Skip to end of metadata
Go to start of metadata

You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 4 Next »

Why Purchase Resources?

Koa is free to use for all students, faculty and staff at any of the campuses in the University of Hawaii system.

While a significant amount of work can be completed using free community resources, some researchers may find that they need more storage or dedicated resources to run longer jobs than the community nodes allow or they find the wait times for community resources to significantly impede their research.

Faculty and staff can purchase or lease compute nodes to power their research and select a long term storage solution for their data.


Computational resources

Researchers have the options of purchasing or leasing computational resources on Koa. Both options provide the following benefits:

  • A private scheduler partition

  • Priority access to those that have been granted access to the lab private partition

    • Private partition jobs will have immediate access compared to community users

  • Maximum of a 30-day job run time

Purchasing computational resources (hardware)

Pre-configured & Custom System Options

When purchasing resources to add to Koa, researchers are able to choose from a set of pre-configured systems that the Information Technology Services Cyberinfrastructure (ITS-CI) team has curated and acquired estimated pricing for.

Pre-configured systems

In some cases, these systems may not meet a researcher's specific needs. In this case, the ITS-CI team is willing to work with the researcher to try and configure something custom that will work with Koa and work for the researcher.

Fiscal & End of Life Options

Purchasing computational resources has two different fiscal components to it. (1) The majority of the cost is classified as equipment, while (2) a small amount of the cost will be classified as a service. The hardware being purchased is not something unique to Koa, allowing researchers the freedom to utilize the hardware outside of Koa after it is used on Koa if they so choose, but the service fee we charge is used to pay for the use of the small amount of networking ports, cables, and data center infrastructure your need would need to live 5 years as part of Koa. Computational resources purchased by a researcher would have three different options once it has been part of Koa for 5 years. The end of life options of (1) Pay to Stay, (2) Donation, and (3) Removal are described here: Computational Resources End of Life

Purchasing Process

This full purchasing process for nodes on Koa can be found here: Node purchasing process, and the node purchasing intent form can be found here: Node purchasing intent form.

Leasing computational resources (service)

In some cases, a researcher just needs a short period of dedicated access to hardware on Koa or is unable to spend a large amount of funds at once. In this case, a service model of leasing computational resources are available to researchers.

Due to the limited supply in Koa, we do not currently lease GPU nodes

Nodes for lease can range from 128 GB of ram with 20 cores up to nodes that have 1.4 TB of ram with 184 cores. Currently our pricing model is as follows:

Node Type

Memory

CPU Cores

CPU type

MPI capable

Cost per month

Basic - Legacy

128GB

20

Intel Ivy-Bridge

No

$110

Basic

128GB

16

AMD Genoa

Yes

$110

Large memory

512 GB

64

AMD Genoa

Yes

$440

XL memory

1TB

128

AMD Genoa

Yes

$880

2XL memory

1.4 TB

190

AMD Genoa

Yes

$1,300

To lease computational resources, please fill out the order form here:


Storage Resources

Koa Research Storage Service

Accelerate your research and store your data on the cluster for faster access.

The University of Hawaiʻi (UH) Koa Research Storage Services offers three different storage solutions for researchers. Find which storage solution fits your research storage needs at the link below.

https://datascience.hawaii.edu/koa-research-storage-service/

  • No labels