direkt zum Inhalt springen

direkt zum Hauptnavigationsmenü

Sie sind hier

TU Berlin

Inhalt des Dokuments

Michael Andersch

Hi! I am a student researcher at the department of embedded systems architectures and participating in several research projects. When I'm not actively researching, you will most often find me in lectures and courses as part of the master's programme in computer engineering.

Between 2010 and 2012, I have actively contributed to the ENCORE project. My main objective has been the investigation of the novel task-based parallel OpenMP Superscalar (OmpSs) programming model with respect to both programmability and application performance. Since 2012, I have been involved with the LPGPU project, where me and my colleagues have developed a cycle-accurate GPU power simulator to conduct research for novel power-efficient GPU architectures that are able to execute control-divergent and latency-limited codes with high performance.

My very own wide-ranging (research) interests include processor architecture, parallel programming for high performance, digital image processing (including computer vision), real-time and photo-realistic computer graphics and signal processing techniques for source coding.

I am going to complete my studies in computer engineering in October 2014. For the time beyond, I have taken a position as a GPU Architect with the NVIDIA Corporation.


On Latency in GPU Throughput Microarchitectures
Citation key andersch_gpu_latency
Author Michael Andersch and Jan Lucas and Mauricio Alvarez-Mesa and Ben Juurlink
Title of Book Proc. IEEE Int. Symposium on Performance Analysis of Systems and Software (ISPASS)
Pages 169-170
Year 2015
DOI 10.1109/ISPASS.2015.7095801
Month March
Download Bibtex entry

Zusatzinformationen / Extras

Quick Access:

Schnellnavigation zur Seite über Nummerneingabe