matlab/matlab-agentic-toolkit / skills-catalog/code-generation/matlab-optimize-gpu-codegen
matlab-optimize-gpu-codegen
UnverifiedOptimize MATLAB design files for GPU Coder to generate faster CUDA code. Iteratively profiles, rewrites, and benchmarks until performance targets are met or diagnostics are resolved. Use when asked to: optimize for GPU Coder, improve GPU codegen performance, profile generated GPU/CUDA code, profile GPU MEX, fix gpuPerformanceAnalyzer diagnostics, speed up GPU MEX, reduce GPU memory transfers, improve kernel parallelism, rewrite MATLAB for CUDA, or run gpuPerformanceAnalyzer.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 18.4 KB |
| manifest.yaml | config | 523 B |
| references/gpu-codegen-functions.md | markdown | 6.4 KB |
| scripts/benchmarkMex.m | other | 2.5 KB |
| scripts/extractDiagnostics.p | binary | 1.2 KB |
scripts/benchmarkMex.m
View on GitHubfunction result = benchmarkMex(mexName, inputs, opts)
%benchmarkMex Convergence-based GPU MEX benchmark using gputimeit.
%
% result = benchmarkMex(mexName, inputs) measures end-to-end execution
% time of a GPU Coder MEX function using gputimeit in an adaptive loop
% that stops when measurements converge.
%
% result = benchmarkMex(mexName, inputs, WarmupIter=5, Tolerance=0.03)
% overrides default parameters.
%
% Inputs:
% mexName - Name of the MEX function (string)
% inputs - Cell array of input arguments to pass to the MEX
%
% Output struct fields:
% result.MedianTime - Median of converged measurements (seconds)
% result.AllTimes - All recorded gputimeit values (seconds)
% result.NumIterations - Total number of gputimeit calls made
% result.Converged - true if CV dropped below tolerance before MaxIter
%
% Example:
% result = benchmarkMex("myDesign_mex", {randn(256), randn(256,1)});
% fprintf("Time: %.4f ms\n", result.MedianTime * 1000);
arguments
mexName (1,1) string
inputs (1,:) cell
opts.WarmupIter (1,1) double {mustBePositive, mustBeInteger} = 3
opts.WindowSize (1,1) double {mustBePositive, mustBeInteger} = 5
opts.MaxIter (1,1) double {mustBePositive, mustBeInteger} = 50
opts.Tolerance (1,1) double {mustBePositive} = 0.02
end
fcn = @() feval(mexName, inputs{:});
% Warmup: lazy init, cuBLAS/cuDNN handle creation, memory pool allocation
for i = 1:opts.WarmupIter
fcn();
end
wait(gpuDevice);
% Convergence-based measurement
times = [];
window = [];
while true
t = gputimeit(fcn);
times(end+1) = t; %#ok<AGROW>
window(end+1) = t; %#ok<AGROW>
if numel(window) > opts.WindowSize
window(1) = [];
end
if numel(times) >= opts.MaxIter
break;
end
if numel(window) >= opts.WindowSize
cv = std(window) / mean(window);
if cv <= opts.Tolerance
break;
end
end
end
result.MedianTime = median(times);
result.AllTimes = times;
result.NumIterations = numel(times);
result.Converged = numel(times) < opts.MaxIter;
fprintf("Benchmark: %.4f ms (median, %d iterations, %s)\n", ...
result.MedianTime * 1000, result.NumIterations, ...
string(tern(result.Converged, "converged", "hit max")));
end
function out = tern(cond, a, b)
if cond
out = a;
else
out = b;
end
end
% Copyright 2026 The MathWorks, Inc.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| brainstormingobra/superpowersYou MUST use this before any creative work - creating features, building components, adding functionality, modifying behavior, or planning anything new, in software or out of it (a talk, a business, a renovation). | Stars: 318k | Updated | Claude CodeCodex | Verified | Install |
| brand-guidelinesanthropics/skillsApplies Anthropic's official brand colors and typography to any sort of artifact that may benefit from having Anthropic's look-and-feel. Use it when brand colors or style guidelines, visual formatting, or company design standards apply. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |
| frontend-designanthropics/skillsGuidance for distinctive, intentional visual design when building new UI or reshaping an existing one. Helps with aesthetic direction, typography, and making choices that don't read as templated defaults. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |
| mcp-builderanthropics/skillsGuide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK). | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |