Andy Pimentel, chair of the group: ‘Our research in the combination of extra-functional behaviour and parallel systems on one chip or one machine is unique in The Netherlands, and even beyond.’
The Parallel Computing Systems (PCS) group is part of the Informatics Institute at the University of Amsterdam. It is the foremost research group in The Netherlands in the field of system optimization of multi-core and multi-processor computer systems. The PCS group looks at system performance, power/energy consumption, reliability, security & safety, but also the degree of productivity to design and program these systems: the extra-functional behaviour of computer systems in full glory.
The top research of the PCS group is indispensable for developments within, for example, Artificial Intelligence. In order to be able to cope with the increasingly demanding calculations in computer science, it is essential that computer systems become faster and more efficient. Without the skills of researchers within computer systems, AI, amongst others, was certainly not where it is today.
Questions that the PCS researchers try to answer are essential in times when we are increasingly dependent on the optimal functioning of computer systems. How to guarantee that an airbag inflates at exactly the right time? How do you make the computer system of a self-driving car so energy efficient that the battery does not drain too quickly? How can a computer system adapt itself automatically as a defense when under attack? All examples of topics in which the PCS group makes a big difference.
All members of staff within the group play a prominent role within their own field at a national and international level. The design methodologies developed by the group over the past 15 years are being increasingly picked up by companies. In recent years the group received a number of grants, including three EU Horizon 2020 grants worth 1.85 million euro and two NWO grants worth 840K euro..
The research focus of the group is fundamental, but always problem-driven. Collaboration through private-public partnerships with large industrial companies such as ASML, Philips, NXP, Thales Nederland and STMicroelectronics they ensure a reality check on their research.
A great example is the iDAPT project in which researchers develop models for ASML that use real-time data to detect and predict whether an error will occur in their chip lithography machines, which basically are huge ‘chip printers’. Important research, because an idle ASML machine costs millions of dollars.
The group works together with CERN on performance engineering. Since the CERN computer systems need to perform real-time processing of vast amounts of experimental data, the researchers are developing techniques to squeeze the last drop of performance out of their systems.
In the coming years, major questions in the field will arise around systems that automatically adapt to their environment or context. Within the ADMORPH project, researchers focus on the security of increasingly complex, networked computer systems. How can you use system adaptability to make a system more robust against external attacks or system errors such as defective processors?
Another trend the group is responding to is the use of the Internet of Things for large calculations. For example, within the TeamPlay project, researchers develop parallel software to perform large calculations on low-energy computer systems as part of the Internet of Things.
These exciting and relevant questions are just a tip of the iceberg.