01 April 2025

On March 3, Mina Niknafs at the Department of Computer and Information Science (IDA) successfully defended her thesis on resource management for multi-core computer systems. "Pursuing a PhD at IDA is an enriching experience that combines academic freedom with a supportive research culture", she says.

Mina Niknafs shows her printed thesis in the lecture hall.
Mina Niknafs

Congratulations on your PhD! What was your background when you started your PhD?

"Thank you! I earned my Master's in IT approximately six years before beginning my PhD. During this period, I worked both in industry as a developer and in academia as a lecturer and faculty member, acquiring valuable experience that connected practical application with theoretical foundations."

What is it like to be a PhD student at IDA?

"Pursuing a PhD at IDA is an enriching experience that combines academic freedom with a supportive research culture. While the journey involves inevitable hurdles and uncertainties, the presence of dedicated supervisors and a collaborative community makes a huge difference. I was fortunate to have Zebo Peng and Petru Eles as my supervisors, whose invaluable feedback and encouragement kept me on track throughout my studies."

What do you want to do next?

"With my PhD completed, I aim to move into the industry, where I can bridge my research background with real-world applications."

Mina Niknafs' summary of her thesis:

In modern embedded heterogeneous architectures, multiple processing resources are often integrated to achieve performance benefits. Different tasks operate on various segments of these systems, resulting in variations in timing, temperature, and energy consumption. A key component for such a system is a resource manager, RM, capable of taking decisions such that resources are used most efficiently and performance requirements are satisfied. Moreover, modern platforms often encounter fluctuating workloads that are not known at design time. This uncertainty makes traditional design-time resource management methods impractical. In such scenarios, runtime resource management decisions can be enhanced by not only considering the current workload of the platform but also by incorporating forecasts of future workloads. From a macro perspective, this thesis focuses on (1) predicting workload, and (2) integrating these predictions into resource management, while accounting for the constraints and features of embedded systems.

More about the subject and doctoral studies

Organisation

Latest news from LiU

Researcher with blue gloves by microscope.

Plastic nerve cells become more advanced – and simpler

An artificial neuron made of conductive plastics that can perform advanced functions similar to those of biological nerve cells has been demonstrated by researchers at LiU.

Headshot of  woman.

Ancient lentils shed light over Canary Islands history

The very first genetic study of archaeological lentils show that lentils grown in the Canary Islands have a history that stretches back almost 2,000 years. These lentils may become valuable for plant breeding in the light of ongoing climate change.

picture of Jan-Ingvar Jönsson

LiU activates crisis management following bus accident on the E4

LiU has activated its crisis management team due to a bus accident on the E4 motorway. It involves a campus bus between LiU’s three campuses in Norrköping and Linköping. "Important to take care of one another", says vice-chancellor Jan-Ingvar Jönsson