Performance Profiling

Instruments decision trees and profiling workflows for CPU, memory, and battery optimization with 3 real-world examples.

When to use: App feels slow, memory grows over time, battery drains fast, want to profile proactively

Key Features

• Performance Decision Tree – Choose the right tool (Time Profiler, Allocations, Core Data, Energy Impact)
• Time Profiler Deep Dive – CPU analysis, hot spots, Self Time vs Total Time distinction, 4x improvement example
• Allocations Deep Dive – Memory growth diagnosis, object counts, leak vs caching, distinguishing real issues
• Core Data Deep Dive – N+1 query detection with SQL logging, prefetching, batch optimization, 7x improvement example
• Real-World Examples (3 complete workflows):
  • Identifying N+1 query problem in Core Data (SQL logging, diagnosis, prefetching fix)
  • Finding where UI lag really comes from (Time Profiler workflow, Self Time vs Total Time)
  • Memory growing vs memory leak (Allocations workflow, distinguishing caching from leaks)
• Pressure Scenarios – App Store deadline pressure, manager authority pressure, misinterpretation prevention

Requirements: Xcode 15+, iOS 14+

Example Prompts

These are real questions developers ask that this skill answers:

• "Scrolling is slow and I need to know if it's Core Data or SwiftUI." → Shows how to use Time Profiler and Core Data instruments to pinpoint the bottleneck

• "My app feels slow but I don't know what to optimize first." → Provides decision tree for choosing the right profiling tool and interpreting results

• "I profiled with Time Profiler but Self Time and Total Time are confusing." → Explains the distinction and shows how to identify real hot spots vs call stacks

• "I have a deadline and my app feels slow. What should I optimize?" → Provides systematic profiling under pressure with professional guidance for trade-offs

• "Memory keeps growing. Is it a leak or normal? How do I tell?" → Covers Allocations instrument decision tree to distinguish caching from actual leaks

Decision Tree

App performance problem?
├─ App feels slow/lags?          → Time Profiler (measure CPU)
├─ Memory grows over time?        → Allocations (find object growth)
├─ Data loading is slow?          → Core Data instrument (if using Core Data)
│                                 → Time Profiler (if computation slow)
└─ Battery drains fast?           → Energy Impact (measure power)

Core Profiling Tools

Time Profiler

Measures CPU time spent in each function. Identifies hot spots and main thread blocking.

When to use: App stalls, UI lag, CPU usage high

Key metrics:

• Self Time = Time spent IN that function
• Total Time = Time in that function + everything it calls
• If Self Time is low but Total Time is high → something it calls is slow

Allocations

Measures memory growth and object creation. Identifies memory leaks vs normal caching.

When to use: Memory grows, want to check for leaks, memory pressure issues

Key metrics:

• Growing count + growing memory = possible leak
• Memory drops under pressure = normal caching
• Memory stays high indefinitely = leak

Core Data Instrument

Analyzes SQLite queries and Core Data performance.

When to use: Using Core Data, data loading is slow, queries look expensive

Setup:

# Edit Scheme → Run → Arguments Passed On Launch
-com.apple.CoreData.SQLDebug 1

Common issue: N+1 queries (fetching parent, then individual query per child)

// ❌ WRONG: Individual query for each relationship
for item in items {
    print(item.relationship.title)  // Extra query each time
}

// ✅ RIGHT: Prefetch the relationship
let request = Item.fetchRequest()
request.relationshipKeyPathsForPrefetching = ["relationship"]
let items = try context.fetch(request)
for item in items {
    print(item.relationship.title)  // Already loaded
}

Energy Impact

Measures power consumption by CPU, GPU, network, location, and other subsystems.

When to use: Battery drains fast, device gets hot

Core Principle

Measure before optimizing. Guessing about performance wastes more time than profiling.

The cost of misdiagnosis:

• Blindly threading code: 2+ hours guessing wrong functions
• Wrong optimization: 4+ hours fixing something that wasn't the bottleneck
• Systematic diagnosis: 15-20 minutes to find the real issue

Real-World Workflow

1. Identify the symptom (slow, memory, battery)
2. Choose the right tool from decision tree
3. Record a profile (cold run first, then warm)
4. Analyze results (read call stacks, check metrics)
5. Identify root cause (which function/query/object is actually the problem)
6. Fix the right thing (not a guess, but evidence-based)
7. Verify improvement (re-profile to confirm the fix worked)

Related Skills

• swiftui-performance – SwiftUI Instrument for view profiling
• memory-debugging – Deep memory leak diagnosis with Instruments
• xcode-debugging – Xcode environment issues

Quick Start

1. What's the symptom? (slow/memory/battery)
2. Which tool? (Time Profiler/Allocations/Core Data/Energy Impact)
3. Record a profile while reproducing the issue
4. Analyze the call stack or statistics
5. Identify the root cause
6. Fix based on evidence, not guesses

Key principle: Self Time vs Total Time distinction prevents optimization rabbit holes.

---

Learn more: See the full skill at /skills/debugging/performance-profiling in Claude Code