SiFive Intelligence & VCIX

on YouTube ↗23m

" summary

SiFive ships the Intelligence X280 as a full 64-bit applications core with RVV 1.0 vector support and 32 512-bit vector registers, delivering up to 512 bits of ALU result per cycle and LMUL=8 chaining to 4K-bit vectors. Customers use it as a companion to larger accelerators; SiFive builds a VCIX: 1024b out / 512b in coprocessor interface that cuts round‑trip latency from ~100 cycles to ~10 and raises sustained bandwidth, yields a reported 24× scalar→vector speedup and an extra 6× with AI/ML extensions, plus Recode NEON migration, LLVM auto‑vectorization and tuned libraries.

AI-written summary · about

" chapters (7)

1. 0:00Introduction and overview
2. 1:03SiFive Intelligence X280 core
3. 3:48X280 companion core use
4. 5:11Vector coprocessor interface
5. 8:39Software support for X280
6. 11:50Cliff Young on accelerators
7. 19:15Q&A session

" related talks

• 13mSiFive Intelligence X280Drew Barbier · embedded world 2022
• 16mSiFive Vector Processor PortfolioAndrew Frame · embedded world 2022
• 17mRISC-V is Ready for Powering the Era of Intelligent General ComputingCharlie Su · RISC-V Summit North America 2025
• 13mKeynote: Taking RISC-V Intelligence EverywhereCharlie Su · RISC-V Summit North America 2023
• 24mAndes RISC-V Processors for Control and Data PathsCharlie Su · RISC-V Global Forum 2020
• 13mUnleashing ML Processing Power Through RISC-V Vectors: Applications, Implementation and...B. Barker · RISC-V Summit North America 2025
• 20mBoosting AI on Semidynamics RISC-V Cores with Custom Tensor Instr...Roger Espasa, José María Arnau · RISC-V Summit North America 2023
• 17mNext-Generation Edge AI with RISC-V Vector Cores for Vision ApplicationsFlorian Zaruba · RISC-V Summit North America 2025