akademischer Mitarbeiter
Hector Munoz Hernandez
VG 1C, Room 1.14
T: +49 (0) 355 69 2877
F: +49 (0) 355 69 2027
hector.munozhernandez(at)b-tu.de
- (2018-ongoing) phD in Embedded Systems/Machine learning. Brandenburg University of Technology Cottbus-Senftenberg, Cottbus - Germany,
- (2016–2018) M. Eng. Embedded Systems for Mechatroncs. Fachhochschule Dortmund, Germany.
- (2010–2014) B.Sc Electronics and telecommunications engineering. Tecnológico de Monterrey Campus Ciudad de México, Mexico city.
Scientific Interests:
- Acceleration and implementation of Machine Learning algorithms
- Hardware acceleration on FPGAs and GPUs
- Digital Design and Computer Architecture
- Image processing applications
Topics for Student Thesis:
Topic 1 - ISA expansion for a soft-core GPU
The FGPU (a soft GPU core that can be implemented in an FPGA) has a working (but limited) OpenCL support, which means that not every instruction can be understood by the FGPU. The objective of this project is to modify the cutsom compiler (LLVM) for the FGPU (a soft-core GPU on an FPGA) in order to add more openCL compatibility.
This thesis is targeted for master students.
As its a complex topic, the student should have previous knowledge of how compilers work. Knowledge of OpenCL, GPUs, and/or LLVM is a plus.