TU Delft
Year
print this page print this page     
NEDERLANDSENGLISH
Organization
2017/2018 Electrical Engineering, Mathematics and Computer Science Master Electrical Engineering
ET4370
Reconfigurable Computing Design
ECTS: 5
Responsible Instructor
Name E-mail
Dr. G.K. Kuzmanov    G.K.Kuzmanov@tudelft.nl
Dr.ir. J.S.S.M. Wong    J.S.S.M.Wong@tudelft.nl
Contact Hours / Week x/x/x/x
0/2/0/0 and 4 hours lab per week
Education Period
2
Start Education
2
Exam Period
2
3
Course Language
English
Expected prior knowledge
System Design with HDLs (ET 4272) or similar
Computer Arithmetic (ET 8019) or similar
Note, starting 2009/10, ET8019 has been replaced by two courses:
ET4170 Computer Arithmetic and
ET4171 Processor Design Project
Course Contents
This course introduces the students to the field of Reconfigurable Computing. The course will provide a balanced insight of both theoretical trends and practical hands-on experience with reconfigurable technology. In addition to the basic concepts, the students are taught advanced topics in reconfigurable technologies, organizations, architectures, programming paradigms, design tools and runtime systems. The particular topics to be covered are, but not limited to: fine and coarse grain reconfigurable technologies, partial and runtime reconfiguration, context switching, runtime algorithms for reconfigurable resource management, tools and methods for system-level and RTL synthesis, reconfigurable-specific compiler optimizations. Special attention is paid on application-specific acceleration in the context of reconfigurable technology. For example, the following application domains will be given special attention: image processing, streaming applications, encryption, compression, bioinformatics, supercomputing, etc.
Study Goals
- The student understands the basics of reconfigurable technology;
- The student is able to design efficiently for reconfigurable technology (in terms of performance and resource utilization);
- The student can identify computationally intensive parts of applications that are suitable for acceleration using reconfigurable computing;
- The student masters state-of-the-art EDA tools for reconfigurable computing;
- The student will be able to effectively use runtime and partial reconfiguration;
- The student will be able to approach the design problems using a top-down methodology (starting from an application, to an architecture, down to a (re)configuration bitstream);
Education Method
Lectures and laboratory work
Literature and Study Materials
Handouts and research papers
Assessment
Open questions written test (50%) + Lab work (50%)