Bachelor and Master Theses
To apply for conducting this thesis, please contact the thesis supervisor(s).
| Title: | Analyzing the latency overheads of acceleration in heterogeneous systems |
| Subject: | Computer science, Embedded systems, Dependable Aeronautics and aerospace |
| Level: | Advanced |
| Description: |
Acceleration is the process of implementing part of a software as hardware that can be synthesized on an FPGA. Running a part of a system on the CPU and another part on the FPGA has some latency overhead due to the required data transfer. The operating system which is handling the software execution and the data transfer will also impose some latency overhead. This thesis aims at analyzing the latency overhead of accelerating various neural networks. The acceleration platform and procedure exist and are ready to use. This is an industrial thesis in collaboration with the ENEA company (https://www.enea.com/). This project can be done by one or two students. Supervisor(s): Mohammad Riazati Examiner: Masoud Daneshtalab |
| Start date: | |
| End date: | |
| Prerequisites: | This thesis is primarily intended for Advanced Level student(s) with good knowledge on C/C++, VHDL, FPGA design. Intermediate knowledge of Neural Network architecture would be a plus. |
| IDT supervisors: | |
| Examiner: | |
| Comments: | |
| Company contact: |