Argonne Leadership Computing Facility

ALCF Resources
Leadership Computing Resources

The ALCF provides users with access to supercomputing resources that are significantly more powerful than systems typically used for open scientific research.
Computing Resources

Aurora

Polaris

Sophia

Crux

ALCF AI Testbed

Evaluation Testbeds

Storage and Networking
Facility Expertise

Facility Expertise
Featured: Aurora
Science and Engineering
Computational Science

The ALCF is accelerating scientific discoveries in many disciplines, ranging from chemistry and engineering to physics and materials science.
Output

Projects

Publications

Case Studies
Allocation Programs

INCITE Program

ALCC Program

Director’s Discretionary

Early Science Program

NAIRR Program
Featured: Engineering
Community and Outreach
Growing the HPC Community

The ALCF is committed to providing training and outreach opportunities that prepare researchers to efficiently use its leadership computing systems, while also cultivating a diverse and skilled HPC workforce for the future.
Partnerships

Industry

Collaborations
Community

NAIRR Pilot

Lighthouse Initiative
Educational Outreach

Women in STEM

Student Programs
About
Accelerating Science

The Argonne Leadership Computing Facility enables breakthroughs in science and engineering by providing supercomputing resources and expertise to the research community.
Get to Know More

Leadership

People

Organizational Chart

User Advisory Council

History
Visit

Visiting ALCF

Tours
Latest

News

Careers
Press Kits

ALCF Media Kit

Aurora Media Kit

Reports Archive
Support
Support Center

The ALCF Support Center assists users with support requests related to their ALCF projects.
Help Desk
Hours: 9:00am-5:00pm CT M-F
Email: mailto:[email protected]
Current

Machine Status

Facility Updates

MyALCF

Training

Training Videos & Slides

Training Overview

Training and Events
Guides

Get Started

Account and Project Management

Data Management

Services

Aurora

Polaris

Sophia

Crux

AI Testbed

Facility Policies
Featured: Get Started

Featured: MyALCF

ALCF User Guides

Home
Account and Project Management
Account and Project Management
- Accounts and Access
- Passcode Tokens
- Accounts and Access FAQs
- Project Management
  Project Management
- Allocations
  Allocations
Data Management
Data Management
- ALCF Community Data Co-Op
- Sharing on Eagle Using Globus
- Transferring Data to Eagle
- Data Transfer
  Data Transfer
  - SFTP and SCP
  - Using Globus
- File System & Storage
  File System & Storage
Services
Services
- Getting Started
- JupyterHub
- Continuous Integration
  Continuous Integration
  - General
  - Gitlab-CI
Running Jobs with PBS at the ALCF
Running Jobs with PBS at the ALCF
Aurora
Aurora
- Getting Started
- Aurora PE
- Sunspot to Aurora
- Known Issues
- Hardware Overview
- Node Performance Overview
- Compiling and Linking
  Compiling and Linking
- Build Tools
  Build Tools
  - CMake
- Running Jobs
- Data Management
  Data Management
  - Copper
  - DAOS
  - Lustre (Flare)
  - Moving data to Aurora
    Moving data to Aurora
    
    DAOS Data Mover
    
    Globus
    
    SCP
- Applications and Libraries
  Applications and Libraries
  - Libraries
    Libraries
    
    Cabana
    
    Spack PE
- Data Science
  Data Science
  - Python
  - Containers
  - Frameworks
    Frameworks
    
    PyTorch
    
    TensorFlow
    
    LibTorch
    
    OpenVINO
    
    oneCCL
- Programming Models
  Programming Models
  - Kokkos
  - Level Zero
  - OpenCL
  - OpenMP
  - SYCL
- Debugging Tools
  Debugging Tools
  - Overview
- Performance Tools
  Performance Tools
  - Advisor
  - VTune
- Services
  Services
  - GitLab
- Workflows
  Workflows
  - SmartSim
Polaris
Polaris
- Getting Started
- System Updates
- Known Issues
- Hardware Overview
  Hardware Overview
  - Polaris Machine Overview
- Compiling and Linking
  Compiling and Linking
- Build Tools
  Build Tools
  - CMake
- Running Jobs
- Applications and Libraries
  Applications and Libraries
  - Applications
    Applications
    
    Amber
    
    GROMACS
    
    LAMMPS
    
    NAMD
    
    NekRS
    
    OpenMM
    
    QMCPACK
    
    Quantum ESPRESSO
    
    VASP
  - Libraries
    Libraries
    
    Math Libraries
    
    Cabana
    
    Spack PE
    
    NCCL
    
    XALT
- Containers
- Data Science
  Data Science
  - Julia
  - Python
  - Frameworks
    Frameworks
    
    TensorFlow
    
    PyTorch
    
    JAX
    
    DeepSpeed
    
    LibTorch
  - Applications
    Applications
    
    Megatron-DeepSpeed
    
    gpt-neox
- Programming Models
  Programming Models
  - OpenMP
  - SYCL
  - Kokkos
- Debugging Tools
  Debugging Tools
  - CUDA-GDB
- Performance Tools
  Performance Tools
  - NVIDIA-Nsight
- Visualization
  Visualization
- Workflows
  Workflows
Sophia
Sophia
- Machine Overview
- Getting Started
- Running Jobs
- Compiling and Linking
- Containers
- Data Science
  Data Science
  - Python
  - Fine-tuning with Autotrain
- Visualization
  Visualization
  - Visualization on Sophia
  - ParaView (Launch from Client)
Crux
Crux
- Machine Overview
- Getting Started
- Running Jobs
- Compiling and Linking
- Containers
- Data Science
  Data Science
  - Python
AI Testbed
AI Testbed
- Getting Started
- Cerebras
  Cerebras
- Graphcore
  Graphcore
- Groq
  Groq
- SambaNova
  SambaNova
- Data Management
Facility Policies
Facility Policies
- Overview of Policies
- ALCF Acknowledgement Policy
- Account Policies
  Account Policies
- Queue and Scheduling Policies
  Queue and Scheduling Policies
- Data and Software Policies
  Data and Software Policies
  - Data Policy
  - Software Policy
Docs Questions/Issues

Flare Filesystem

Flare is a 91 PB Lustre Filesystem with 160 OSTs, 40 MDTs and 48 Gateway nodes mounted at /lus/flare/projects/ with a peak theoritical performance of 650GB/s. You should launch jobs only from this flare space.

Home is 12 PB Gecko Lustre Filesystem with 32 OSTs and 12 MDTs.

Follow this link for more basic information on I/O optimization for the Lustre Filesystem I/O