Write, debug, and optimize SLURM job scripts for HPC clusters.
npx skills add https://github.com/michaelrizvi/claude-config --skill slurm-michaelrizvi
Please help me install this Agent Skill. Skill: slurm Source: https://github.com/michaelrizvi/claude-config/tree/main/skills/slurm Command: npx skills add https://github.com/michaelrizvi/claude-config --skill slurm-michaelrizvi
This Skill helps researchers and engineers write, debug, and optimize SLURM job scripts for HPC clusters, reducing time-to-submit and minimizing common resource-related errors.
Create a simple sbatch script that requests 1 GPU for 4 hours and explain how to submit it to the cluster.
High-intent search queries and answers about installing and using this skill.
To write a SLURM job script for GPU training, you define job directives and resource requests like CPUs, GPUs, and RAM in an sbatch script. This ensures proper resource allocation and includes safety checks for your HPC cluster submission.
Yes, you can set up dependencies between multiple SLURM sbatch jobs to manage multi-step training workflows. This involves configuring job directives within your scripts to handle execution order and manage failure handling across the cluster.
Resource allocation for SLURM salloc and squeue workflows involves specifying CPU, GPU, and RAM requests while considering fair-share guidelines. Proper validation prevents common resource-related errors and streamlines job execution on HPC clusters.
SLURM job submission often fails due to misconfigured resource requests or missing safety checks in your job scripts. Debugging with commands like sacct and nvidia-smi helps identify allocation issues and minimize these errors on HPC clusters.
The best way to debug a failed SLURM job script is to use built-in debugging commands like sacct to check job accounting data and nvidia-smi for GPU status. This identifies resource bottlenecks and optimizes your HPC cluster workflows.
yale-som-hpcFetch WRDS data, query REST APIs, and scrape sources on the Yale SOM HPC cluster with secure credentials.
carbonscottAnswer Slurm questions on SLAC's S3DF cluster with read-only commands.
dongzhuoyaoSubmit GPU training jobs to Slurm clusters with preflight validation and monitoring.
googleAutomate batch and HPC workload deployment on Google Kubernetes Engine.
