Architektur eingebetteter Systeme

Dr.-Ing. Daniel Maier

Forschung

  • Approximate computing
  • Software and compiler optimizations

Master/Bachelor Projects

Completed Projects

Memory-aware Weight Pruning for Deep Neural Networks
Thomas Hartenstein (results published)

Automatic Transformations for Local Memory Usage in OpenCL Kernels
Stefan Gebauer

Automation of the optical beamline and integration with the network
Keerthana Ranganathan (Helmholtz-Zentrum Berlin)

Effects of Approximate Computing on Workload Character
Stefan Schirmeister (results published)

The Impact of Reconstruction in Approximate Computing
Jahn Michel

Hardware Software Co−Design of Neural Network Accelerators
Muhammad Usama Javed (Bosch)

Preise & Auszeichnungen

  • Outstanding Paper Award Runner Up auf der International Conference on High Performance Computing & Simulation (HPCS 2019) für das Paper "Approximating Memory-bound Applications on Mobile GPUs"
  • Best Paper Award auf der Eighth EAI International Conference on Simulation Tools and Techniques (SIMUTools 2015) für das Paper "Deterministic Models of the Physical Layer through Signal Simulation"

Publikationen

2023

Maier, Daniel
Approximate computing by perforation and reconstruction on modern multicore and massively parallel architectures
Dissertation
Dissertation
Berlin
2023

2022

Maier, Daniel; Schirmeister, Stefan; Juurlink, Ben
Effects of Approximate Computing on Workload Characteristics
In Schulz, Martin and Trinitis, Carsten and Papadopoulou, Nikela and Pionteck, Thilo, Editor, Architecture of Computing Systems, Seite 85–99
In Schulz, Martin and Trinitis, Carsten and Papadopoulou, Nikela and Pionteck, Thilo, Editor
Herausgeber: Springer International Publishing, Cham
2022
ISBN
978-3-031-21867-5
Maier, Daniel; Juurlink, Ben
Model-Based Loop Perforation
Euro-Par 2021: Parallel Processing Workshops, Seite 549–554
Herausgeber: Springer International Publishing, Cham
2022
ISBN
978-3-031-06156-1

2021

Maier, Daniel; Cosenza, Biagio; Juurlink, Ben
ALONA: Automatic Loop Nest Approximation with Reconstruction and Space Pruning
In Leonel Sousa and Nuno Roma and Pedro Tomás, Editor, Euro-Par 2021: Parallel Processing, Seite 3–18
In Leonel Sousa and Nuno Roma and Pedro Tomás, Editor
Herausgeber: Springer International Publishing, Cham
2021
ISBN
978-3-030-85665-6

2020

Loni, Mohammad; Zoljodi, Ali; Maier, Daniel; Majd, Amin; Daneshtalab, Masoud; Sjödin, Mikael; Juurlink, Ben; Akbari, Reza
DenseDisp: Resource-Aware Disparity Map Estimation by Compressing Siamese Neural Architecture
IEEE World Congress On Computational Intelligence (WCCI) 2020
Juli 2020

2019

Maier, Daniel; Mammeri, Nadjib; Cosenza, Biagio; Juurlink, Ben
Approximating Memory-bound Applications on Mobile GPUs
2019 International Conference on High Performance Computing & Simulation (HPCS)
IEEE
2019
Hartenstein, Thomas; Maier, Daniel; Cosenza, Biagio; Juurlink, Ben
Memory-aware Weight Pruning for Deep Neural Networks.
PARS-Mitteilungen, (to appear)
2019

2018

Maier, Daniel; Cosenza, Biagio; Juurlink, Ben
Local Memory-Aware Kernel Perforation
Proceedings of the 2018 International Symposium on Code Generation and OptimizationausCGO '18
Herausgeber: ACM
2018

2015

Maier, Daniel; Moser, Steffen; Slomka, Frank
Deterministic Models of the Physical Layer Through Signal Simulation
Proceedings of the 8th International Conference on Simulation Tools and TechniquesausSIMUTools '15, Seite 175–182
Herausgeber: ICST (Institute for Computer Sciences, Social-Informatics and Telecommunications Engineering), ICST, Brussels, Belgium, Belgium
2015
ISBN
978-1-63190-079-2

2014

Maier, Daniel
Konzeption und Entwurf einer Protokoll-/Signal-Co-Simulation für Ad-Hoc-Netzwerke
Diplomarbeit
Diplomarbeit
Universität Ulm
2014