
Unreal Engine C++ Performance Profiling & Optimization
Profile and optimize Unreal Engine C++ bottlenecks with AI-powered analysis
What You Can Do
This skill turns raw Unreal Engine profiler data into actionable performance improvements. Paste your profiling results, frame stats, or code metrics, and Claude analyzes the data like a performance engineer—identifying bottlenecks, explaining root causes, and recommending specific optimizations. Works with Insight profiler dumps, Stat reports, RenderDoc data, and custom performance metrics.
Features
Parse stat commands and identify rendering, physics, and game logic bottlenecks
Detect leaks, fragmentation, and asset bloat patterns with optimization targets
Pinpoint CPU-intensive functions and their call chains from profiler dumps
Get targeted, prioritized fixes with estimated performance gains
Optimize workload distribution and identify graphics bottlenecks
Diagnose slow asset streaming and initialization sequences
Compare performance characteristics and suggest C++ conversions
Extend profiling to game-specific counters and KPIs
Example Output
Frame Rate Analysis
Input: Stat FPS 45 fps, 22ms per frame (18ms CPU, 4ms GPU)
Output:
- ⚠️ CPU bottleneck (18ms of 22ms budget)
- Main thread: 12ms in GetAllActorsOfClass queries
- Physics: 4ms in FPhysScene::Simulate
- GC spikes: 2ms (irregular, mid-frame)
Recommendations:
1. Cache actor queries (saves ~8ms) — HIGH impact/EASY effort
2. Reduce physics actors 30% via spatial partitioning (saves ~1.2ms)
3. Defer GC to loading screens (eliminates 2ms spikes)
Estimated gain: +15 fps (22ms → 15ms)
Memory Optimization
Input: 2.5 GB total, textures 1.2GB, skeletal meshes 0.6GB
Output:
Quick wins (36% total reduction):
- Enable texture streaming: −300 MB
- Consolidate duplicate skeletons: −400 MB
- Stream audio instead of loading: −200 MB
Target achieved: ~1.6 GB (within budget)
Code Hotspot Analysis
Input: Profiler top function: ACharacter::Tick (18% CPU)
Output:
Root cause: Full IK animation LOD disabled every frame
Fix: Enable animation LOD 1 at distance (disables IK)
Expected: 8% CPU reduction, easy implementation
Priority: HIGH
What's Included
- SKILL.md: Complete performance analysis and optimization workflow
- Profiler data parsing templates: Unreal Insight CSV, Stat dumps, RenderDoc exports
- Performance optimization checklist: CPU, GPU, memory, I/O, streaming
- Common bottleneck patterns guide: Physics, animation, rendering, garbage collection
- C++ code review template: For performance-critical code candidates
- Optimization verification checklist: Before/after measurement methodology
- Case study templates: Documenting optimizations for team handoff
Who It's For
- C++ gameplay programmers optimizing game logic and actor queries
- Graphics programmers reducing frame time and GPU stalls
- Performance engineers conducting optimization sprints and regression analysis
- Technical leads reviewing bottlenecks and prioritizing fixes
- Engine programmers tuning core systems and subsystems
Best For
- Diagnosing frame rate stutters and identifying CPU vs GPU bottlenecks
- Memory usage reduction from profiler snapshot analysis
- Optimizing hot code paths identified in profiler output
- Planning refactors by identifying low-hanging fruit and quick wins
- Performance regression analysis to find what changed
- Threading and load balancing recommendations for CPU-bound code






