RISC-V is Ready for Powering the Era of Intelligent General Computing

on YouTube ↗17m

" summary

Claims concrete hardware and software wins: Andes offers AX46MPV with up to 1024-bit vector and memory-path optimizations that boost vector bandwidth ~40% and deliver ~5.65 SPECint/GHz; AX66 adds 128-bit vectors and shows up to 9.6x NN kernel speedups and ~3.5x average vector gains; the high-end Cuzco targets 20 SPECint/GHz when configured as a six‑way super out‑of‑order core. The stack includes Andes AI, ACE custom extensions and llama.cpp ports that report ~26–30× LLM speedups on their silicon.

AI-written summary · about

" chapters (9)

1. 0:00The era of intelligent computing
2. 0:14RISC-V enabled AI innovation
3. 2:36Maturing RISC-V ecosystem
4. 3:22Open-source development status
5. 4:51Andes processor roadmap
6. 6:26Deep dive into AX46MPV and AX66
7. 11:29Cuzco processor innovation
8. 14:08AndesAIRE AI software stack
9. 15:14OS support and conclusion

" related talks

• 20mRISC-V Spotlight: Expanding the RISC-V Horizon and BeyondCharlie Su · RISC-V Summit 2022
• 13mKeynote: Taking RISC-V Intelligence EverywhereCharlie Su · RISC-V Summit North America 2023
• 17mThe Real Challenge for RISC-V Vector ProcessorsJohn Min · RISC-V Forum: Vector & Machine Learning 2021
• 24mAndes RISC-V Processors for Control and Data PathsCharlie Su · RISC-V Global Forum 2020
• 40mKeynote Panel: The Future of High Perfor...Nick Brown, Wei-han Lien, Luisa Gonzalez, Travis Lanier · RISC-V Summit North America 2024
• 10mRISC-V Revolutionizing in AIVince Wu · RISC-V at embedded world 2024
• 30mRISC-V: Enabling Open Physical AILuca Benini · RISC-V Summit Europe 2026
• 24mReal World Results using Thousands of RISC-V Cores for AI and BeyondDave Ditzel · RISC-V Summit 2022