ASR

Atomic Simulation Recipes (ASR) is a Python framework for defining and executing simulation workflows. Developed at DTU, it allows users to define "recipes" (workflows) that link calculations (typically using ASE and GPAW) to results. AS…

9. FRAMEWORKS 9.5 Specialized Analysis Frameworks VERIFIED 1 paper
Back to Mind Map Official Website

Overview

Atomic Simulation Recipes (ASR) is a Python framework for defining and executing simulation workflows. Developed at DTU, it allows users to define "recipes" (workflows) that link calculations (typically using ASE and GPAW) to results. ASR is the engine behind the C2DB database. It emphasizes caching, reproducibility, and automatic presentation of results via a web interface.

Reference Papers (1)

Full Documentation

Official Resources

  • Homepage: https://asr.readthedocs.io/
  • Documentation: https://asr.readthedocs.io/
  • Source Repository: https://gitlab.com/ase/asr
  • License: GNU General Public License v3.0

Overview

Atomic Simulation Recipes (ASR) is a Python framework for defining and executing simulation workflows. Developed at DTU, it allows users to define "recipes" (workflows) that link calculations (typically using ASE and GPAW) to results. ASR is the engine behind the C2DB database. It emphasizes caching, reproducibility, and automatic presentation of results via a web interface.

Scientific domain: Workflow automation, database generation, reproducibility
Target user community: ASE/GPAW users, C2DB contributors

Capabilities (CRITICAL)

  • Recipes: Modular Python scripts defining calculations (e.g., relax.py, bandstructure.py).
  • Caching: Smart caching of results ("smart make" functionality) to avoid recomputing.
  • Web UI: Automatically generates a web page for browsing results in a directory.
  • CLI: Command line interface (asr run relax).
  • Dependency: Automatic handling of task dependencies.

Sources: ASR documentation, Adv. Theory Simul. 2, 1900080 (2019)

Inputs & Outputs

  • Input formats: ASE atoms objects, CLI arguments
  • Output data types: Pickled result files (results-relax.pkl), JSON

Interfaces & Ecosystem

  • ASE: Core object model.
  • GPAW: Primary calculator supported (but code agnostic in principle).
  • MyQueue: Scheduler integration.

Workflow and Usage

  1. Create a directory for a material.
  2. Run recipes:
    asr run relax
    asr run bandstructure
    
  3. View results: asr browser starts a local web server.

Performance Characteristics

  • Very low overhead
  • Excellent for managing "high-throughput" studies of a moderate number of materials (100s-1000s)

Application Areas

  • Database generation (C2DB)
  • Systematic material studies
  • Reproducible thesis work

Community and Support

  • Developed by CAMD (DTU)
  • Active development

Verification & Sources

Primary sources:

  1. Documentation: https://asr.readthedocs.io/
  2. GitLab: https://gitlab.com/ase/asr

Confidence: VERIFIED

Verification status: ✅ VERIFIED

  • Website: ACTIVE
  • Documentation: COMPREHENSIVE
  • Source: OPEN (GitLab)
  • Development: ACTIVE
  • Applications: Workflows, recipes, caching

Related Tools in 9.5 Specialized Analysis Frameworks