Linux Kernel 7.2-rc2 Increases RISC-V CPU Limit to 256
💡Critical infrastructure update for RISC-V server adoption, essential for scaling AI and high-performance computing.
⚡ 30-Second TL;DR
What Changed
Default NR_CPUS for RISC-V 64-bit increased from 64 to 256.
Why It Matters
This update signals the maturation of the RISC-V ecosystem for data center and AI infrastructure, allowing developers to deploy more powerful, multi-core RISC-V clusters for large-scale computing tasks.
What To Do Next
If you are building AI workloads on RISC-V hardware, recompile your kernel with the updated NR_CPUS limit to leverage high-core-count server architectures.
🧠 Deep Insight
AI-generated analysis for this event.
🔑 Enhanced Key Takeaways
- •The increase to 256 CPUs specifically addresses the 'NR_CPUS' configuration limit, which dictates the maximum number of processors the kernel can manage at boot time.
- •This adjustment is part of a broader effort to align RISC-V kernel configurations with other high-performance architectures like ARM64, which often support significantly higher core counts.
- •Developers noted that the memory overhead for increasing this limit is negligible on 64-bit systems, as the kernel uses sparse memory allocation for per-CPU structures.
- •The patch was submitted by the RISC-V Linux maintainers to ensure that upcoming server-grade SoCs, which are expected to exceed the previous 64-core threshold, can boot without custom kernel recompilation.
- •This change specifically targets the 'defconfig' (default configuration) for RISC-V, allowing distributions to support larger hardware out-of-the-box without requiring users to manually adjust kernel parameters.
🛠️ Technical Deep Dive
- The NR_CPUS parameter defines the compile-time constant for the maximum number of CPUs supported by a specific kernel binary.
- Increasing NR_CPUS impacts the size of static per-CPU arrays and bitmasks used for CPU affinity and scheduling.
- On RISC-V 64-bit, the kernel utilizes the 'smp_ops' framework to handle multi-core initialization, which remains compatible with the increased limit.
- The memory footprint increase is primarily associated with the 'cpumask' structures and per-CPU data sections, which scale linearly with the number of CPUs but remain well within the limits of modern server memory capacities.
🔮 Future ImplicationsAI analysis grounded in cited sources
⏳ Timeline
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: IT之家 ↗