How does it help with memory-constrained environments?

It calculates available RAM and recommends out-of-core computing strategies using libraries like Zarr or H5py if your dataset size exceeds 50% of available memory.

Does this skill support Apple Silicon GPUs?

Yes, it fully detects M-series chips and provides specific recommendations for using the Metal Performance Shaders (MPS) backend with PyTorch and JAX.

When should I run this skill in my workflow?

It is best practice to run this at the start of any computationally intensive task or project initialization to inform your architectural and library choices.

What dependencies are required to run the detection?

The detection script requires the 'psutil' Python package; all other functionality relies on the Python standard library.

Scientific Resource Detector

Name: Scientific Resource Detector
Author: Zehong-Wang

byZehong-Wang

0•

데이터 과학 및 ML

Detects hardware capabilities and provides optimized computational strategies for resource-intensive scientific tasks.

The Scientific Resource Detector skill automates the identification of system hardware, including CPU cores, GPU backends (NVIDIA CUDA, AMD ROCm, Apple Silicon Metal), and memory constraints. It generates a detailed JSON report with actionable recommendations for high-performance computing, such as determining when to use out-of-core processing with Dask, selecting the correct GPU acceleration backend for PyTorch or JAX, and optimizing parallel worker counts to prevent system bottlenecks during large-scale data analysis.

주요 기능

01Memory-aware strategy suggestions for out-of-core computing (Dask, Zarr)

02Standardized JSON output for resource-informed architectural decisions

030 GitHub stars

04Smart recommendations for parallel processing and worker count optimization

05Automated identification of Apple Silicon, NVIDIA, and AMD GPU backends

06Cross-platform detection of CPU, GPU, RAM, and disk availability

사용 사례

01Calculating the most efficient number of workers for parallel joblib tasks

02Identifying the optimal GPU backend and library for training neural networks

03Determining if a 50GB dataset requires Dask/Zarr or can fit in local memory

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add zehong-wang/kosmos get-available-resources

For use in Claude.ai and ChatGPT

Download Skill