Architectures and Algorithms for Energy-Efficient IoT and HPC Applications
Perugia (Italy) – September 3rd to 6th, 2019
The Summer School on Architectures and Algorithms for Energy-Efficient IoT and HPC Applications, bringing together international scientists from academia, research centers and industry, offers the opportunity to young researchers to be aware of the state of art on the development of energy efficient computer architectures, algorithms and their applications to the Internet of Things (IoT) and High Performance Computing (HPC) domains.
This is the tenth Summer School organised by NiPS Laboratory, and the second in collaboration with the E.C. funded project OPRECOMP.
World class speakers:
Scientific program:
Day 1 (Tuesday, September the 3rd)
- Memory Systems and Memory-Centric Computing Systems: Challenges and Opportunities (Onur Mutlu – ETH Zurich, Switzerland) | Videos
- PULP: A Multi-Core Platform for Micropower In-Sensor Analytics (Davide Rossi – University of Bologna, Italy) | Slides
Day 2 (Wednesday, September the 4th)
- Energy efficient heterogeneous design (Luca Benini – University of Bologna, Italy) | Slides
- Approximate (Trans-Precision) Computing Without Re-Inventing the Wheel (Andreas Burg – EPFL, Switzerland) | Slides
- Summer of Code presentations (Elevator speeches by the SoC winners)
- Ivan Fernadez Vega, Time series analysis using transprecision computing
- Tuba Ayhan, Employing Transprecision Computing Techniques on JPEG Compression System
- Jorge Castro-Godínez, Approximate Software for Accurate Hardware
- Davide Zoni, Design and implementation of an efficient Floating Point Unit (FPU) in the IoT era
- Matthew Tang, Circuit Design and Analysis for Approximate Integer Format
- Daniele Cattaneo, Value Range Analysis and Feedback-Driven Optimization for a Mixed Precision Compiler
- Marco Widmer, Approximate Computing with Unreliable Memory: FPGA-Based Emulation for PULP
Day 3 (Thursday, September the 5th)
- Energy Efficient Inter-Chip Communication in Heterogeneous Application Domains (Vasilis Pavlidis – University of Manchester, United Kingdom) | Slides
- ConvNets on the Edge: Optimization before Acceleration (Andrea Calimera – Politecnico di Torino, Italy) | Slides
-
Efficient Memory Controller Architectures and their Integration in Transprecision Compute Platforms (Christian Weis – University of Kaiserslautern, Germany) | Slides
- Understanding performance numbers in HW Design (Frank K. Gürkaynak – ETH Zurich, Switzerland) | Slides
Day 4 (Friday, September the 6th)
- Dynamic Application Autotuning for Approximate Computing (Gianluca Palermo – Politecnico di Milano, Italy) | Slides
-
CNN on Gap8 – From NN training to full network deployment on an IoT embedded processor (Francesco Paci – GreenWaves Technologies, France) | Slides
- Enabling Machine Learning on FPGA (Mirko Mariotti – University of Perugia, Italy) | Slides
During the school we will give chances to young researchers to present they works through a poster session which will span all days of the school. Moreover, winners of the current edition of the OPRECOMP Summer of Code (SoC) will present their projects and results, with the aim to stimulate discussions and collaborations among participants.
Steering committee:
- Igor Neri (University of Perugia, Italy)
- Frank K. Gürkaynak (ETH Zurich, Switzerland)
- Cristiano Malossi (IBM Zurich Research Laboratory, Switzerland)
- Enrique Salvador Quintana Ortí (Universidad Politécnica de Valencia, Spain)
- Davide Rossi (University of Bologna, Italy)
- Henri-Pierre Charles (CEA, France)
- Giuseppe Tagliavini (University of Bologna, Italy)
Scientific Director:
Luca Gammaitoni, NiPS Laboratory, Dipartimento di Fisica e Geologia, Università di Perugia
All logistic information and the registration procedure can be found here.