The winners of the Facebook AI System Hardware/Software Co-Design research awards have just been announced. Congratulation to Onur and the SAFARI Research Group for their proposal on Realistic Benefits of Near-Data Processing for Emerging ML Workloads!
Read more about this prestigious award, and the other winners on the Facebook website:
Onur talked at the Design Automation Summer School last week on “Enabling Computation with Minimal Data Movement: Changing the Computing Paradigm for High Efficiency”. The talk slides are now online:
Slides (pptx)
Related Overview Paper
Onur‘s slides from his Visionary Talk “Using Commodity Memory Devices to Support Fundamental Security Primitives”at the workshop on Energy-Secure System Architectures (ESSA), in conjunction with the 12th IEEE International Symposium on Hardware Oriented Security and Trust (HOST), McLean, VA, are now online: [Slides (pptx)]
Onur‘s course on “Memory Systems and Memory-Centric Computing Systems” taught at the HiPEAC ACACES Summer School 2018 is now online: https://safari.ethz.ch/memory_systems/ACACES2018/.
Take a refresher before the next summer school!
https://www.eurolab4hpc.eu
https://www.hipeac.net
https://www.tetramax.eu
Onur gave a keynote talk on “Processing Data Where It Makes Sense in Modern Computing Systems: Enabling In-Memory Computation” at the 29th ACM Great Lakes Symposium on VLSI GLSVLSI in Washington.
Onur Mutlu,
“Processing Data Where It Makes Sense in Modern Computing Systems: Enabling In-Memory Computation”
Keynote Talk at 29th ACM Great Lakes Symposium on VLSI (GLSVLSI), Washington, DC, USA, May 2019.
[Slides (pptx)]
[Related Overview Paper]
Juan Gómez Luna gave a keynote talk on Wednesday 22 May on “From GPGPUs to Processing-in-Memory: High Performance and Energy Efficiency for Computer Vision Workloads” at the British Machine Vision Association (BMVA) Computer Vision + High Performance Computing meeting in London. His slides are now online: https://bit.ly/30OrMUW
Congratulations to Lara Lazier, a bachelor student in our group, and her Team Kassiopeia, who took home the prize for “Most Creative Policy Response Alternative” during this year’s Cyber 9/12 Strategy Challenge hosted by the Geneva Centre for Security Policy: www.gcsp.ch/cyber-912-strategy-challenge-2019
Read more: Team Kassiopeia wins “Most Creative Policy Response Alternative”, D-INFK
Congratulations Lara and Team Kassiopeia!
Our new paper “NAPEL: Near-Memory Computing Application Performance Prediction via Ensemble Learning”, the first machine-learning-based prediction framework for Near-Memory Computing to be presented at the Design Automation Conference in June is now out:
Gagandeep Singh, Stefano Corda, Geraldo Francisco de Oliveira, Juan Gomez-Luna, Giovanni Mariani, Sander Stujik, Onur Mutlu, and Henk Corporaal, “NAPEL: Near-Memory Computing Application Performance Prediction via Ensemble Learning” Proceedings of the 56th Design Automation Conference (DAC), Las Vegas, NV, USA, June 2019.
We’ve also released Ramulator-PIM, a processing-in-memory simulation framework, used to develop NAPEL:
Source Code: https://github.com/CMU-SAFARI/ramulator-pim
