Optimization of Image processing Algorithms for TI DSP

Lindau, Bodensee  ‐ Vor Ort
Dieses Projekt ist archiviert und leider nicht (mehr) aktiv.
Sie finden vakante Projekte hier in unserer Projektbörse.

Schlagworte

Beschreibung

1) Algorithm profiling and runtime analysis of image processing algorithms
* Setup of evaluation board simulation environment
* Setup of bitexactness tests between target and reference implementation
* Setup of runtime instrumentation for runtime profiling

2) Algorithm optimization with respect to runtime and memory useage
* Proposal of optimization concept based on runtime profiling analysis
* identification of code to be optimized for DSP core
* identification of code to be optimized for EVE accelerator
* identification of code to be optimized for ARM core
* estimation of runtime reduction for proposed optimization concept
* implementation of code optimizations
* verification of achieved runtime target
* verification of functional correctness of algorithms using bitexactness tests against reference implementation

Required background :

* proficient in C, C++ programming
* good teamworking abilities
* language skills : good command of german and english (esp. for coordinator)
* basic knowledge of image / signal processing
* experience with runtime and memory optimization of C/C++ code
* knowledge of SW engineering process
* experience with version management systems

Desired background :

* embedded experience, ideally with multi-core architectures
* experience with DSP optimization, ideally for TI DSPs and ARM cores
* experience with algorithm optimization for vector engines, ideally for EVE
* experience with embedded operating systems
* knowledge of Conti simulation environment

Location: Onsite, Back-Office

Schlagwörter:
image processing, algorithm, DSP (TI), C, C++, ARM


Kontaktperson: Niederlassungsleiter Herr Robert Willems


FERCHAU Engineering GmbH
Niederlassung Friedrichshafen
Otto-Lilienthal-Straße 6
88046 Friedrichshafen
Start
01.10.2013
Dauer
30 Wochen
Von
FERCHAU GmbH
Eingestellt
16.09.2013
Ansprechpartner:
Daniel Gogel
Projekt-ID:
599462
Vertragsart
Freiberuflich
Um sich auf dieses Projekt zu bewerben müssen Sie sich einloggen.
Registrieren