Demo: TraceLLM - Harness the Full Potential of your RISC-V Systems with an AI Based, Real-time, RISC-V Embedded Application Debug and Trace Analysis Engine
Talkat RISC-V Summit North America 2024 · on Tuesday, 2024-10-22· at 12:55· Expo Hall - Exhibit Hall A - Demo Theater·
summary
TraceLLM integrates an LLM analysis engine with the RiscFree IDE and Ashling Vitra‑XS probe to capture non‑intrusive, real‑time traces and reconstruct full call stacks over long runs. It answers natural‑language queries like “which function used the most time,” returns function hot‑spot tables, call counts and code links, and reports exact code‑coverage hits (examples show 100% code coverage results). Demo streams traces into a 10 GB buffer storage and the TraceLLM engine produces immediate, actionable outputs.
AI-written summary · about
chapters (9)
- 0:00Introduction to Ashling Microsystems
- 1:30Ashling’s debug and trace tools
- 2:56RIS-free IDE overview
- 3:59VRA access debug and Trace probe
- 5:00Trace insights with VRA access
- 6:45Non-intrusive real-time tracing demo
- 8:10AI-driven TraceLLM introduction
- 9:13TraceLLM demo with natural language
- 10:58Q&A and booth visit invitation
related talks
44mAccelerating Software Development: The LLM (R)evolution- 44mJoint Keynote Address: Accelerating Software Development: The LLM (R)evolution
30mDevRelling in the age of GenAI
12mBeyond Generic Spans: Distributed Tracing for Actio...
16mIs that LLM feature any good?
01mSpeed Up Your Code With LLM's!
17mRLM: Recursive Language Models for Large Codebases
45mDebugging tools for LLM apps in JS