
Performance Profiling & Bottleneck Analysis
Analyze game performance bottlenecks from profiler data and telemetry logs
3.8(32 reviews)100+ downloadsUpdated Sep 2026Verified SafeSecurity VerifiedThis skill was analyzed by our AI security scanner for harmful content including data exfiltration, system manipulation, credential theft, and prompt injection. No threats were detected.
What You Can Do
You can transform raw profiler data from Unity Profiler, Unreal Insights, PIX, or NVIDIA FrameView into structured performance analysis reports. This skill helps you isolate root causes of frame time spikes, distinguish transient dips from systemic issues, and create optimization priority matrices that guide developer focus. You'll generate reproducible test cases and validation metrics that accelerate performance fix implementation.
Features
Parse profiler output to isolate CPU/GPU function calls, draw calls, and memory allocations causing frame time spikes
Distinguish between transient performance dips and systemic bottlenecks using statistical analysis of telemetry
Generate reproducible test cases that reliably trigger identified performance issues
Create optimization priority matrices ranking issues by impact, frequency, and fix complexity
Compare performance across builds and platforms with baseline metrics and regression detection
Produce QA reports with actionable recommendations mapped to specific code sections or systems
Validate performance fixes against pre- and post-optimization telemetry data
Example Output
Input: Frame profiler log showing 45ms frame spike with 32ms in DrawCall batch and 10ms in physics.simulate()
Output:
code
Bottleneck Priority Report
1. CRITICAL: GPU Draw Call Batching
- Current: 2,847 draw calls/frame
- Baseline: ~1,200
- Root Cause: Excessive material state changes in vegetation system
- Recommendation: Implement material atlasing
- Estimated Impact: -15ms frame time
- Reproducible Case: Load Forest_Arena scene, pan camera across trees
2. HIGH: Physics Simulation
- Current: 10ms/frame (physics.simulate)
- Spike Pattern: Occurs when 40+ dynamic bodies active
- Recommendation: Implement physics LOD for distant objects
- Estimated Impact: -4ms frame time
- Reproducible Case: Run PhysicsStress test at 60 active rigidbodies
3. MEDIUM: Memory Allocation Spikes
- Current: 2.3MB allocations/frame (GC pressure)
- Recommendation: Object pooling for projectiles and effects
What's Included
- SKILL.md instruction file with structured analysis workflows:
- Profiler Data Analysis Template for parsing Unity, Unreal, and PIX output formats:
- Performance Bottleneck Classification Matrix (CPU/GPU/Memory/I/O taxonomy):
- Optimization Priority Checklist for ranking issues by impact and complexity:
- Reproducible Test Case Framework with setup steps and validation metrics:
- Performance Report Template with baseline comparison and recommendation structure:
Who It's For
- Game QA Testers — Translate profiler data into actionable developer feedback
- Performance Engineers — Systematize performance analysis and bottleneck prioritization
- Technical Leads — Generate performance reports for sprint planning and resource allocation
- Build/Optimization Engineers — Validate performance fixes and track regressions across builds
- Game Producers — Understand performance blockers and priority trade-offs for scheduling
Best For
- Analyzing frame time spikes and stuttering during gameplay
- Prioritizing multiple performance issues for developer teams
- Investigating CPU, GPU, memory, and I/O bottlenecks from telemetry
- Creating and validating performance regression baselines across builds or platforms
- Generating QA reports with reproducible test cases and specific optimization recommendations
You might also like
$45.00







