Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Learn how to optimize virtual Performance Monitoring Unit (vPMU) implementation on ARM architecture in this 29-minute conference talk from KVM Forum. Discover the performance overhead issues caused by KVM's current vPMU implementation that traps and emulates the entire PMU, and explore innovative solutions to address these challenges. Examine how modern ARM CPU features including PMUv3 and Fine Grain Traps (FGT) enable selective untrapping of common PMU registers and features, allowing guests direct hardware access to dramatically reduce overhead and improve performance. Understand the technical implementation details and review notable performance improvements achieved through this optimization approach, with insights drawn from months of development work and testing results shared on the kvmarm mailing list.
