You are here : Home > APPLIED RESEARCH > STRATEGIC AXES > Energy Efficiency of ICTs

Article | Laboratory | Micro-nanotechnologies | Materials


Energy Efficiency

Energy Efficiency

ICT energy efficiency is one of the major challenges of our time. According to the worst scenario, ICTs could represent up to 51% of the world's energy consumption and 23% of greenhouse gases in 2030. 

Published on 1 March 2017

Energy efficiency at the device, chip, board and system levels: ensuring the most efficient use of energy resources

ICT energy efficiency is one of the greater challenges of our time. According to worst case scenarios, ICTs could represent up to 51% of the world's energy consumption and 23% of greenhouse gases in 2030. 

Energy consumption of high-performance computer systems, already at 20 MW, is a major technological and economic issue. To compensate the increase in power consumption of mobile devices that accompanies the increase in  their computing capacity, long term research is ongoing to improve battery performance. Short term solutions must thus include drastically lowering the energy budget of intelligent systems used in the Internet of Things (IoT) to guarantee operation below 1 mW. This includes lowering standby power consumption, as IoT devices operate at very low duty cycles. Rethinking the whole family of technologies, from devices to applications, will allow us to meet these challenges and maximize energy efficiency.

Leti is developing technologies and architectures to address ICT power management such as low-power CMOS (FD-SOI, TFET, etc.), neuromorphic circuits based on ReRam memories, computation modules using 3D integration, silicon-based photonics, stacked intelligent 3D imagers, etc. The capacity of both hardware and software to adapt dynamically to environmental conditions and different application requirements is also under study to curtail energy waste.



Click to enlarge 

Multi-core module based on a 3D interposer and FD-SOI chiplets ​Multi-core architecture based on a Silicon Photonics electro-optical interposer ​Photograph of a 28 FDSOI based DSP VLIW functioning at 460MHz and 400mV

Eye diagram of an optical modulator amplifier functioning at 10Gbps

Photos ©CEA

Referenced publications in our annual research reports:

Beigné, E.; Valentian, A.; Miro-Panades, I.; Wilson, R.; Flatresse, P.; Abouzeid, F.; Benoist, T.; Bernard, C.; Bernard, S.; Billoint, O.; Clerc, S.; Giraud, B.; Grover, A.; Le Coz, J.; Noel, J.-P.; Thomas, O. & Thonnart, Y. (2015), 'A 460 MHz at 397 mV, 2.6 GHz at 1.3 V, 32 bits VLIW DSP g FMAX Tracking', IEEE Journal of Solid-State Circuits 50(1), 125-136

Belleville, M.; Thomas, O.; Valentian, A. & Clermidy, F. (2013), 'Designing digital circuits with nano-scale devices: Challenges and opportunities', Solid-State Electronics 84, 38-45

E Vianello, O Thomas, G Molas, O Turkyilmaz, N Jovanović, D Garbin, .et al., "Resistive Memories for Ultra-Low-Power embedded computing design,Electron Devices Meeting" (IEDM) 2014 IEEE International, 6.3.1-6.3.4
  
O.Weber et al., “14nm FDSOI Upgraded Device Performance for Ultra-Low Voltage Operation”, VLSI Technology Symposium, 2015

O. Weber et al., (invited) “Static and Dynamic Power Management in 14nm FDSOI Technology”, ICICDT conference, 2015

O. Weber et al., (invited) “14nm FDSOI Technology for High-Speed and Energy-Efficient CMOS”, ECS transactions 2015 - issue 20