matlab-optimize-performance

What problem does it solve?

This Skill helps you systematically optimize MATLAB performance without guessing, so you can find true bottlenecks and avoid incorrect or unverified speedup claims.

Core Features & Use Cases

• 7-step performance optimization workflow: baseline measurement, profiling, bottleneck identification, targeted optimization, re-measurement, correctness verification, and reporting.
• Robust MATLAB timing guidance: emphasizes using timeit for warmup/averaging and appropriate GPU timing practices (e.g., gputimeit concepts).
• Profiling-first decision making: focuses optimization effort on self-time hotspots and high-impact bottlenecks.
• Scope guardrails: clarifies when not to use this workflow (e.g., Simulink simulation profiling, compiled MEX/C bottlenecks, purely I/O-bound issues) and points you to the right alternative for performance testing.

Quick Start

Use the matlab-optimize-performance workflow to measure baseline timing, profile to locate bottlenecks, implement one targeted change, then verify correctness and confirm end-to-end speedup.