Exploration of Computation near Storage
What you will do
Data movement between processing cores and memories is becoming a critical bottleneck. In that context, in recent high-performance multi-processors, moving the data from the memory to the processor costs more than performing the computation itself. In systems where the main memory capacity is insufficient for the working set of the running applications, the critical bottleneck moves to the storage interface, where the bandwidth is lower and energy per transaction is higher. This dramatically reduces the energy efficiency of processing cores running data-centric workloads. Thereby, both academia and industry are looking for solutions to improve the computation efficiency of such architectures.
One promising solution is the integration of computing units directly inside the memory hierarchy creating the concept of in- or near-memory computing, depending on where the computing capabilities are integrated. These solutions consist on performing the computation as close as possible as where the data is located. Thereby, long data transfer latency and energy overhead is mitigated, more bandwidth is available, and the processors are freed to perform other tasks. As the selected candidate for this postdoc, you will
- Propose solution to tackle the Memory wall and data transaction overhead by means of innovative architectural “Compute Near Storage Solutions”
- Explore how the compute should be integrated with conventional and emerging storage technologies, for several data-dominated applications, analyzing the trade-offs in performance and power.
- Integrate proposed architectures into a larger full-system simulation framework, based on gem5, so that the resulting system power and performance benefits can be accurately assessed.
- Be a part of a multi-disciplinary team, in close interaction with technology and circuit-level considerations, that will provide realistic inputs to ensure accurate and reliable simulations.
- Set the roadmap for the computer systems of the future.
What we do for you
We offer you the opportunity to join one of the world’s premier research centers in nanotechnology at its headquarters in Leuven, Belgium. With your talent, passion and expertise, you’ll become part of a team that makes the impossible possible. Together, we shape the technology that will determine the society of tomorrow.
We are proud of our open, multicultural, and informal working environment with ample possibilities to take initiative and show responsibility. We commit to supporting and guiding you in this process; not only with words but also with tangible actions. Through imec.academy, 'our corporate university', we actively invest in your development to further your technical and personal growth.
We are aware that your valuable contribution makes imec a top player in its field. Your energy and commitment are therefore appreciated by means of a competitive scolarship.
Who you are
- You have a Ph.D. degree in Electrical or Computer Engineering
- We need your experience with memory subsystems and computer architecture
- We are looking for your knowledge in next generation memory technologies
- You have experience in development of architectural simulator (gem5)
- We value your experience with industry-standard EDA tools like Cadence/Synopsys, specifically system modeling tools and languages: SystemC, Platform Architect or similar.
- We are looking for a good team player and your ability to work independently
- You have excellent English communication skills and enjoy working in a multicultural environment
This postdoctoral position is funded by imec through KU Leuven. Because of the specific financing statute which targets international mobility for postdocs, only candidates who did not stay or work/study in Belgium for more than 24 months in the past 3 years can be considered for the position (short stays such as holiday, participation in conferences, etc. are not taken into account).