MSc projects

The following lists possible MSc thesis projects: (more coming - check back regularly)

  • Implement MOLEN co-processors: The MOLEN processor comprises a general-purpose processor with a reconfigurable hardware unit that functions as a co-processor to speed up the processing of special kernels of applications running on the general-purpose processor. In the past, we have worked on many multimedia and network processing related kernels. We would like to extend this set of accelerators to cover more applications. In this project, we will determine the most interesting application at the time and define a project to build dedicated MOLEN co-processors.
  • Extend the functionality of the MOLEN prototype The MOLEN prototype is currently implemented on the Virtex II-Pro and Virtex4 platforms. However, it is still an ongoing project to further extend the project to make the prototype more functional and accessible. In this project, we will define the needed functionality we need to extend the protoype with and work to achieve it.
  • Extending the ρ-VEX processor: The ρ-VEX processor can be stand-alone extensible VLIW processor or attached to the MOLEN as a co-processor. In either case, the ρ-VEX processor is still in its early stages and its functionality and performance needs to be further improved. (Check out this article for more details). In this project, we will define the precise work on how to improve the ρ-VEX processor and work to achieve this.
  • Implement the MOLEN concept on an AMD Opteron-Hyper transport platform: Reconfigurable hardware in increasingly moving closer to the general-purpose processor and that is the reason why reconfigurable hardware nowadays can be found connected to the Hypertransport bus attaching it closely to the AMD Opteron processor. In this project, we will implement the MOLEN concepts on the Bypertransport.
  • Implement grid computing hardware functionalities: In collaborative grid computing, many computing nodes work together to achieve a common goal (e.g., large-scale scientific computing). It is expected that these nodes will contain reconfigurable hardare as they are getting increasingly more performant. In this project, we will work on either specific application acceleration on reconfigurable grid nodes or design network-related hardware functionalities to improve communication between the reconfigurable collaborative computing nodes.
  • Implement OS scheduling techniques for reconfigurable computing: (description coming)
  • Investigate interruptible hardware on reconfigurable computing platforms: (description coming)
  • Investigate dynamic VLIW code scheduling: (description coming)

Email me if you are interested. The descriptions are purposely broad so students are giving freedom to find interesting topics (software, hardware, or both combined) within the described projects to work on. NOTE: Guidance of students outside the TU Delft is not possible.

Stephan Wong
Stephan Wong (Associate Professor)