Print Print Send link Bookmark and Share

Computing platforms

Competence

Computing platforms are the corner stones of computer-based systems and modern information society. VTT has a long research background in processor and computer architectures originating from hardware (ASIC) design, embedded system design, and computer science. The main competences are related to parallel computing and especially parallel processor architectures, multicore solutions based on integrated System on Chip and Network on Chip, and distributed and ubiquitous computing paradigms, performance and quality evaluation of integrated computing platforms, and to capability to evaluate, demonstrate and prototype different types of platforms including support for parallel programming languages and compilers. From application domains we have very close relationship to communication systems and products, and processor architectures needed in their efficient realisations.

Challenges

The main technical challenge in computing platforms research is identifying the right organisation and composition of architectural techniques supporting efficient management of complexity, energy efficiency, interoperability, programmability, and performance issues. All these are also related to cost management due to the increased capabilities offered by technology development and increased pressures from markets in the form of an increased number of product variations and need for still shorter product life cycles.

The computing platforms are facing four major disruptive changes due to (or forced by) technology developed.

  • High performance computing will be done by parallel computers and parallel programs because the fundamental limits of sequential computing in terms of clock speed have been reached. Increasing performance is no longer possible by increasing the speed of electronic circuitry.
  • The shift to parallel platforms has an irreversible impact on programming paradigms, which need to go parallel as well. Architectural support for easy-to-use and easy-to-migrate parallel computing paradigms will be a key to their successful exploitation in computers and electronics products.
  • The complexity of systems and energy-efficiency requirements are forcing the computing platform architectures to more modular and specialised solutions. Approaches that combine specialised subsystems allow the distribution of R&D work more efficiently and lead to better innovation potential.
  • Fourthly, the cost of putting at least simple computers that have wireless communication capabilities has decreased to a level that allows the emergence of ubiquitous intelligence as soon as interoperability and business model related problems are solved. Applications, services and systems that are based on separate devices and their capabilities in our environment will bring a new dimension to information society

Solutions

We can provide solutions such as:

  • heterogeneous multicore, complex memory organisations and reconfiguration based platforms
  • embedded multicore and parallel processors
  • dynamically reconfigurable electronics
  • parallel processing based system and application SW
  • open source components based SW platforms, virtual computing platforms
  • open interfaces inside subsystems, services, and applications
  • interoperability in component, service and information level
  • system co-modelling, evaluation and synthesis
  • general-purpose computing architectures realising advanced parallel computing paradigms
  • easy-to-use and low-overhead parallel languages and compiling tools.

Benefits

Our solutions are applicable for a wide range of products and systems due their general propose nature. The main benefits are increased levels of:

  • scalability
  • flexibility
  • interoperability
  • modularity
  • security
  • energy-aware performance
  • cost efficiency
  • design productivity.

References and merits

  • software radio principle based multi-standard radio design and prototyping
  • efficient parallel processor architecture for telecommunication applications
  • evaluation methods and tool for mobile phone implementation architecture performance monitoring and analysis
  • participation in NoTA development and the NoTA community
  • base station and microwave radio implementation architectures
  • DSP and Telecom SW and HW for 3rd generation base stations
  • 4th generation radio modem prototype
  • WiMAX radio and network HW and SW design and prototyping
  • general-purpose CMP architecture realizing the PRAM model
  • parallel programming language with programming tools.


Additional information

Markus Taumberger
Research Team Leader
+358 20 722 2358

Juha-Pekka Soininen
Research Professor
+358 20 722 2280

Additional information

Markus Taumberger
Research Team Leader
+358 20 722 2358

Juha-Pekka Soininen
Research Professor
+358 20 722 2280

See also