perfetto-sql

You can query Android trace data by translating natural language intents into syntactically valid Perfetto SQL. This allows you to extract performance metrics like slice durations and thread states without manual SQL construction.

To analyze UI jank with Perfetto traces, you can request calculations for total slice durations matching specific render thread patterns. This identifies performance bottlenecks across trace files using standard library schemas.

Yes, you can extract app startup times and CPU frequency from local trace files. The query generation leverages pre-built Perfetto standard library modules to calculate these complex performance metrics automatically.

To calculate total RenderThread slice duration, provide the trace file and ask the system to compute the duration of all matching slices. It generates an idempotent SQL query adhering to Perfetto schemas for execution.

Yes, this approach supports memory usage analysis for Android profiling. It translates natural language data intents into valid SQL queries to extract memory usage metrics directly from local Perfetto trace files.

Query generation for trace analysis is limited by adherence to Perfetto Standard Library schemas and requires syntactically valid SQL execution. Complex intents must map to existing schema structures to yield accurate performance metrics.