Write EmbodiChain benchmark scripts measuring runtime, CPU/RAM, GPU/VRAM, and correctness.
npx skills add https://github.com/DexForce/EmbodiChain --skill benchmark-dexforce
Please help me install this Agent Skill. Skill: benchmark Source: https://github.com/DexForce/EmbodiChain/tree/main/skills/benchmark Command: npx skills add https://github.com/DexForce/EmbodiChain --skill benchmark-dexforce
This Skill solves the problem of producing consistent, accurate benchmark scripts so you can quantify speed, memory usage, and quality for EmbodiChain modules.
Ask the AI to generate a benchmark script that measures throughput and memory for a specific EmbodiChain component and writes a single Markdown report with the required three-table format.
High-intent search queries and answers about installing and using this skill.
To benchmark EmbodiChain modules, generate a script that uses time.perf_counter for consistent timing, tracks memory via psutil and torch CUDA peak tracking, and outputs a single Markdown report with three structured tables.
Yes, you can compare CPU vs GPU performance for solvers, samplers, and FK/IK routines by generating benchmark scripts that evaluate both implementations and report performance and correctness metrics side by side.
The required Markdown format for EmbodiChain performance profiling reports consists of exactly three tables: Time and Memory metrics, Success and Other Metrics, and a success-rate-sorted Leaderboard.
When CUDA hardware is unavailable, benchmark scripts apply skip-safe execution to bypass GPU tests gracefully while still measuring CPU and RAM behavior, ensuring the benchmark completes without hardware errors.
Yes, EmbodiChain benchmarking requires Apache 2.0 headers in all generated benchmark scripts to maintain consistent licensing conventions across solvers, samplers, metrics, and data-pipeline components.
The best way to profile data-pipeline components and metrics in EmbodiChain is to use benchmark scripts that quantify runtime, memory, and correctness, applying time.perf_counter for timing and psutil for memory tracking.
igamenovoerConfigure CUDA build environments for Pixi projects with CMake and nvcc.
erencodingFetch real-time AI model rankings from Artificial Analysis using Playwright.
dz306271098Profile CPU, GPU, memory, or interconnect targets and generate structured performance reports.
Omar-ObandoProfile application layers and validate optimizations with before/after benchmarks.
