SwiftUI Performance Analysis

This skill helps find and fix SwiftUI performance issues. When triggered, follow this workflow:

Recommended Workflow

1. Review Code → 2. Record Trace (if needed) → 3. Analyze → 4. Fix

Start with code review (immediate results), then offer trace recording for deeper insights.

When to Use This Skill

Use this skill when user says things like:

• "Find and fix SwiftUI performance issues"
• "My app UI is slow/laggy"
• "Optimize my SwiftUI views"
• "Fix the re-render issues in my app"
• "Why is my view updating so much?"
• "Profile my SwiftUI app"
• "Help with SwiftUI performance"

Mode 1: Code Review (No Trace Required)

Review SwiftUI source files directly for performance anti-patterns. This is the fastest way to find issues.

What to Look For

Scan Swift files for these anti-patterns:

Pattern	Problem	Fix
@ObservedObject on large objects	All views re-render on any property change	Use @Observable (iOS 17+) or split state
Passing whole model to child views	Unnecessary re-renders when unused properties change	Pass only needed primitives
Timer/animation at parent level	Cascades updates to all children	Isolate to affected view only
Missing Equatable on data-driven views	Can't skip unnecessary re-renders	Implement Equatable
@State for derived values	Redundant state, sync issues	Use computed properties
@EnvironmentObject for local state	All consumers re-render together	Scope to minimal views

Code Review Process

1. Find SwiftUI view files:

   find . -name "*.swift" -exec grep -l "struct.*View" {} \;
   

2. For each view, check:

   • What state properties does it observe? (@State, @ObservedObject, @EnvironmentObject)
   • Does it pass whole objects to child views?
   • Does it have timers or frequent updates?
   • Could it implement Equatable?

3. Apply fixes from references/fix-patterns.md

Example Review

// ANTI-PATTERN: Whole object passing
struct UserList: View {
    @ObservedObject var viewModel: AppViewModel  // Large object

    var body: some View {
        ForEach(viewModel.users) { user in
            UserRow(user: user)  // Passes entire User object
        }
    }
}

// FIXED: Pass only needed data
struct UserList: View {
    @ObservedObject var viewModel: AppViewModel

    var body: some View {
        ForEach(viewModel.users) { user in
            UserRow(name: user.name, avatarURL: user.avatarURL)
        }
    }
}

Mode 2: Record & Analyze Trace (Deeper Insights)

For quantitative data on actual re-render counts and update chains, help the user record and analyze a trace.

Step 1: Check Prerequisites

# Check if swiftuice is installed
swiftuice version

# If not installed:
go install github.com/greenstevester/swiftui-cause-effect-cli/cmd/swiftuice@latest

Step 2: Find the App Bundle ID

Help the user find their app's bundle ID:

# From Xcode project
grep -r "PRODUCT_BUNDLE_IDENTIFIER" *.xcodeproj/project.pbxproj 2>/dev/null | head -1

# Or check the Info.plist
find . -name "Info.plist" -exec grep -A1 "CFBundleIdentifier" {} \; 2>/dev/null | head -5

# If app is running in simulator
xcrun simctl listapps booted 2>/dev/null | grep -A1 CFBundleIdentifier | head -10

Step 3: Record the Trace

Guide the user through recording:

# Record for 15 seconds - user should interact with the app during this time
swiftuice record -app <BUNDLE_ID> -time 15s -out trace.trace

Important instructions for the user:

1. Make sure the app is running on a simulator or device
2. During the 15 seconds, perform the UI actions that feel slow
3. Focus on the specific screens/flows with performance issues

Step 4: Analyze the Trace

# Analyze with source correlation
swiftuice analyze -in trace.trace -source . -stdout

Alternative: Manual Recording in Instruments

If swiftuice record doesn't work, guide the user through Instruments:

1. Open Instruments: Xcode → Open Developer Tool → Instruments
2. Choose the SwiftUI template
3. Select target device/simulator and app
4. Click Record (red button)
5. Interact with the app for 10-15 seconds
6. Click Stop
7. File → Save as .trace file

Then analyze:

swiftuice analyze -in /path/to/saved.trace -source . -stdout

Workflow

Step 1: Identify the Input

Determine what input is available:

Input Type	Command
.trace file	swiftuice analyze -in path/to/trace.trace
Exported directory	swiftuice analyze -in path/to/exported/
Need to record	swiftuice record -app <bundle-id> first

Step 2: Run Analysis

Execute the analysis with source correlation if Swift source is available:

# With source correlation (recommended)
swiftuice analyze -in <trace-or-export> -source <swift-project-root> -stdout

# Without source correlation
swiftuice analyze -in <trace-or-export> -stdout

The -stdout flag outputs JSON directly for parsing.

Step 3: Parse the JSON Output

The output contains these key sections:

{
  "summary": {
    "performance_score": 65,
    "health_status": "warning",
    "issues_found": 3,
    "critical_issues": 1
  },
  "issues": [...],
  "source_correlations": [...],
  "agent_instructions": {...}
}

Step 4: Implement Fixes

For each issue in the issues array:

1. Check the severity (critical, high, medium, low)
2. Review suggested_fixes with code_before and code_after examples
3. Use source_correlations to navigate to the affected files
4. Apply the fix pattern that best matches the codebase

Issue Types and Fix Patterns

Excessive Re-render (excessive_rerender)

Problem: A view updates too frequently (>10 times per user action)

Fix approaches:

1. Implement Equatable on the View to control when it re-renders
2. Extract subview to isolate frequently-updating content
3. Use @Observable (iOS 17+) for fine-grained observation

Example fix:

// Before
struct ItemRow: View {
    let item: Item
    var body: some View { ... }
}

// After - Add Equatable
struct ItemRow: View, Equatable {
    let item: Item
    var body: some View { ... }

    static func == (lhs: ItemRow, rhs: ItemRow) -> Bool {
        lhs.item.id == rhs.item.id
    }
}

Cascading Update (cascading_update)

Problem: Single state change triggers many view updates

Fix approaches:

1. Use derived state with computed properties
2. Split state into focused objects
3. Scope ObservableObject to only views that need it

Timer Cascade (timer_cascade)

Problem: Timer/animation causes broad UI updates

Fix approaches:

1. Use TimelineView for time-based content
2. Limit timer scope to only the view that needs it
3. Extract animated content into isolated subview

Whole-Object Passing (whole_object_passing)

Problem: Passing entire model objects causes unnecessary re-renders

Fix approaches:

1. Pass primitives instead of whole objects
2. Define focused protocols with minimal data requirements
3. Use @Bindable (iOS 17+) for specific properties

Recording a New Trace

If user needs to create a trace:

# Record for 15 seconds
swiftuice record -app com.company.appname -time 15s -out trace.trace

# Then analyze
swiftuice analyze -in trace.trace -source ./MyApp -stdout

Human-Readable Summary

For user-facing summary instead of JSON:

swiftuice summarize -in <trace-or-export> -out summary.md -dot graph.dot

This creates:

• summary.md: Markdown summary with top issues
• graph.dot: Graphviz diagram (render with dot -Tpng graph.dot -o graph.png)

Verification After Fixes

After implementing fixes:

1. Record a new trace with the same user flow
2. Run analysis again
3. Compare performance_score (should increase)
4. Verify issues_found decreased
5. Check specific view update counts reduced

Troubleshooting

"no parseable Cause & Effect data found"

• The SwiftUI instrument may not have captured data
• Try recording longer or with more UI interaction
• Open trace in Instruments to verify data exists

swiftuice not found

• Install with: go install github.com/greenstevester/swiftui-cause-effect-cli/cmd/swiftuice@latest
• Or add Go bin to PATH: export PATH=$PATH:$(go env GOPATH)/bin

Source correlation shows 0 matches

• Verify -source points to the Swift project root
• Check that .swift files exist in the directory
• Source correlation uses symbol matching, works best with clear View names

References

For detailed fix patterns with more code examples, see:

• references/fix-patterns.md - Complete fix pattern catalog
• references/issue-types.md - Detailed issue type documentation