Within this project you will make Calimero OSGI (Open Services Gateway Initiative) ready. OSGI provides a framework and service platform for deploying modular Java components, so called bundles. OSGI is a platform that can be used to create home automation gateways. Bundles have a lifecycle within an OSGI container and provide only a defined scope of their classes to other bundles within the container, which eases and enforces proper modularization of large Java projects. The goal of this project is to make Calimero OSGI ready, which means to create a bundle activator responsible for establishing connections to a KNX bus system. Furthermore, the API classes and interfaces that should be visible for other bundles need to be defined as exported classes and any dependencies on other software bundles need also to be defined and specified in the according manifest attributes of a .jar file. Within the project the build process to create an OSGI bundle should be automated and automated unit tests should verify that all the Calimero features are working within an OSGI container.

Benefit for the Student

Work on a well-established and industry acknowledged open source project. Gain experience in automation technology KNX, OSGI, build tools and automatic unit testing.

Benefit for the Project

Calimero is used in several other open source projects. A direct OSGI support within Calimero would allow to easily use Calimero as OSGI bundle in other software projects and avoid the redundancy of each project defining its own OSGI bundle for Calimero.

Requirements

Good Java programming skills
Know-how in OSGI, Build Tools (Ant/Maven/Gradle) and automatic unit testing (JUnit) is a plus.

Mentors

Wolfgang Kastner, Markus Jung

Contact

Mentors are regularly around in our GSoC IRC channel #TU-CSE-SoC at irc.freenode.net. You can also reach us via the mailinglist – send an email to soc2013@iue.tuwien.ac.at using the prefix [CALIMERO] (a subscription is required).

More Information

http://www.iue.tuwien.ac.at/cse/wiki/doku.php?id=osgi_module
http://sourceforge.net/p/calimero/wiki/Home/

Eclipse Array Explorer allows developers to explore large arrays efficiently. The tabular view provided by this tool offers an overview of the values of all fields in all elements in the array. Specific values in this table can be spot using interaction or by means of additional views (such as bar charts).

More details: http://www.cvast.tuwien.ac.at/projects/visualdebugging/ArrayExplorer

Public Source code with BSD license: https://github.com/bilalsal/ArrayExplorer

Short article: http://publik.tuwien.ac.at/files/PubDat_207584.pdf

Video: http://www.youtube.com/watch?v=7gTv9yUtFBc

Benefit for the Student

Learning how to develop plugins for the Eclipse IDE. Eclipse is one of the most widely used IDEs for Java, and supports a variety of other languages. A variety of plugins exist for eclipe to enrich its functionalities. These plugins profit from the extensive functionalities and Eclipse already supports.

Learning Eclipse Rich Client Platform (RCP), an advanced platforms for frontend development in Java.

Learning how to develope a tool with a productive quality, and gaining credits as a very good developer in case the tool was adpoted by the Eclipse community.

Learning how to program basic visualizations that are highly usabile and reliable.

Benefit for the Project

Releasing the Eclipse Array Explorer with a productive quality. This will enable releasing the tool to the public and get hopefully adpoted by the Eclipse community.

The tool has a potential to make debugging programs having large arrays and collections much easier than with the existing Eclipse debugger.

Mentors

Bilal Alsallakh, Markus Bögl

Contact

Mentors are regularly around in our GSoC IRC channel #TU-CSE-SoC at irc.freenode.net. You can also reach us via the mailinglist – send an email to soc2013@iue.tuwien.ac.at using the prefix [CVAST](a subscription is required).

More information

http://www.iue.tuwien.ac.at/cse/wiki/doku.php?id=eclipse_array_explorer

EclipseTracer allows developers to adapt selected Eclipse break-points and watch-points into trace-points while debugging Java programs. Instead of stopping the execution, trace-points only record information about the current state when they are hit and store them in per-instance traces. The "silent" hits can be visualized over a time-line for post-mortem examination of the recorded information.

More details: http://www.cvast.tuwien.ac.at/projects/visualdebugging/EclipseTracer

Public Source code with BSD license: https://github.com/bilalsal/EclipseTracer

Short article: http://publik.tuwien.ac.at/files/PubDat_207587.pdf

Video: http://www.youtube.com/watch?v=GNuRbPrhvrw

Benefit for the Student

Learning how to develop plugins for the Eclipse IDE. Eclipse is one of the most widely used IDEs for Java, and supports a variety of other languages. A variety of plugins exist for eclipe to enrich its functionalities. These plugins profit from the extensive functionalities and Eclipse already supports.

Learning Eclipse Rich Client Platform (RCP), an advanced platforms for frontend development in Java.

Learning how to develope a tool with a productive quality, and gaining credits as a very good developer in case the tool was adpoted by the Eclipse community.

Learning how to program basic visualizations that are highly usabile and reliable.

Benefit for the Project

Releasing the EclipseTracer with a productive quality. This will enable releasing the tool to the public and get it hopefully adpoted by the Eclipse community.

The tool has a potential to make understading traces and variable histories easier than with the existing Eclipse debugger.

Mentors

Bilal Alsallakh, Markus Bögl

Contact

Mentors are regularly around in our GSoC IRC channel #TU-CSE-SoC at irc.freenode.net. You can also reach us via the mailinglist – send an email to soc2013@iue.tuwien.ac.at using the prefix [CVAST] (a subscription is required).

More information

http://www.iue.tuwien.ac.at/cse/wiki/doku.php?id=eclipse_tracer

The gateway currently keeps all oBIX objects and data point histories in memory. If the gateway is restarted all previous data points are lost. Within this project you will create a persistence layer based on a ORM framework (e.g. JPA) and allow the storage of oBIX objects in a database. Due to the aim of deploying the gateway on constrained devices (e.g. Raspberry PI) a light-weight database concept should be developed and implemented.

Benefit for the Student

Dive into latest technologies and emerging protocols for the Internet of Things/Web of Things. Gain hands on experience in Java based software development and latest Java persistence APIs and frameworks.

Benefit for the Project

A persistence layer based on a relational database is a starting point for long term observation and monitoring of buildings integrated by the IoTSyS gateway. The database can be extended in the future to be used for data mining and data analysis.

Requirements

Strong skills in Java based software development, relational data base design and Java Persistence APIs and frameworks are required.

Mentors

Markus Jung, Andreas Fernbach

Contact

Mentors are regularly around in our GSoC IRC channel #TU-CSE-SoC at irc.freenode.net. You can also reach us via the mailinglist – send an email to soc2013@iue.tuwien.ac.at using the prefix [IOTSYS] (a subscription is required). 

More Information

http://www.iue.tuwien.ac.at/cse/wiki/doku.php?id=gateway_persistence_layer
http://code.google.com/p/iotsys
https://www.oasis-open.org/committees/obix/ (oBIX)
http://www.oracle.com/technetwork/java/javaee/tech/persistence-jsp-140049.html (Java Persistence API)

IoTSyS is currently based on a simple oBIX object container, where most of the dependency between different components and services is wired hardcoded together. This design should be refactored to use the Inversion of Control or Dependency Injection pattern in order to increase the maintainability and extensibility of the oBIX server. Currently the protocol handlers, the message encoders/decoders and the oBIX server are hard-wired together. There is a great potential to refactor the whole system design in order to apply state of the art component oriented design patterns. Furthermore, a smart object represented as oBIX object should be identified through different addressing schemes within the internal object broker. Currently, this design is strongly based on extensions of the oBIX Toolkit and need to be redesigned in order to support multiple addressing schemes to identify an oBIX object through different addressing schemes (e.g. IPv6 global unicast address, building topology, functional domain).

Benefit for the Student

Dive into latest technologies and emerging protocols for the Internet of Things/Web of Things. Gain hands on experience in Java based software development in design and realization of component oriented execution containers. Apply state of the art knowledge for the design of component containers and design patterns for server development.

Benefit for the Project

The maintainability and extensibility of the gateway project will be improved allowing to easily adding new components and services in between the existing request processing.

Requirements

Strong skills in Java based software development are necessary.
Know-how about OSGI, IoC container design and implementation, Java byte code modification frameworks is a plus.

Mentors

Markus Jung, Andreas Fernbach

Contact

Mentors are regularly around in our GSoC IRC channel #TU-CSE-SoC at irc.freenode.net. You can also reach us via the mailinglist – send an email to soc2013@iue.tuwien.ac.at using the prefix [IOTSYS] (a subscription is required). 

More Information

http://www.iue.tuwien.ac.at/cse/wiki/doku.php?id=obix_server_container_redesign
http://code.google.com/p/iotsys
http://martinfowler.com/articles/injection.html (IoC, Dependency Injection)
https://www.oasis-open.org/committees/obix/ (oBIX)

IoTSyS uses a simple oBIX serve for the gateway that partly implements the oBIX 1.1 WD 06 specifications. Within this project you will complete the implementation and add additional features to the gateway. This includes the full support of oBIX watches, histories and alarming. Furthermore, automated unit testing should be introduced in order to test the compliance with the standard specification.

Benefit for the Student

Dive into latest technologies and emerging protocols for the Internet of Things/Web of Things. Gain hands on experience in Java based software development and implementation based on standard specifications. oBIX provides a convenient way to integrate home and building automation technologies through RESTful Web service interfaces based on HTTP and XML.

Benefit for the Project

The gateway currently supports only parts of the oBIX specification. A more complete implementation would provide the basis for more evaluations and beneficial for any further standardization activities. Furthermore the project can act as a basis for a reference implementation for new standard enhancements.

Requirements

Strong skills in Java based software development are necessary.
Know-how about XML data processing and RESTful Web services is an advantage.

Mentors

Markus Jung, Andreas Fernbach

Contact

Mentors are regularly around in our GSoC IRC channel #TU-CSE-SoC at irc.freenode.net. You can also reach us via the mailinglist – send an email to soc2013@iue.tuwien.ac.at using the prefix [IOTSYS] (a subscription is required).

More Information

http://www.iue.tuwien.ac.at/cse/wiki/doku.php?id=obix_server_specification_1.1_wd_06_implementation
http://code.google.com/p/iotsys
https://www.oasis-open.org/committees/obix/ (oBIX)

The IoTSyS gateway provides an oBIX Web service interface to access heterogeneous existing building automation technologies and smart meters (e.g. KNX, BACnet, ZigBee, Wireless M-Bus). The gateway provides a Web service protocol binding to SOAP and RESTful HTTP and CoAP Web service endpoints. Furthermore, virtual IPv6 endpoints are provided for all devices behind the gateway making them globally accessible in the Internet. Within this project state of the art Web service security technologies and access control technologies should be applied to protect these Web service endpoints. Transport layer security should be provided for HTTP (TLS/SSL) and CoAP (DTLS), but also message layer security should be considered. For the SOAP WS-endpoint the WS-Security stack can be used for securing message exchange by applying signatures and encryption. For the RESTful Web service endpoints XML Signature and XML Encryption should be used on the payload. Finally, access control mechanisms should be applied to allow fine-grained access control on oBIX objects for certain clients (e.g. based on XACML). The resulting outcome should be an IoTSyS security OSGI bundle that can be deployed on the gateway and protects the incoming and outgoing requests to the gateway.

Benefit for the Student

Dive into latest technologies and emerging protocols for the Internet of Things/Web of Things. Gain hands on experience in Java based software development and security for Web services. Apply theoretical knowledge on cryptography and security in practice and enhance and tighten your security skills.

Benefit for the Project

Security is currently a strong and open issue for the IoTSyS project which has highest priority. Due to the security and privacy issues arising through offering public access to automation devices a contribution on this topic would be a huge gain for the project.

Requirements

Strong skills in Java based software development are necessary.
Know-how about OSGI, IoC container design and implementation, Java byte code modification frameworks is a plus.

Mentors

Markus Jung, Thomas Hofer

Contact

Mentors are regularly around in our GSoC IRC channel #TU-CSE-SoC at irc.freenode.net. You can also reach us via the mailinglist – send an email to soc2013@iue.tuwien.ac.at using the prefix [IOTSYS] (a subscription is required).

More Information

http://www.iue.tuwien.ac.at/cse/wiki/doku.php?id=security
http://code.google.com/p/iotsys
http://datatracker.ietf.org/doc/draft-ietf-core-coap/ (CoAP)
https://tools.ietf.org/html/rfc6347 (DTLS)
https://www.oasis-open.org/committees/tc_home.php?wg_abbrev=wss (WS-Security)
http://www.w3.org/TR/soap/ (SOAP)
https://www.oasis-open.org/committees/obix/ (oBIX)
https://www.oasis-open.org/committees/tc_home.php?wg_abbrev=xacml (XACML)

The student migrates the current MOST (most.bpi.tuwien.ac.at) MySQL database schema to a relational/ NoSQL hybrid solution using MariaDB and Cassandra. The current implementation heavily depends on stored procedures for data preprocessing. The student implements a Java-based framework for database independent data preprocessing.

Benefit for the Student

The student works with state of the art database technologies and learns how to use highly scalable data storages. Additionally, the work focuses on performance optimization.

Benefit for the Project

This work is a first step towards a plug and play building data collection. Combined with other projects the goal is to realize a cloud-based building monitoring data storage mechanism.

Requirements

Good Java programming skills, MySQL. NoSQL knowledge is not mandatory, but nice to have

Mentors

Harald Hofstätter, Robert Zach, Rainer Bräuer, Stefan Glawischnig

Contact

Mentors are regularly around in our GSoC IRC channel #TU-CSE-SoC at irc.freenode.net. You can also reach us via the mailinglist – send an email to soc2013@iue.tuwien.ac.at using the prefix [MOST] (a subscription is required).

More information

http://www.iue.tuwien.ac.at/cse/wiki/doku.php?id=nosql_solution_for_building_data_warehouses

Image by Oliver Widder, available under a Creative Commons Attribution 3.0 Unported License.

The student implements a Java-based gateway that runs on the Raspberry Pi to enable a plug and play mechanism for building sensors. New sensors are recognized and registered automatically in the MOST framework. The final parameter setting is done via an application, that will either be an Android or cross-platform web application.

Benefit for the Student

The student will participate in a real world home automation scenario. This includes the opportunity to work with technologies form various disciplines (distributed systems, mobile development).

Benefit for the Project

This work is a first step towards a plug and play building sensor registration. Combined with other projects the goal is to realize a cloud-based building monitoring service with minimal configuration efforts.

Requirements

Good Java programming skills, basic knowledge in distributed systems

Mentors

Harald Hofstätter, Robert Zach, Stefan Glawischnig

Contact

Mentors are regularly around in our GSoC IRC channel #TU-CSE-SoC at irc.freenode.net. You can also reach us via the mailinglist – send an email to soc2013@iue.tuwien.ac.at using the prefix [MOST] (a subscription is required).

More information

http://www.iue.tuwien.ac.at/cse/wiki/doku.php?id=raspberry_pi_-_mobile_plug_and_play_home_automation

Description

Currently, nomacs supports simple file and directory access such as read, write, update or thumbnail preview. However, simple batch processing would improve nomacs file management. The task is to implement a GUI for file filtering and batch processing. The filtering should be applicable for file names, extensions and exif tags. In addition a batch editing tool should be implemented that allows for editing exif tags of multiple files, renaming multiple files or converting files to other image formats (e.g. converting multiple files from *.nef to *.jpg)

Benefit for the Student

The student will gain experience in Qt GUI development. Moreover, basic knowledge of image processing libraries such as Exiv2 or OpenCV will be developed.

Benefit for the Project

File handling of nomacs will be improved. Thus, nomacs can be used to manage large amounts of image data.

Requirements

Good skills in object-oriented C++ and Qt are required.

Mentors

crop toolnomacs basically supports cropping. This includes a simple-to-edit rectangle that allows for rotating/cropping the current image. We want to improve the user experience by adding a crop toolbar with tools such as zoom/crop and units preview. In addition a Hough Transform based snapping to lines present in an image will be implemented that allows for an exact cropping of lines. Cropping and rotating images may result in image regions that are not defined. Currently, these image regions are not filled at all (transparent) or they are filled with a default color. Filling these regions with parts that are replicated from the original image improves the visual appearance. The student should investigate state-of-the-art texture synthesis methods such as Kwatra et al. for a best possible visual appearance of cropped images.

Benefit for the Student

The student will gain experience in state-of-the-art texture synthesis algorithms and computer vision. Moreover, basic knowledge of image processing libraries such as OpenCV will be developed.

Benefit for the Project

The user experience of the crop tool will be improved. Additionally, texture synthesis improves the usability of the current crop functionality.

Requirements

Good skills in object-oriented C++ are required accompanied by basic experience in computer vision. Qt or OpenCV knowledge would be beneficial.

Mentors

Florian Kleber, Markus Diem, Stefan Fiel

fake miniatur of bejingThe goal of this project is to design a plug-in system for nomacs. The basic idea of nomacs was to develop a fast and stable image viewer with synchronizing capabilities. On user requests features like supporting HDR images, RAW images, etc. have been implemented to the basic viewer. The plug-in system should allow to extend nomacs with such features while keeping the basic image viewer small. The student should define the interface of the plug-in system and write a small demo plug-in to create fake miniatures of images. A Matlab sample code for creating fake miniatures already exists.

Benefit for the Student

The student will deepen knowledge in software and plug-in architecture. In addition, basic knowledge of image processing will be gained.

Benefit for the Project

The plug-in system will keep a basic image viewer, whereas the user can extend functionalities on his demand.

Requirements

Good skills in C++, Experience in writing plug-ins or designing a plug-in system.

Mentors

Florian Kleber, Markus Diem, Stefan Fiel

Description

Based on the explicit data models of the different heterogeneous tools/data sources, a generic approach for the definition of queries should be implemented. Based on this generic approach, the queries then need to be transformed into a query language the target data source can understand and handle - first of all we target at supporting data sources capable of dealing with the SQL syntax. For testing purposes, data from different projects will be provided.

Benefit for the Student

Learning about the transformation of queries to a set of heterogeneous data sources, and more generally about model-driven systems engineering.

Benefit for the Project

By now, queries to the different heterogeneous data sources/tools integrated need to be written and transformed manually to their local representations. The automation of this definition and transformation process would dramatically increase the usability and acceptance of the framework, since even non-experts would be capable of defining and executing complex queries.

Requirements

Excellent skills in understanding and programming Java code are necessary. Useful would be knowledge about OSGi, database technologies (e.g., SQL), and about semantic web technologies (e.g., OWL).

Mentors

Richard Mordinyi, Andreas Pieber, Dietmar Winkler

Contact

Mentors are regularly around in our GSoC IRC channel #TU-CSE-SoC at irc.freenode.net. You can also reach us via the mailinglist – send an email to soc2013@iue.tuwien.ac.at using the prefix [OPENENGSB] (a subscription is required).

More information

http://www.iue.tuwien.ac.at/cse/wiki/doku.php?id=query_transformation

Description

Software testing is a very important part in the software engineering process, like in the context of remote (distributed) component types in component-based systems which provide/use services over the network via RPC/RMI communication calls or messaging. Based on explicit data models of the different heterogeneous software tools taking part in the integration process, a generic approach for the generation of test cases has to be implemented. Tools connected to the OpenEngSB framework are seen as components for which test cases simulating the tool's behavior have to be derived. There are two distinct ways of simulating a tool’s behavior. In the first one, the tool is generating the test-cases from the specification of the corresponding components. In the second approach, the tool is recording the communication between the components and creates from this network information’s the test-cases.

Benefit for the Student

Learning about component-based systems, testing of such systems, semantic models, and model-driven engineering.

Benefit for the Project

By now, there is no concept enabling a test-driven engineering approach which takes into account existing knowledge about tools, processes, and data. The automatic derivation of test cases using this knowledge would increase the quality of the integration solution.

Requirements

Excellent skills in understanding and programming Java code are necessary. Useful would be knowledge about OSGi, component-based systems, testing approaches, and about semantic web technologies (e.g., OWL).

Mentors

Richard Mordinyi, Andreas Pieber, Dietmar Winkler

Contact

Mentors are regularly around in our GSoC IRC channel #TU-CSE-SoC at irc.freenode.net. You can also reach us via the mailinglist – send an email to soc2013@iue.tuwien.ac.at using the prefix [OPENENGSB] (a subscription is required).

More information

http://www.iue.tuwien.ac.at/cse/wiki/doku.php?id=test_case_generation

The aim of this project is to refine and improve the parallelization of routines used in the {OpenPixi simulator}[http://www.openpixi.org/]. The Particle-In-Cell simulations use a large number of particles within a fixed grid. The grid and the particles are split so that they can run on different cores of a CPU, or even on different computers. The first step of the project, which is written in Java, is to refine and improve the existing parallelization routines that use the IBIS framework. On top of this, an implementation which utilizes OpenCL should be developed in order to make full use of the calculation power of GPUs (graphics processing units) on personal computers, or larger GPU clusters dedicated to large-scale simulations. In any case, automatable test cases must ensure that the results agree with the single threaded version of the code.

Benefit for the student

The student will learn about physics simulations and how to parallelize them for best performance. The student will learn about the IBIS framework and about OpenCL.

Benefit for the project

The project will receive a fast parallel implementation of existing routines.

Requirements

Good skills of Java and C/C++. Knowledge of MPI or OpenCL is of advantage.

Task for the application process

For the application process, please fix one of the issues at {GitHub}[https://github.com/organizations/openpixi/dashboard/issues], or provide a small demo application that uses OpenCL and implements one of the routines of the Java version of OpenPixi in parallel.

Mentors

Andreas Ipp, Ognen Kapetanoski, Kirill Streltsov

"Split" is an algorithm that compresses a given propositional resolution proof by first selecting a literal "L", then splitting the refutation into a proof of "L" and proof of "(not L)", and finally constructing a new refutation by resolving these two proofs. This process can be repeated until no further compression occurs. We have reasons to believe that the current heuristic for selecting the literal "L" is far from optimal. Furthermore, it seems to be possible to improve compression by using recursion instead of mere repetition of the process. 

Benefit for the Student

The student will acquire practical experience and be in touch with cutting-edge research in the fields of automated deduction and applied proof theory. He will be mentioned as a co-author of any paper that might benefit from his implementation. He will have the pleasure of programming in the awesome language Scala.

Benefit for the Project

"Split" and "Reduce&Reconstruct" are among the most compressing algorithms available, but they are far from being the most efficient. If the ideas discussed above are successfully implemented, Skeptik will be able to provide more efficient algorithms to the research community.

Requirements

Basic knowledge of (propositional) logic is required. Basic knowledge of Scala or experience with other object-oriented (e.g. Java, C++,...) and functional (e.g. Haskell, OCaml,...) programming languages and willingness to learn Scala is required. Experience with data structures for proofs or directed acyclic graphs is desirable.

Mentors

Bruno Woltzenlogel Paleo, Joseph Boudou

Contact

Mentors are regularly around in our GSoC IRC channel #TU-CSE-SoC at irc.freenode.net. You can also reach us via the mailinglist – send an email to soc2013@iue.tuwien.ac.at using the prefix [SKEPTIK] (a subscription is required).

More Information

http://www.iue.tuwien.ac.at/cse/wiki/doku.php?id=heuristics_and_improvements_for_split_and_reduce_reconstruct
Instructions on how to improve your chances of getting accepted are listed on Skeptik's wiki (https://github.com/Paradoxika/Skeptik/wiki/GSoC-Instructions). In our issue tracker you can also find other ideas.

There are a few proof compression algorithms that are not yet implemented in Skeptik, even though they have already been described in the literature and implemented elsewhere. We would like these algorithms to implemented, generalized and improved.

One of them is "RecycleUnits" which implements an efficient restricted kind of subsumption in which the subsuming clause is unit. Besides implementing the original "RecycleUnits", we would like to generalize it to larger non-unit subsuming clauses.

Benefit for the Student

The student will acquire practical experience and be in touch with cutting-edge research in the fields of automated deduction and applied proof theory. He will be mentioned as a co-author of any paper that might benefit from his implementation. He will have the pleasure of programming in the awesome language Scala.

Benefit for the Project

If all mentioned algorithms are implemented, Skeptik will contain practically all known propositional resolution proof compression algorithms known to this date.

Requirements

Basic knowledge of (propositional) logic is required. Basic knowledge of Scala or experience with other object-oriented (e.g. Java, C++,...) and functional (e.g. Haskell, OCaml,...) programming languages and willingness to learn Scala is required. Experience with data structures for proofs or directed acyclic graphs is desirable.

Mentors

Bruno Woltzenlogel Paleo, Joseph Boudou

Contact

Mentors are regularly around in our GSoC IRC channel #TU-CSE-SoC at irc.freenode.net. You can also reach us via the mailinglist – send an email to soc2013@iue.tuwien.ac.at using the prefix [SKEPTIK] (a subscription is required).

More Information

http://www.iue.tuwien.ac.at/cse/wiki/doku.php?id=recycleunits_subsumption_and_intermediate_rp-rpi_algorithms
Instructions on how to improve your chances of getting accepted are listed on Skeptik's wiki (https://github.com/Paradoxika/Skeptik/wiki/GSoC-Instructions). In our issue tracker you can also find other ideas.

The C++ Standard Template Library (STL) defines a set of basic algorithms such as sort().
Such functionality is, where reasonable, provided for NVIDIA GPUs with thrust, but there is no extensive and independent implementation available in OpenCL. Thus, the goal of this project is to implement some of the most popular algorithms from the C++ STL for GPUs in ViennaCL.

Benefit for the Student

Fame and glory! :-) Invaluable experience in programming a robust high-level interface for targeting CPUs and GPUs on the low level will be gained.

Benefit for the Project

Providing efficient STL-type functionality will enable a lot of additional algorithms with portable performance to be added to ViennaCL in the future. In addition, users can easily compose operations not natively provided by the library.

Requirements

Advanced knowledge of C++ is essential, including familiarity with meta-programming techniques. Experience with GPU programming is a plus.

Mentors

Karl Rupp, Philippe Tillet

Contact

Mentors are regularly around in our GSoC IRC channel #TU-CSE-SoC at irc.freenode.net. You can also reach us via the mailinglist – send an email to soc2013@iue.tuwien.ac.at using the prefix [VIENNACL] (a subscription is required).

More information

http://www.iue.tuwien.ac.at/cse/wiki/doku.php?id=gpu-accelerated_stl_functionality

Current vendor libraries for basic linear algebra operations (BLAS) use a C-style function interface. Even though a high-level approach such as ViennaCL requires much clearer and less error-prone code, the availability of a C-style interface for the various operations will aid potential users who are required to stick with C or Fortran. Also, this would be the first BLAS library with portable high performance on hardware from different vendors.

Benefit for the Student

The student will learn about the challenges of GPU and accelerators programming and how to write and maintain shared libraries for multiple platforms.

Benefit for the Project

ViennaCL will be the first project to offer a low-level BLAS library for GPUs and accelerators explicitly targeting multiple vendors and providing high performance on each of them.

Requirements

A solid understanding of C and C++ is required. Ideally, the student has some experience with CMake and shared libraries on multiple platforms, particularly Windows and Linux.

Mentors

Karl Rupp, Philippe Tillet

Contact

Mentors are regularly around in our GSoC IRC channel #TU-CSE-SoC at irc.freenode.net. You can also reach us via the mailinglist – send an email to soc2013@iue.tuwien.ac.at using the prefix [VIENNACL] (a subscription is required).

More information

http://www.iue.tuwien.ac.at/cse/wiki/doku.php?id=low-level_c_interface

ViennaCL currently allows users to conveniently code linear algebra operations for GPU inside C++. On the other hand, the script language Python is very popular for high-level programming approaches.
The aim of the project is to bridge the gap between these two approaches, making the functionality in ViennaCL also available within Python.

Benefit for the Student

The student will work on the boundary between C++ and Python, thus learning a lot about the strengths and limits of both languages.

Benefit for the Project

A Python wrapper will make the rich functionality in ViennaCL available to a wide range of scientists using Python for their everyday projects.

Requirements

A solid understanding of both C++ and Python is required. Ideally, the student has some experience with working with shared libraries on multiple platforms, particularly Windows and Linux.

Mentors

Karl Rupp, Philippe Tillet

Contact

Mentors are regularly around in our GSoC IRC channel #TU-CSE-SoC at irc.freenode.net. You can also reach us via the mailinglist – send an email to soc2013@iue.tuwien.ac.at using the prefix [VIENNACL] (a subscription is required).

More information

http://www.iue.tuwien.ac.at/cse/wiki/doku.php?id=python_wrapper

GPUs are able to provide high performance for the operation C = A * B for dense matrices A, B, and C. However, if A and B are both sparse matrices, a lot of additional tricks are required to get reasonable performance. The aim of this project is to implement a toolkit of algorithms analyzing the sparsity patterns, which are then composed to yield a fast sparse matrix-matrix multiplication.
Moreover, the implementations should be tuned to GPUs from NVIDIA and AMD as well as Intel's MIC platform.

Benefit for the Student

The student will get hands-on experience in GPU programming using both OpenCL and CUDA. In particular, the student will learn the various tricks required to obtain high performance.

Benefit for the Project

The sparse matrix-matrix multiplication is a key building block for algebraic multigrid solvers and preconditioners. A fast sparse matrix-matrix multiplication will directly improve the efficiency of such methods significantly.

Requirements

Experience in either OpenCL or CUDA is desired.

Mentors

Karl Rupp, Philippe Tillet

Contact

Mentors are regularly around in our GSoC IRC channel #TU-CSE-SoC at irc.freenode.net. You can also reach us via the mailinglist – send an email to soc2013@iue.tuwien.ac.at using the prefix [VIENNACL] (a subscription is required).

More information

http://www.iue.tuwien.ac.at/cse/wiki/doku.php?id=sparse_matrix-matrix_multiplication

Simulation meshes generated by CAD application often are of bad quality, sometimes even broken. For example the hull of an object isn't fully closed, duplicate elements exist or elements intersect each other. All of those (and many more) errorsare fatal for most meshing algorithms and following simulations.

The student should check, if there is any open source library, capable of fixing broken meshes. If available, an interface to ViennaMesh should be written. Otherwise a ViennaMesh module should be implemented which finds defects and repairs them if possible.

ViennaMesh is currently completly reimplemented from scratch. The current source code available on sourceforge does not reflect the current state of the project. For more information please contact the mentors via IRC or the mailinglist soc2013@iue.tuwien.ac.at.

Benefit for the Student

The student will get in touch with the challenging topic of 3D meshing and will learn basic mesh element operations.

Benefit for the Project

The Mesh Doctor will allow ViennaMesh to detect and handle broken meshes instead of just failing to mesh them.

Requirements

The student has to provide good skills in C++ and basic knowledge in geometry.

Mentors

Florian Rudolf, Dieter Pahr, Josef Weinbub

Contact

Mentors are regularly around in our GSoC IRC channel #TU-CSE-SoC at irc.freenode.net. You can also reach us via the mailinglist – send an email to soc2013@iue.tuwien.ac.at using the prefix [VIENNAMESH] (a subscription is required).

More information

http://www.iue.tuwien.ac.at/cse/wiki/doku.php?id=mesh_doctor

In scientific computing, many mesh related algorithms have a great need for meshes with high quality. Although the type of quality of a mesh depends on the algorithm operating on it, most requirements are similar. Especially size and shape of mesh elements is a central topic when speaking of mesh quality. Mesh generators often struggle with constraints and tend to produce low quality mesh elements. To compensate this, mesh optimizer are invented. These algorithms take a mesh and try to optimize the quality of this mesh. Popular mesh optimizers like Stellar or Mesquite were implemented.

The goal is to write an interface for such libraries to ViennaMesh. Additionally some low-level optimizations based on the ViennaGrid data structure might be implemented.

ViennaMesh is currently completly reimplemented from scratch. The current source code available on sourceforge does not reflect the current state of the project. For more information please contact the mentors via IRC or the mailinglist soc2013@iue.tuwien.ac.at.

Benefit for the Student

The student will get in touch with the challenging topic of 3D meshing and will learn the importance of good mesh quality.

Benefit for the Project

With these mesh optimization modules, ViennaMesh is able to perform basic automatic mesh generation and produce meshes with high quality.

Requirements

The student has to provide good skills in C++ and interest in geometry and computer science, mathematical and geometric knowledge is advantageous.

Mentors

Florian Rudolf, Dieter Pahr, Josef Weinbub

Contact

Mentors are regularly around in our GSoC IRC channel #TU-CSE-SoC at irc.freenode.net. You can also reach us via the mailinglist – send an email to soc2013@iue.tuwien.ac.at using the prefix [VIENNAMESH] (a subscription is required).

More information

http://www.iue.tuwien.ac.at/cse/wiki/doku.php?id=mesh_optimizing

ViennaMesh provides a solid C++ interface for meshing 2D and 3D geometries. Using C++ templates, the runtime of ViennaMesh is reduced while holding a good level of flexibility. But C++ templates have an unpopular side effect: hight compilation times. To compensate this problem and allow easy-to-use mesh scripting, a Python interface should be implemented which wraps the ViennaMesh interface.

ViennaMesh is currently completly reimplemented from scratch. The current source code available on sourceforge does not reflect the current state of the project. For more information please contact the mentors via IRC or the mailinglist soc2013@iue.tuwien.ac.at.

Benefit for the Student

The student will work on the boundary between C++ and Python, thus learning a lot about the strengths and limits of both languages.

Benefit for the Project

A Python wrapper will allow easy, flexible and dynamic mesh scripting without long compilation times.

Requirements

The student has to provide good skills in both C++ and Python. Ideally, the student has some experience with working with shared libraries on multiple platforms, particularly Windows and Linux.

Mentors

Florian Rudolf, Dieter Pahr, Josef Weinbub

Contact

Mentors are regularly around in our GSoC IRC channel #TU-CSE-SoC at irc.freenode.net. You can also reach us via the mailinglist – send an email to soc2013@iue.tuwien.ac.at using the prefix [VIENNAMESH] (a subscription is required).

More information

http://www.iue.tuwien.ac.at/cse/wiki/doku.php?id=python_interface

A vital part of computational science is the step of visualization. In this project, the student should extend the current visualization layer based on the VTK library. The goal is to utilize various visualization algorithms to provide advanced visualization capabilities to be used by ViennaMOS to visualize the simulation results. Different algorithms can be investigated based on the students interests, such as the visualization of iso-surfaces and vector fields, clipping, slicing, block extraction, or streamlines.

ViennaMOS is currently completly reimplemented from scratch. The current source code available on sourceforge does not reflect the current state of the project. For more information please contact the mentors via IRC or the mailinglist soc2013@iue.tuwien.ac.at.

Benefit for the Student

The student will obtain skills in 3D visualization and basic skills in GUI programming. More concretly, the student will work with the well-known open source libraries VTK and Qt.

Benefit for the Project

Due to the student's work, the project's visualization capabilities will be advanced considerably, allowing for improved capabilities to interpret the simulation results generated by ViennaMOS.

Requirements

Good C++ skills are required as well as an interest in scientific visualization and meshes.

Mentors

Josef Weinbub, Florian Rudolf

Contact

Mentors are regularly around in our GSoC IRC channel #TU-CSE-SoC at irc.freenode.net. You can also reach us via the mailinglist – send an email to soc2013@iue.tuwien.ac.at using the prefix [VIENNAMOS] (a subscription is required).

More information

http://www.iue.tuwien.ac.at/cse/wiki/doku.php?id=3d_visualization

Almost every data set include missing information but methods for analyzing data almost always needs complete information.

Visualization of incomplete data allows to simultaneously explore the data
and the structure of missing values. This is helpful for learning about the distribution
of the incomplete information in the data, and to identify possible structures of the
missing values and their relation to the available information that can be used for better imputation of missing values.
Imputation of incomplete data is necessary to obtain a complete data set. Imputation methods varies from distance-based methods to model-based iterative procedures.

The R package VIM already include basic tools for visualization and imputation of missing values and it also provides a point and click graphical user interface.

The aim of the GSoC project is to
- integrate a selection of available imputation methods from R in the VIM system
- the tcltk-based graphical user interface should be replaced by a graphical user interface based on Gtk2.
- to allow to work with complex survey objects from the survey package and to generate survey objects. In addition basic operations from the survey package should be accessed through the GUI

Benefit for the Student

The student will get deep knowledge in programming in R and Gtk2 but also basic knowledge in complex sampling designs.
He will understand the needs of imputation in data from complex survey designs.

Requirements

Basic knowledge in complex survey methodology. Advanced knowledge in any scripting language and at least basic knowledge in R.

Mentors

Matthias Templ, Peter Filzmoser

Contact

Mentors are regularly around in our GSoC IRC channel #TU-CSE-SoC at irc.freenode.net. You can also reach us via the mailinglist – send an email to soc2013@iue.tuwien.ac.at using the prefix [VIM] (a subscription is required).

More information

http://www.iue.tuwien.ac.at/cse/wiki/doku.php?id=integration_of_survey_design_facilities_into_a_gtk2-based_r-package_vim