Guide for benchmarking FlashInfer kernels with CUPTI timing
Install
npx skills add flashinfer-ai/flashinfer --skill benchmark-kernelnpx skills add flashinfer-ai/flashinfer --skill benchmark-kernel -a claude-code -gnpx skills add flashinfer-ai/flashinfer --skill benchmark-kernel -a codex -ggemini skills install https://github.com/flashinfer-ai/flashinfer.git --path .claude/skills/benchmark-kernelgh skill install flashinfer-ai/flashinfer benchmark-kernelgit clone --depth 1 https://github.com/flashinfer-ai/flashinfer /tmp/flashinfer-ai__flashinfer
mkdir -p ~/.claude/skills ~/.agents/skills
cp -R /tmp/flashinfer-ai__flashinfer/.claude/skills/benchmark-kernel ~/.claude/skills/benchmark-kernel # Claude Code
cp -R /tmp/flashinfer-ai__flashinfer/.claude/skills/benchmark-kernel ~/.agents/skills/benchmark-kernel # Codex, Cursor, Gemini CLI, Copilot, OpenCode