Postdoctoral Researcher in Programming Models and Processing-in-Memory technologies

As a postdoctoral researcher with a focus on programming models, your main task will be to link our programming model efforts in the SSF PRIDE, and the EuroHPC EUPilot and eProcessor projects. The focus of PRIDE are processing-in-memory technologies. The focus of EUPilot and eProcessor are long vector multicore architectures. The goal of our team is to develop effective runtime support to program these architectures using programming models such as OpenMP and SYCL.

The core team currently consists of seven members, five of which are women. We are part of the computing systems unit (4 professors) within the unit of Computing Networks and Systems, which itself belongs to the Compute Science and Engineering department. Our team contributes to many Swedish and European projects. The team also has a long running collaboration with Riken-CCS in Kobe, Japan. The team meets once a week to discuss updates and research topics. In addition the team leader holds individual weekly meetings with all PhD students and postdocs.

Our research infrastructure is provided mainly by NAISS (National Academic Infrastructure for Supercomputing in Sweden) and C3SE (Chalmers Centre for Computational Science and Engineering). Our partners and funding projects provide access to hardware testbeds, such as RISC-V vector processors, Fujitsu A64FX, Graviton3, FPGAs, etc.

As a postdoc you will be given significant autonomy and responsibility. You will participate in the co-supervision of PhD students, and you will be involved in teaching. Courses on supervision and teaching skills are offered by Chalmers. The goal is to effectively prepare you for a future academic career. The position is designed for a two-year period, with a possibility to extend the employment to three years.

Major responsibilities

• Lead our research on moldability and locality in OpenMP and SYCL
• Research how to compile SYCL programs to long vector multicore architectures
• Research extensions to OpenMP and SYCL to support the PIM (processing-in-memory) paradigm
• Teaching and Co-supervision of PhD students
• Support the PI in leading two workpackages in two separate European projects.