domains / software-engineering / performance-profiling / performance-profiling

performance-profiling

Skill card

combined

The three copies share the name 'performance-profiling' but target three different domains: Apple platform apps (Instruments/MetricKit), web performance (Lighthouse/Core Web Vitals), and computational simulations (Python scripts for scaling/memory). None is a reproduction of the others; each contains material steps, tools, and constraints the others lack.

Repositories

Other copies

Combined skill

# Performance Profiling

Use this skill to diagnose performance issues systematically, pick the right profiling workflow, apply targeted fixes, and verify the change with real measurements.

## Domain Selection

Determine the target domain before proceeding:

- **Apple Platform Apps**: Use Instruments, Xcode diagnostics, and MetricKit. Profile on device with Release builds.
- **Web Applications**: Use Lighthouse, DevTools, and RUM. Focus on Core Web Vitals (LCP, INP, CLS).
- **Computational Simulations**: Use Python scripts for timing, scaling, and memory analysis. Requires Python 3.8+.

## Workflow

1. **Baseline**: Measure current state using the appropriate tool for the domain.
2. **Identify**: Find the bottleneck using domain-specific analysis.
3. **Fix**: Apply the smallest targeted fix that addresses the measured bottleneck.
4. **Validate**: Re-profile to confirm the improvement.

## Apple Platform Apps

- Profile on device with Release configuration.
- Use Time Profiler for CPU/hangs, Allocations/Leaks for memory, App Launch for launch time, Energy Log for battery/thermal.
- Add `os_signpost` markers for ongoing timing visibility.
- Use MetricKit for production monitoring of launch, responsiveness, and memory.

## Web Applications

- Measure Core Web Vitals: LCP (<2.5s), INP (<200ms), CLS (<0.1).
- Use Lighthouse for page load, Bundle analyzer for bundle size, DevTools for runtime and memory.
- Prioritize: enable compression, lazy load images, code split routes, cache static assets.

## Computational Simulations

- Run timing analyzer on simulation logs to identify dominant phases.
- Run scaling analyzer on multi-run data to assess parallel efficiency.
- Run memory profiler on simulation parameters to estimate memory requirements.
- Run bottleneck detector to combine analyses and get actionable recommendations.

## Ground Rules

- Profile with representative data, not empty databases or toy assets.
- Keep measurements before and after the fix so the outcome is concrete.
- Do not guess at problems; profile first.
- Fix the biggest issue, not micro-optimizations.