Mobile Performance

Mobile Performance

Purpose

Plan and diagnose mobile performance: locate the real bottleneck (lists, re-renders, images/memory, startup, jank) via measurement and plan safe, behavior-preserving fixes on a representative build.

When to Use

• When a screen/list is janky, slow to start, or memory-heavy.
• Before a performance-motivated refactor.
• Not as blanket optimization without a measured bottleneck.

Inputs

• The slow scenario (screen, action, platform, build).
• Profiling data or a plan to obtain it.

Discovery Questions

• What is slow, under what action/scale, on which platform?
• Debug or release/production-like build?
• Is there a measurement, or must one be taken?
• Low-end device behavior?

Responsibilities

• Locate the bottleneck: list rendering (FlatList config, keys, memoized rows), re-renders, image/memory pressure, startup, or animation jank.
• Plan safe, targeted fixes preserving behavior.
• Validate on a representative build; capture before/after.
• Escalate structural fixes to ../../refactor-planning.

Required Workflow

1. Reproduce + measure on a representative build.
2. Trace the bottleneck to specific code.
3. Plan minimal behavior-preserving fixes.
4. Re-measure; capture before/after.
5. Record findings + residual risks.

Decision Rules

• No optimization without a measured bottleneck.
• Prefer small measurable wins over rewrites.
• Judge from release-like conditions, not debug.
• Don't add heavy libraries (e.g. list libs) without measured justification.

Rules

• Preserve behavior; performance work is not a feature change.
• Measure before/after; quote numbers.
• Flag unrun measurements 'unverified until run'.

Anti-Patterns

• Premature memoization / blind micro-optimization.
• Judging performance from a debug build.
• Rewriting instead of fixing the measured hotspot.
• Adding a list/image library without data.

Validation Checklist

• Scenario reproduced + measured (representative build).
• Bottleneck traced to code.
• Fixes minimal + behavior-preserving.
• Re-measured; before/after captured.
• Residual risks noted.

Definition of Done

A recorded mobile-performance report: measured bottleneck, safe fix plan (or applied fixes), before/after numbers from a representative build, and residual risks — aligned with ../../performance-review.

Related Skills

../../performance-review, ../../refactor-planning, mobile-state-management, mobile-vector-icons, mobile-camera-media, mobile-unit-testing

Related Knowledge

../../../knowledge/ (hot paths).

Related References

../../../references/mobile/screens/, ../../../references/mobile/state/ when populated.

Context Loading Guidance

• Requires: the slow scenario, measurement (or plan), the hot-path code.
• Does not require: unrelated screens, the full mobile skill set, unrelated references.
• May load: ../../refactor-planning if the fix is structural.
• Stop when: before/after is recorded (or measurement flagged unrun).

Token Efficiency Guidance

Read only the measured hot path; quote decisive metrics rather than full profiler dumps.