An Adaptive Interrupt Architecture for Extremely Timing-Critical Applications
Talkat RISC-V Summit North America 2024 · on Tuesday, 2024-10-22· at 14:15· Grand Ballroom H (Level 1)·
summary
A scalable interrupt microarchitecture duplicates most general-purpose registers and CSRs and adds a hardware preemption controller plus a per-interrupt lookup table to pick between conventional handling and an immediate ISR path, implemented in silicon and measured to produce about a 30% reduction in total control time. Device drivers expose modified IRQ APIs while keeping applications unchanged, and debug access uses privileged instructions or memory mappings to inspect the duplicated contexts, proving the approach works and remains orthogonal to existing optimizations.
AI-written summary · about
chapters (7)
- 0:00RISC-V at Samsung LSI
- 4:25Timing-critical challenges
- 7:10Immediate interrupt handling
- 9:37Avoiding fragmentation
- 10:59Adaptive architecture design
- 13:57Conclusion and summary
- 15:34Q&A session
related talks
31mFast Interrupts for RISC-V
01h 01mConcurrency Challenges of Interrupt Service Routines
47mWhat Every Driver Developer Should Know about RT
30mLook Ma’, No (Real) Interrupt Controller!
42mUnderstanding Linux Interrupt Subsystem
15mOptimizing Real-Time Application Requirements on ARC-V Processors Leveraging RISC-V Ex...
17mDemo: RISC-V AIA Expanding Interrupts: Applications, Implementation and Verification
57mCortex-M(etal): Hyper-optimized Zephyr-Friendly Context Switching for a Mature Architecture
