Scope
The “Advanced Runtime Analysis” workshop brings together researchers and practitioners from both the hardware and software domains to explore how fine‑grained execution monitoring can be turned into provably safe, high‑performance optimisations. The event focuses on the entire measurement‑to‑action pipeline: from on‑chip telemetry (counters, event traces, and custom CSRs) that expose micro‑architectural behaviour, through runtime‑driven compilation and adaptation, to the enforcement of safety, security, and correctness properties. By situating these techniques squarely on RISC‑V platforms, the workshop offers a concrete substrate on which formal methods can be applied “close to hardware,” enabling participants to reason about the semantics of monitored events, generate verified instrumentation, and close the loop with hardware‑assisted de‑optimisation or recovery mechanisms.
Topics
Key topics include (i) hardware‑supported monitoring primitives such as RISC‑V Performance‑Monitoring Units, Trace Port Interface (TPI), and bespoke custom CSR extensions; (ii) just‑in‑time (JIT) and adaptive compilation frameworks that consume live profile data to perform speculative inlining, specialization, or vectorization while preserving correctness guarantees; (iii) safe‑checking mechanisms for memory safety, control‑flow integrity, and information‑flow enforcement that are generated from formal specifications and verified by model‑checking or theorem‑proving; (iv) secure‑checking pipelines that integrate hardware‑rooted attestation, runtime policy enforcement, and hardened debugging interfaces capable of stepping through instrumented code without compromising isolation; and (v) formal‑methods‑driven workflows—specification languages for monitors, automatic synthesis of hardware‑assisted checks, and verification of the composition of monitoring, optimisation, and de‑optimisation actions. The workshop also invites case studies, tool demonstrations, and research on cross‑layer debugging interfaces that combine hardware trace, software debuggers, and dynamic instrumentation to provide end‑to‑end visibility and correctness assurance.
In this workshop, we aim to explore these questions through presentations and focused discussions, bringing together experts from the corresponding fields.