this can be done optimally, a greedy algorithm known as list scheduling is used in practice in most production compilers. <br /><br /> List scheduling is generally regarded as being near-optimal in practice, provided a good choice of heuristic is used. However, previous work comparing a list scheduler against an optimal scheduler either makes the assumption that an idealized architectural model is being used or uses too few test cases to strongly prove or disprove the assumed near-optimality of list scheduling. It remains an open question whether or not list scheduling performs well when scheduling for a realistic architectural model. <br /><br /> Using constraint programming, we developed an efficient optimal scheduler capable of scheduling even very large blocks within a popular benchmark suite in a reasonable amount of time. I improved the architectural model and optimal scheduler by allowing for an issue width not equal to the number of functional units, instructions that monopolize the processor for one cycle, and non-fully pipelined instructions. I then evaluated the performance of list scheduling for this more realistic architectural model. <br /><br /> I found that when scheduling for basic blocks when using a realistic architectural model, only 6% or less of schedules produced by a list scheduler are non-optimal, but when scheduling for superblocks, at least 40% of schedules produced by a list scheduler are non-optimal. Furthermore, when the list scheduler and optimal scheduler differed, the optimal scheduler was able to improve schedule cost by at least 5% on average, realizing maximum improvements of 82%. This suggests that list scheduling is only a viable solution in practice when scheduling basic blocks. When scheduling superblocks, the advantage of using a list scheduler is its speed, not the quality of schedules produced, and other alternatives to list scheduling should be considered.

This degree project concludes the Innovation and Design Engineering programme at Karlstad University. The project was carried out by David Brocker, Erik Hallberg and Andreas Hertzman during the spring of 2006 and corresponded to 15 weeks of work per student. Assigner was Martin Larsson at Arexor and instructor at Karlstad University was Monica Jakobsson. Arexor has a patent to a construction of an electrical skirting board with moveable wall sockets, called Electrical Track System. The lowest parts of the interior walls along the floor are today almost always covered by a conventional skirting board. Arexors patented product replaces that with the possibility to move and increase the number of wall sockets by choice along an electrical skirting board. Conventional wall sockets are limited due to fixed positions and are not transferable. It causes a problem when the number of fixed wall sockets controls the possibilities and not the users demand. The assignment commissioned by Arexor to the students was to improve and develop the Electrical Track System because the patent did not fulfil the requirements for CE certification. The project results were to be used by Arexor as the basic data when the product was tested by ETL SEMKO. The objective of the project was to present directions for production to a functioning prototype of the Electrical Track System, within the estimated time period. The prototype was to be shown at the exhibition for degree projects at Karlstad University in May 2006. The objective also included to create a product brochure and a display case. The development method were divided into five phases: preparation phase, research phase, idea generating phase, conception development phase and concretisation phase. They were carried out linear up to the conception development phase. Then iteration between the research phase, idea generating phase, conception development phase was repeated until satisfied result was achieved. The result included a number of functioning prototypes, a display case, a product brochure, CAD drawings, renderings and this academic report. The prototypes were manufactured by Modellteknik in Eskilstuna, Sweden. Parts of the result cannot be presented due to that the solutions was not, at the time, protected by the patent. Arexor announced in the end of the project that they would apply for a new patent that included our solutions. Because of that these solutions could not be shown in public. Some parts in the report therefore refer to secrecy. / Detta examensarbete var avslutningen på Innovations- och designingenjörsprogrammet vid Karlstads universitet på fakulteten för teknik- och naturvetenskap. Arbetet genomfördes av David Brocker, Erik Hallberg och Andreas Hertzman under våren 2006 på uppdrag av Arexor och omfattade 15 högskolepoäng per student. Uppdragsgivare på Arexor var Martin Larsson och handledare var Monica Jakobsson från Karlstads universitet. Arexor har patenterat en konstruktion av en elektrisk golvlist med flyttbara vägguttag, kallat Electrical Track System. Längs med golvet utmed väggarna sitter i fastigheter i dag nästan alltid en golvlist. Arexors patenterade produkt ersätter golvlisten och ger möjlighet att bland annat flytta, montera och öka antalet vägguttag längs med listen efter eget behov. Traditionella uttag begränsas av att de har fasta positioner och inte kan omplaceras. Det innebär problem då antalet fasta uttag styr möjligheterna och inte behovet. Uppdraget som Arexor gav projektgruppen var att vidareutveckla Electrical Track System, då patentet inte uppfyllde kraven för CE-certifiering. Resultatet av arbetet skulle sedan användas av Arexor som underlag vid test hos ETL SEMKO. Målet var att ta fram tillverkningsunderlag för en slutgiltig, fungerande prototyp av ETS inom tidsramen för projektet. Den skulle sedan visas på examensutställningen vid Karlstads universitet den 30 maj 2006. Delmål var även att ta fram en produktbroschyr och en utställningsmonter. Arbetet delades in i stegen arbetsplan, förstudie, idégenerering, konceptutveckling och konkretisering. De genomfördes linjärt upp till konceptutvecklingen, sedan skedde en iteration mellan stegen förstudie, idégenerering och konceptutveckling. Det betyder att processen upprepades tills det att önskat resultat uppnåddes. Resultatet blev en mängd olika fungerande prototyper, en utställningsmonter, en produktbroschyr, CAD-ritningar, renderingar och denna akademiska rapport. Prototyperna tillverkades i samarbete med Modellteknik i Eskilstuna. Delar av resultatet redovisas inte på grund av att lösningarna inte vid tidpunkten skyddades av rådande patent. Arexor meddelande i slutet av projektet att företaget skulle ansöka om ett nytt patent där dessa lösningar inkluderades. Det medförde att lösningarna inte fick offentliggöras innan patentansökan hade lämnats in. Därför hänvisas till sekretess i vissa delar av rapporten.

This thesis discusses cache oblivious data structures. These are structures which have good caching characteristics without knowing Z, the size of the cache, or L, the length of a cache line. Since the structures do not require these details for good performance they are portable across caching systems. Another advantage of such structures isthat the caching results hold for every level of cache within a multilevel cache. Two simple data structures are studied; the array used for binary search and the linear list. As well as being cache oblivious, the structures presented in this thesis are space efficient, requiring little additional storage. We begin the discussion with a layout for a search tree within an array. This layout allows Searches to be performed in O(log n) time and in O(log n/log L) (the optimal number) cache misses. An algorithm for building this layout from a sorted array in linear time is given. One use for this layout is a heap-like implementation of the priority queue. This structure allows Inserts, Heapifies and ExtractMaxes in O(log n) time and O(log nlog L) cache misses. A priority queue using this layout can be builtfrom an unsorted array in linear time. Besides the n spaces required to hold the data, this structure uses a constant amount of additional storage. The cache oblivious linear list allows scans of the list taking Theta(n) time and incurring Theta(n/L) (the optimal number) cache misses. The running time of insertions and deletions is not constant, however it is sub-polynomial. This structure requires e*n additional storage, where e is any constant greater than zero.

Modern architectures allow multiple instructions to be issued at once and have other complex features. To account for this, compilers perform instruction scheduling after generating the output code. The instruction scheduling problem is to find an optimal schedule given the limitations and capabilities of the architecture. While this can be done optimally, a greedy algorithm known as list scheduling is used in practice in most production compilers. <br /><br /> List scheduling is generally regarded as being near-optimal in practice, provided a good choice of heuristic is used. However, previous work comparing a list scheduler against an optimal scheduler either makes the assumption that an idealized architectural model is being used or uses too few test cases to strongly prove or disprove the assumed near-optimality of list scheduling. It remains an open question whether or not list scheduling performs well when scheduling for a realistic architectural model. <br /><br /> Using constraint programming, we developed an efficient optimal scheduler capable of scheduling even very large blocks within a popular benchmark suite in a reasonable amount of time. I improved the architectural model and optimal scheduler by allowing for an issue width not equal to the number of functional units, instructions that monopolize the processor for one cycle, and non-fully pipelined instructions. I then evaluated the performance of list scheduling for this more realistic architectural model. <br /><br /> I found that when scheduling for basic blocks when using a realistic architectural model, only 6% or less of schedules produced by a list scheduler are non-optimal, but when scheduling for superblocks, at least 40% of schedules produced by a list scheduler are non-optimal. Furthermore, when the list scheduler and optimal scheduler differed, the optimal scheduler was able to improve schedule cost by at least 5% on average, realizing maximum improvements of 82%. This suggests that list scheduling is only a viable solution in practice when scheduling basic blocks. When scheduling superblocks, the advantage of using a list scheduler is its speed, not the quality of schedules produced, and other alternatives to list scheduling should be considered.

A study of packing constructs a critique of the everyday: a dialogue between chaos and order, surface and area, interior and exterior, gravity and lightness. In search of tangible expression of the spatial processes I am responsible for, I have become both master architect and expert packer. I have composed this thesis the same way I pack: I have assembled piles of fragments, regrouped them, reconsidered, edited, alloted them more or less space. Things have become more and less valuable. Quotes and images are precious, like artifacts, tucked delicately between text-filled pages. Each word I write, each line I draw, creates a boundary, a parcel, a unit of space set apart from the white of the page. This book is my suitcase.

In this dissertation, advanced channel decoding techniques based on bit-level soft information are studied. Two main approaches are proposed: bit-level probabilistic iterative decoding and bit-level algebraic soft-decision (list) decoding (ASD). In the first part of the dissertation, we first study iterative decoding for high density parity check (HDPC) codes. An iterative decoding algorithm, which uses the sum product algorithm (SPA) in conjunction with a binary parity check matrix adapted in each decoding iteration according to the bit-level reliabilities is proposed. In contrast to the common belief that iterative decoding is not suitable for HDPC codes, this bit-level reliability based adaptation procedure is critical to the conver-gence behavior of iterative decoding for HDPC codes and it significantly improves the iterative decoding performance of Reed-Solomon (RS) codes, whose parity check matrices are in general not sparse. We also present another iterative decoding scheme for cyclic codes by randomly shifting the bit-level reliability values in each iteration. The random shift based adaptation can also prevent iterative decoding from getting stuck with a significant complexity reduction compared with the reliability based parity check matrix adaptation and still provides reasonable good performance for short-length cyclic codes. In the second part of the dissertation, we investigate ASD for RS codes using bit-level soft information. In particular, we show that by carefully incorporating bit¬level soft information in the multiplicity assignment and the interpolation step, ASD can significantly outperform conventional hard decision decoding (HDD) for RS codes with a very small amount of complexity, even though the kernel of ASD is operating at the symbol-level. More importantly, the performance of the proposed bit-level ASD can be tightly upper bounded for practical high rate RS codes, which is in general not possible for other popular ASD schemes. Bit-level soft-decision decoding (SDD) serves as an efficient way to exploit the potential gain of many classical codes, and also facilitates the corresponding per-formance analysis. The proposed bit-level SDD schemes are potential and feasible alternatives to conventional symbol-level HDD schemes in many communication sys-tems.

Introduction The number of the people who needs dialysis escalates rapidly year each. The rising cost of dialysis is relative. Because Bureau National Health Insurance endures the co-payment of hemodialysis, patients have freedom of choice. Two critical factors that influence patient¡¦s decision in medical care are access and quality. Facing pressure from competition, many hemodialysis service providers believe that high quality service will retain patients to return as well as maintain a healthy physician-patient relationship. Therefore, they apply many quality management tools, which also include service pathways. There are few investigations that discuss service pathways. Thus, this research constructs a model to analyze the effect of service pathways in hemodialysis service providers. Method From Nov. 1, 2003 to Dec. 31, 2004, the researcher interviewed hemodialysis nurses to develop flow chart, customer encounter, and check list for hemodialysis. We also collected the check list records of the hemodialysis patients to examine the effect. Conclusion First, the construction of total service pathways comprised three critical interlocking phases: flow chart, customer encounter, and check list. After that we shall evaluate and correct them continuously. Secondly, the researcher found that most patients are routine members, and come from long-term care center and respiratory care ward (RCW). Thirdly, the indicators influencing hemodialysis results and error incidents were steady. Finally, in the value-added services dimensions, the indicators of nurses¡¦ performance on greeting and asking patients if they want to have a meal were significant, but the indicators of telephone and inpatient interview were unfavorable.

Αποτελεί γεγονός ότι τα κινητά τηλέφωνα χρησιμοποιούνται με όλο και αυξανόμενο ρυθμό για τη συλλογή μεγάλου όγκου δεδομένων αποσκοπώντας στην ανάλυση της ανθρώπινης συμπεριφοράς. Όσο τα κινητά τηλέφωνα εξελίσσονται και τα smartphones διαμορφώνουν μια έντονα ανταγωνιστική και επικερδή αγορά, δημιουργούνται νέα δεδομένα στις λειτουργίες και τις εφαρμογές των κινητών. Η λίστα επαφών μπορεί πλέον να αποτελείται από χιλιάδες εγγραφές καθιστώντας αρκετά χρονοβόρα την αναζήτηση κάποιας επαφής. Στην παρούσα διπλωματική εργασία παρουσιάζεται η πιθανότητα ο χρήστης να καλέσει μια επαφή του βάσει της τρέχουσας τοποθεσίας και του αρχείου κλήσεών του. Η εξαγωγή αυτών των πιθανοτήτων απαιτεί την ομαδοποίηση των κλήσεων του χρήστη βάσει της τοποθεσίας του. Αυτό επιτυγχάνεται με τη χρήση κατάλληλου αλγορίθμου clustering, ο οποίος κατηγοριοποιεί γεωγραφικά τις κλήσεις του χρήστη. Επίσης, έχοντας πρόσβαση στο αρχείο κλήσεων του χρήστη επιστρέφονται από τον αλγόριθμο ποιες επαφές έχουν κληθεί από κάθε τοποθεσία που έχει κάνει κλήση ο χρήστης. Αξιοποιώντας αυτή την πληροφορία υπολογίζεται η πιθανότητα να κληθεί κάθε μία από τις επαφές του από τη συγκεκριμένη τοποθεσία. Ο αλγόριθμος που παρουσιάζεται στη διπλωματική εργασία θα μπορούσε να χρησιμοποιηθεί για την δημιουργία μιας πιο εύχρηστης λίστας επαφών. Τα περισσότερα κινητά διαθέτουν την επιλογή να ορίσει ο χρήστης ποιες επαφές του είναι αυτές που χρησιμοποιεί πιο συχνά αλλά και το αρχείο κλήσεων που εμφανίζει τις κλήσεις με χρονολογική σειρά. Σε καμία από αυτές τις επιλογές δεν αποτελεί παράγοντας η γεωγραφική τοποθεσία του χρήστη. Ο άνθρωπος ως μέλος της κοινωνίας χρησιμοποιεί το κινητό του τηλέφωνο ως μέσο επικοινωνίας και αλληλεπίδρασης με άλλους ανθρώπους. Κατά τη διάρκεια της ημέρας οι τοποθεσίες που επισκέπτεται ο χρήστης διαφέρουν ανάλογα με την επαγγελματική του ιδιότητα, την προσωπικότητα, την ηλικία αλλά την περίοδο του χρόνου. Τις ώρες που ο άνθρωπος βρίσκεται στο χώρο εργασίας του, είναι πιο πιθανό να επικοινωνεί τηλεφωνικά με συνεργάτες του ενώ το βράδυ όταν βρίσκεται στην εστία του είναι αναμενόμενο να συνομιλεί με οικείους και φίλους. Επίσης ειδικές περίοδοι του χρόνου όπως οι διακοπές των γιορτών ή οι διακοπές το καλοκαίρι, συνεπάγονται οι άνθρωποι να ταξιδεύουν στις πόλεις από όπου κατάγονται. Καθίσταται σαφές ότι σε αυτές τις περιπτώσεις η επικοινωνία του χρήστη διαφοροποιείται καλώντας επαφές που δεν χρησιμοποιεί στην καθημερινότητα του. Η τοποθεσία του ατόμου αποτελεί συνεπώς κύριο παράγοντα στην επιλογή της επαφής με την οποία θα επικοινωνήσει. Τέλος, αξιολογήθηκαν τα αποτελέσματα και η ακρίβεια της προσέγγισης που παρουσιάστηκε στη διπλωματική εργασία και εξήχθησαν τα αντίστοιχα συμπεράσματα. / It is a fact that mobile phones are used with an increasing pace to collect large amounts of data in order to analyze the human behavior. As mobile phones evolve and smartphones are shaping a highly competitive and lucrative market, new data on functions and mobile applications are created. The contact list is now composed of thousands of records making it quite time consuming to search a contact. This thesis presents the possibility for the user to call a contact under the current location and call log. The algorithm presented in this thesis could be used to create a more user-friendly contact list. Most mobile devices give the user the option to set which contacts are those used most often and the call log that shows calls in chronological order. None of these options take under consideration the geographical location of the user. Man as a member of society uses a mobile phone as a medium of communication and interaction with other people. During the day the sites the user visits vary depending on the professional status, personality, age, but the time period. The time that the man is in the workplace, they are more likely to communicate by phone with colleagues whereas in the evening when at home, it is expected to talk with relatives and friends. Also specific periods of time such as vacation holiday or summer vacation, involve people to travel to cities where they originate from. It is clear that in these cases the communication of the user differs calling contacts not used in everyday life. The location of the individual is therefore a key factor in the selection of contact which will communicate. Finally, the results and the accuracy of the approach presented in the thesis were evaluated and we exported the corresponding conclusions.

No description available.

This dissertation investigates three topics related to imagereconstruction from list-mode Anger camera data. Our mainfocus is the processing of photomultiplier-tube (PMT)signals directly into images. First we look at the use of list-mode calibration data toreconstruct a non-parametric likelihood model relating theobject to the data list. The reconstructed model can thenbe combined with list-mode object data to produce amaximum-likelihood (ML) reconstruction, an approach we calldouble list-mode reconstruction. This trades off reducedprior assumptions about the properties of the imaging systemfor greatly increased processing time and increaseduncertainty in the reconstruction. Second we use the list-mode expectation-maximization (EM)algorithm to reconstruct planar projection images directlyfrom PMT data. Images reconstructed by EM are compared withimages produced using the faster and more common techniqueof first producing ML position estimates, then histogramingto form an image. A mathematical model of the human visualsystem, the channelized Hotelling observer, is used tocompare the reconstructions by performing the Rayleigh task,a traditional measure of resolution. EM is found to producehigher resolution images than the histogram approach,suggesting that information is lost during the positionestimation step. Finally we investigate which linear parameters of an objectare estimable, in other words may be estimated without biasfrom list-mode data. We extend the notion of a linearsystem operator, familiar from binned-mode systems, tolist-mode systems, and show the estimable parameters aredetermined by the range of the adjoint of the systemoperator. As in the binned-mode case, the list-modesensitivity functions define ``natural pixels'' with whichto reconstruct the object.