PhD student position in HPC Programming Models and Computer Architectures

Updated: almost 2 years ago
Deadline: 13 Jun 2022

Welcome to join a team of researchers developing a homegrown European computing infrastructure based on open source software and hardware technologies. We are looking for a highly motivated team-player who can develop novel programming and architectural techniques to drive next-generation heterogeneous HPC infrastructure. This position will develop your skills in high performance languages and computer architectures, and you will also learn how to communicate research ideas following scientific practice.

Information about the research group

The Computer Architecture Research Group, within the Division of Computer Networks and Systems under the leadership of Professor Per Stenström, is conducting research on design principles for the next generation of computer systems. To ease the programming task and yet achieve high computational performance at as high energy efficiency as possible is an important objective in the knowledge generation process that the group contributes with. To this end, the group has a solid track record and a long-term focus on contributing to design principles of parallel computers. Four senior faculty members, six postdocs and twelve Ph.D students are engaged in research in Computer Architecture. The group is a founding partner of the HiPEAC European Network of Excellence and engaged in several EU and national projects with strong ties to industry, among them the European Processor Initiative (EPI).

Chalmers is involved in the EPI project in several workpackages. One line of the work focuses on the engineering of the memory subsystem. A second line focuses on the development and optimization of parallel programming models. The team is also involved in an SSF project related to Processing-in-Memory (PIM) technolgies.

Major responsibilities

Computer architectures are becoming increasingly parallel and heterogeneous, requiring novel programming model techniques to program and manage this diversity. At the same time, data communication costs are becoming dominant in parallel computing leading to novel PIM-like designs that tightly integrate processors with stacked DRAM. Yet languages and computer architectures provide very limited support to effectively manage data. In this project you will develop novel techniques to jointly manage compute and memory heterogeneity, on top of open source programming models such as OpenMP or SYCL.

Qualifications

To qualify as a PhD student, you must have a master's level degree corresponding to at least 240 higher education credits in a field related to Computer Engineering. You are highly motivated, self-propelled, energetic, independent and with a well-developed analytical problem-solving ability. As a person, you are empathetic, loyal and have high ethical standards. Your communication skills in English (literal as well as oral) are of a high quality.



Similar Positions