🔥36氪•Stalecollected in 20m
Alibaba DAMO Teams Up on RISC-V Push
💡Alibaba advances open RISC-V CPU for AI; key for custom chip builders.
⚡ 30-Second TL;DR
What Changed
Strategic cooperation signed March 24 at XuanTie conference
Why It Matters
Strengthens open-source alternatives to proprietary AI chips, potentially lowering costs for custom AI hardware.
What To Do Next
Download Xiangshan RISC-V sources and benchmark against Arm for your AI edge inference needs.
Who should care:Researchers & Academics
🧠 Deep Insight
AI-generated analysis for this event.
🔑 Enhanced Key Takeaways
- •The collaboration marks a significant shift in China's semiconductor strategy, moving from fragmented RISC-V development to a centralized, state-backed effort to create a 'national' high-performance CPU architecture.
- •The integration of Alibaba's XuanTie technology into the Xiangshan project aims to bridge the gap between academic research and commercial-grade industrial application, specifically targeting data center and AI acceleration workloads.
- •This partnership is designed to mitigate the impact of international export controls on advanced chip architectures by fostering a self-reliant, open-source hardware and software stack within the domestic Chinese market.
🛠️ Technical Deep Dive
- •Xiangshan (Mountain) Architecture: An open-source, high-performance RISC-V processor core developed by the Beijing Open Source Chip Academy, designed to scale from embedded to high-performance computing (HPC) applications.
- •SMT (Simultaneous Multithreading): The collaboration focuses on implementing robust SMT support to improve instruction-level parallelism and resource utilization in multi-core RISC-V designs.
- •On-chip Interconnects: Development of high-bandwidth, low-latency coherent interconnect fabrics to support multi-core scaling and heterogeneous computing requirements.
- •Unified Push Technology: A specialized cache-coherency or data-movement mechanism intended to optimize memory access patterns for AI workloads, reducing latency between the CPU and accelerators.
🔮 Future ImplicationsAI analysis grounded in cited sources
Xiangshan will achieve performance parity with mid-range ARM Neoverse cores by 2027.
The pooling of resources between Alibaba's commercial engineering teams and CAS's research capabilities significantly accelerates the development cycle of the Xiangshan architecture.
Alibaba will transition its internal cloud infrastructure to RISC-V based servers within three years.
The strategic focus on high-performance interconnects and software ecosystem compatibility indicates a clear path toward replacing proprietary architectures in Alibaba's data centers.
⏳ Timeline
2021-06
Beijing Open Source Chip Academy releases the first version of the Xiangshan RISC-V processor.
2021-10
Alibaba DAMO Academy open-sources the XuanTie RISC-V series IP cores.
2023-03
Xiangshan 'Nanhu' (second generation) architecture is taped out, demonstrating improved performance and scalability.
2025-05
Alibaba and CAS initiate preliminary discussions on integrating XuanTie software stacks with Xiangshan hardware.
📰
Weekly AI Recap
Read this week's curated digest of top AI events →
👉Related Updates
AI-curated news aggregator. All content rights belong to original publishers.
Original source: 36氪 ↗