AgentSkillsCN

profile-code

测量代码的执行时间与内存占用，深入剖析程序的性能特征与瓶颈所在。

SKILL.md
--- frontmatter
name: profile-code
description: "Measure execution time and memory usage of code. Use when analyzing performance characteristics."
mcp_fallback: none
category: analysis
tier: 2

Profile Code

Use profiling tools to measure CPU time, memory allocation, and identify performance bottlenecks in code.

When to Use

  • •Finding performance bottlenecks
  • •Measuring CPU vs memory tradeoffs
  • •Understanding where code spends most time
  • •Optimizing hot code paths

Quick Reference

bash
# Python CPU profiling with cProfile
python3 -m cProfile -s cumulative script.py | head -30

# Memory profiling
pip install memory-profiler
python3 -m memory_profiler script.py

# Detailed call graph
pip install graphviz
python3 -m pstats /tmp/profile.prof

Workflow

  1. •Select profiler: Choose appropriate tool (cProfile for CPU, memory_profiler for memory)
  2. •Run with instrumentation: Execute code with profiling enabled
  3. •Capture metrics: Record timing and memory data
  4. •Analyze output: Identify top time consumers and memory hogs
  5. •Report findings: Document bottlenecks with before/after context

Output Format

Profiling report:

  • •Top functions by execution time
  • •Call count for each function
  • •Memory allocation per function
  • •Call graph/call tree
  • •Percentage of total time/memory per function
  • •Recommendations for optimization

References

  • •See suggest-optimizations skill for improvement recommendations
  • •See benchmark-functions skill for measuring improvements
  • •See CLAUDE.md > Performance for optimization guidelines