Performance Tuning & Workstation Optimization

DevDiff is optimized to run efficiently alongside resource-heavy IDEs, Docker containers, and local compilers. This guide details techniques for tuning execution speed, memory footprint, and token budgeting on your development workstation.

---

⚡ Performance Optimization Checklist

1. Enable Fast-Path Memory Queries

DevDiff builds a persistent codebase index (.devdiff/memory/codebase-index.json). Querying persistent memory avoids re-indexing unchanged source files, reducing response latency from 4.5s down to < 50ms.

# Initialize memory index once per workspace
devdiff memory init

2. Configure .devdiffignore File Filters

Exclude heavy build output, lockfiles, and generated assets from AST processing:

# Exclude vendor & build output
dist/
build/
node_modules/
*.log
package-lock.json
pnpm-lock.yaml

3. Tune AI Model Selection

• Use lightweight models (ollama://llama3.2:3b or openai://gpt-4o-mini) for routine pre-commit checks.
• Reserve larger models (anthropic://claude-3-5-sonnet or gemini://gemini-1.5-pro) for major release merges.

4. Configure IDEGuardian Memory Ceiling

In VS Code, adjust background worker thread memory settings in .devdiff/config.json:

{
  "performance": {
    "memoryCapMb": 256,
    "idleDetectionSeconds": 5,
    "taskTimeoutSeconds": 120
  }
}