skills.rest
EmaRimoldi
EmaRimoldiCOMMUNITY
Explore all skills by this author →
GitHub

compute-planner

Generates SLURM scripts and computes resource allocations from experiment matrices and budgets.

Updated Mar 27, 2026
Search Tags:#slurm#resource-estimation#compute-planning#experiment-design#gpu-scheduling#cluster-scripts#mit-engaging
One-click install
npx skills add https://github.com/EmaRimoldi/Claude-scholar-extended --skill compute-planner
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: compute-planner
Source: https://github.com/EmaRimoldi/Claude-scholar-extended/tree/main/skills/compute-planner
Command: npx skills add https://github.com/EmaRimoldi/Claude-scholar-extended --skill compute-planner

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Plan compute resources and develop deployment plans for run-ready SLURM scripts based on an experiment plan and smoke-test validation timing.

Core Features & Use Cases

  • Resource estimation: translate an experiment-matrix into per-run budgets (GPUs, memory, time) and aggregate totals.
  • Partitioning & scheduling: select MIT Engaging partitions and design dependency chains to optimize throughput.
  • Script generation: produce production-ready SLURM scripts and a master launch script to orchestrate phases.
  • Use Case: After experiment design and validation, generate a complete compute plan to execute the full sweep on the cluster.

Quick Start

Provide an initial compute plan for the given experiment plan and validation outputs.

Frequently Asked Questions about compute-planner

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate SLURM scripts for a GPU experiment sweep?▼

Generate SLURM scripts by providing an experiment plan and smoke-test validation timing to translate the experiment matrix into per-run GPU allocations, partition assignments, dependency chains, and executable batch scripts.

How do I estimate compute resources for a cluster experiment matrix?▼

Estimate compute resources by translating an experiment matrix into per-run budgets including GPUs, memory, and wall-time, then aggregating totals while enforcing defaults like a 20% time margin and single GPU per job.

Can I use compute planning for the MIT Engaging cluster specifically?▼

Yes, compute planning selects MIT Engaging partitions and designs dependency chains to optimize throughput, producing scripts suitable for MIT Engaging, standalone execution, or integration into a larger pipeline.

What's the best way to schedule GPU jobs with dependencies in SLURM?▼

Schedule GPU jobs by designing dependency chains that orchestrate execution phases, assigning a single GPU per SLURM job and outputting a master launch script alongside a compute-plan.md file to manage the workflow.

Do I need validation timing data before planning cluster compute budgets?▼

Yes, validation timing data from a smoke test is required alongside the experiment plan to accurately calculate per-run resource allocations, apply a 20% wall-time margin, and generate runnable batch scripts.

Why does my compute plan assign only one GPU per SLURM job?▼

Assigning one GPU per SLURM job is an enforced default in compute planning to ensure reliable resource allocation and scheduling within the experiment matrix execution on the cluster.

Related Skills

Explore More Education & Research Skills→
Mang30Mang30

experiment-plan

Generate a claim-driven experiment roadmap with run order and compute estimates.

Community
Advanced
Zayne-spragueZayne-sprague
5

run-job

Automate SLURM job execution from sbatch generation to artifact upload.

Community
Advanced
dogekikidogekiki

ablation-planner

Design and manage machine learning ablation studies for model component validation.

Community
Advanced
caseywdunncaseywdunn
2

dunnlab-hpc

Locate Dunn Lab HPC cluster specifications for SLURM batch scripts.

Community
Basic