broad-performance-investigation

To identify performance optimization opportunities in a Truffle language implementation, you need systematic analysis across benchmarking, code analysis, CPU sampling, memory profiling, and IGV inspection to generate actionable theories.

Analyzing GraalVM CPU sampler data requires coordinating findings with code analysis, memory profiling, and IGV inspection to ensure equal weight, generating tier-specific theories ready for implementing performance fixes.

Yes, Truffle performance analysis requires a BENCHMARK_BASELINE.md file and timing data as mandatory inputs to establish current performance baselines and produce prioritized optimization theories.

IGV analysis contributes to uncovering Truffle performance optimizations by acting as one of four mandatory analysis steps, inspecting compiler graphs to surface findings that form actionable performance theories.

When Truffle performance investigation findings are uncertain, the process orchestrates a deep-performance-investigation to resolve ambiguities and deliver a ready-to-implement plan for the current severity tier.

A broad performance investigation for GraalVM is limited to producing tier-specific PERFORMANCE_THEORIES.md, meaning it identifies opportunities but requires a subsequent implementing-performance-fixes step to execute changes.