Skip to main content

NSF Access

ACCESS is an advanced computing and data resource program supported by the U.S. National Science Foundation (NSF). Please refer to https://allocations.access-ci.org/ for more information on how to get access to NSF ACCESS resources.

note

Allocations are absolutely free of cost and you do not need NSF, or funding from any agency to receive one.

Here are some useful links to get started:

Recommended Resources for CIROH projects:

JetStream2​

Jetstream2 is a powerful hybrid-cloud platform designed for researchers and educators. It offers a range of flexible, on-demand, and programmable infrastructure tools, from interactive virtual machines (VMs) to advanced infrastructure and orchestration services. The primary resource consists of AMD Milan 7713 CPUs with 128 cores per node and 512 GB RAM per node, all connected by a high-speed 100 gbps wthernet.

jetstream2 logoImage Source: https://docs.jetstream-cloud.org/

Use cases:​

Jetstream2 is ideal for researchers with diverse needs:

  • On-demand virtual machines: It is ideal for research that requires on demand virtual machine services. It is also best for researchers needing to create their own customized virtual machine environment for specific software needs.
  • Always-on research infrastructure: It can host research-supporting infrastructure services that require continuous operation.
  • Educational support: It can be used to provide virtual machines for student use in research or coursework.

For information about available instance sizes, visit the JetStream2 VM Sizes page.

To get started with JetStream2:​

info

For a more detailed information on JetStream2, visit the official NSF ACCESS Jetstream2 website here.

Anvil​

Anvil is a powerful supercomputer, offering computing power for demanding research problems. Purdue's Anvil cluster consists of 1,000 nodes with two 64-core AMD EPYC "Milan" processors each and delivers over one billion CPU core hours each year. With a peak performance of 5.1 petaflops and a speed of 100 Gbps interconnect, Anvil ensures rapid data transfer and processing for efficient research workflows. Standard Anvil nodes have 256GB of DDR4-3200 memory each, ideal for most research tasks.

jetstream2 logoImage Source: https://www.rcac.purdue.edu/anvil

Use cases:​

  • General-Purpose CPU Power: Anvil's powerful CPUs (with 128 cores per node) are ideal for computationally intensive tasks suitable for modeling and simulation across scientific and engineering fields.
  • Memory-Intensive Workloads: The dedicated large memory nodes (with 1TB of DDR4-3200 memory per node) works bestfor research that demands significant memory resources.
  • Composable Subsystem: It is a private cloud built on Kubernetes and consists of bothe CPU and GPU nodes and S3 data storage. It is suitable for applications such as model inference service (via NVIDIA Triton), Specialized LLMs, dataset hosting, science gateways and web application hosting, and classroom and training applications via interactive access interfaces.
Anvil Documentation
info

For a more detailed information on Anvil, visit the official NSF ACCESS website here.

NSF NCAR Derecho​

NCAR's Derecho supercomputer is a high-performance computing system with 19.87 petaflops of processing power. It comprises 2,488 nodes, each equipped with two 64-core AMD EPYC 7763 Milan processors, totaling 323,712 processor cores. Each node has 256 GB of DDR4 memory and is interconnected by an HPE Slingshot v11 high-speed network in a dragonfly configuration. The system supports Earth system science research for U.S. institutions.

Use Cases:​

  • Earth System Science Research: Designed specifically for tasks related to climate modeling, weather prediction, and environmental studies.
  • Large-Scale Simulations: Optimal for computationally demanding research in Earth sciences that requires extensive core-hours.
  • Educational Use: Available for classroom and instructional purposes, supporting students, postdocs, and new faculty without external funding.

To learn more, visit the NCAR HPC Documentation page.

NSF NCAR HPC Documentation
info

For additional information on NSF NCAR Derecho and its allocation process, refer to the NSF ACCESS Resources page.