Mark Sagi, M.Sc.

Wissenschaftlicher Mitarbeiter  

Technische Universität München
Fakultät für Elektrotechnik und Informationstechnik
Lehrstuhl für Integrierte Systeme
Arcisstr. 21
80290 München

Tel.: +49.89.289.25287
Fax: +49.89.289.28323
Gebäude: N1 (Theresienstr. 90)
Raum: N2117
Email: mark.sagi@tum.de

Curriculum Vitae

Before joining the Chair for Integrated Systems in 2016, I worked as patent advisor at Leonhard & Collegen. My master's thesis was about "Joint Channel Coding in Heterogeneous Radio Environments" in cooperation with Intel in 2015. Furthermore, I worked as student at Rohde & Schwarz, Infineon and ESG (BMW) and tutored in "adveisor", math, programming in C and a practical course on electrical engineering. I also studied for a semester at ETH Zürich.


I speak fluently german, english and hungarian.

Angebotene Arbeiten

BAMAIDPFPIPHSSHK
Titel
------

Multicore Power Proxies and Models

Multicore Power Proxies and Models

Stichworte:
multicore, processor, modeling, power proxy, linear regression

Beschreibung

To optimize the power consumption of multicore processors, accurate power infrormation is needed. Such information can be either directly measured (sensors) or indirectly determined through so called power proxies. With sensors being very area expensive, power proxies are commonly used. Such power proxies combine a design-time power model with run-time activity information, e.g. performance counters. In this seminar, you will learn about state-of-the-art power proxies and identify the currenty challenges in using power proxies.

Kontakt

mark.sagi@tum.de

Betreuer:

------

Mult-Agent Reinforcement Learning for Multicore Processors

Mult-Agent Reinforcement Learning for Multicore Processors

Stichworte:
Machine learning, multi-agent reinforcement learning, multicore, power, temperature

Beschreibung

Reducing the power consumption of multicore processors is an ongoing challenge for industry and academia alike. Many different machine learning algorithms (Reinforcement learning (RL), supervised learning, unsupervised learning) have been proposed to manage power and performance of multicore processors. However, the interaction of multiple reinforcement learning algorithms running in parallel is an open research area. In this seminar, you will identify state-of-the-art reinforcement algorithms for multicore power management and investigate if there are any methods for Multi-Agent-Reinforcement-Learning in use today.

Kontakt

mark.sagi@tum.de

Betreuer:

Laufende Arbeiten

Hauptseminare

Representation Learning for Multicore Power/Thermo Features

Representation Learning for Multicore Power/Thermo Features

Stichworte:
Machine learning, representation learning, multicore, power, temperature

Beschreibung

 

Reducing the power consumption of multicore processors is an ongoing and challenging task for processor designers. With increasing transistor count, power and thermal information is increasingly difficult to obtain. To obtain more and reliabe power/thermo information, designers are starting to use novel machine learning algorithms. In this seminar, you will investigate different representation learning algorithms – a subset of machine learning algorithms – for identifying power/thermo features on chip. You will make an overview over related work for multicore reprenstation learning and finally make an educated guess which representation learning algorithms are best suited for identiyfing power/thermo features.

 

Kontakt

mark.sagi@tum.de

Betreuer:

Forschung

My research focus is on power management for multi-core processors which are constrained by dark silicon.
Due to increasing power density, a decreasing ratio of transistors can be powered on simultaneously.
At 16nm up to 46% of a chip might need to be powered down, i.e. stay dark.

The dark silicon trend further increases the importance of pioneering research in power management.
Our goal is to research and build novel power management systems which will ensure that Moore's law can live on.

Motivational literature:
Dark Silicon and the End of Multicore Scaling
Intels tick tock seemingly dead becomes process architecture optimization

Publikationen

  • Santiago Pagani, Lars Bauer, Qingqing Chen, Elisabeth Glocker, Frank Hannig, Andreas Herkersdorf, Heba Khdr, Anuj Pathania, Ulf Schlichtmann, Doris Schmitt-Landsiedel, Mark Sagi, Éricles Sousa, Philipp Wagner, Volker Wenzel, Thomas Wild, Jörg Henkel: Dark silicon management: an integrated and coordinated cross-layer approach. it - Information Technology 58 (6), 2016, 297–307 mehr… BibTeX Volltext ( DOI )