GreenWaves Presentation

on YouTube ↗21m

" summary

GreenWaves describes GAP8, a RISC-V MCU with an 8-core parallel cluster and integrated power management targeting battery-operated edge AI. Claim: cluster gives ~20x better power efficiency vs state-of-the-art for signal-processing use, doing a CIFAR-10 8-bit CNN in 1.5 million cycles (15.4 MHz) vs 21.4M on an STM32H7, cutting active power from ~60 mW to ~3.7 mW and down to 0.78 mW using a convolution accelerator. Tools: open-source SDK, GCC extensions and an OtterTail code generator. Product ships as GAP8 dev boards and silicon.

AI-written summary · about

" chapters (11)

1. 0:00Introduction
2. 1:30The Very Edge
3. 2:45Preprocessing
4. 4:59Gap
5. 10:39Instruction Set
6. 15:33Counting People
7. 16:23Industrial
8. 18:02Security applications
9. 19:04Wearables
10. 19:32Pulp
11. 20:45Availability

" related talks

• 18mThe Next Step in Low Energy Edge DSP and AIMartin Croome · embedded world 2022
• 21mNext Generation of GAP8 - IoT App Processor for Inference at the Very EdgeMartin Croome · RISC-V Summit 2019
• 40mProgramming a 144-computer chip to minimize powerChuck Moore · Strange Loop 2013
• 06mEmbedded Executive LIVERich Nass, Steve Brightfield · embedded world North America 2026
• 09mKeynote: RISC-V - A User's PerspectiveLoic Lietar · RISC-V Global Forum 2020
• 16mRVEdge-Vision: A Fully Open,Ultra-Efficient On-Device AI Platform for Smart EyewearMichele Magno · RISC-V Summit Europe 2026
• 12mKeynote: 100G Intelligent Edge Innovation with RISC-V (The Future is Now)Bruce Weyer · RISC-V Summit North America 2023
• 30mFrom Eyewear to Silicon: RISC-V for Low-Power AI in Next-Generation Smart GlassesMarco Fariselli · RISC-V Summit Europe 2026