Rates

The number and size of jobs allowed on Oscar vary with both partition and type of user account. All Brown users can request a free exploratory account with basic access to a resources on our Batch and GPU partitions. Researchers with more intensive computing needs can upgrade to priority accounts, which provide access to more GPU & CPU cores, memory, walltime and faster start times.

Batch Partition

The batch partition is designed for high-throughput computing jobs that require many CPU cores. Upgrading to priority accounts allows users to run larger jobs with more CPU cores, memory, walltime, faster start times, and limited standard GPUs.

GPU Partition

The gpu partition is designed for jobs that prioritize GPUs. The nodes are equipped with our standard GPUs such as NVIDIA Quadro RTX or A5000. This partition is ideal for standard parallel workloads, model training, and GPU-accelerated visualization. Upgrading to priority accounts allows users to run larger jobs with more GPU & CPU cores, memory, walltime and faster start times.

High-End GPU Partition

The gpu-he partition provides access to GPUs with higher memory and compute performance such as NVIDIA Tesla V100 or A6000 GPUs. Recommended for large-scale deep learning, molecular dynamics, and scientific simulations requiring high memory or compute throughput.

Note, these values are subject to periodic review and changes. For more technical details, review the system hardware documentation.

*Max Resource Walltime: The time limit when using the maximum of available CPUs or GPUs of an account. Using less of the allotted CPUs or GPUs allows for longer walltimes. This is because the resource allocation is actually governed by a fixed total compute budget measured in CPU-minutes or GPU-minutes that is distributed across your concurrent jobs. For example, HPC Priority accounts have a compute budget of 19,968 core-hours. Using all 208 cores results in a walltime of 96 hours (19,968 / 208); however using half of the available cores with 104 CPUs would result in double the walltime at 192 hours (19,968 / 104).

A condo is an account that gives users priority access to computing resources, specifically CPU cores and GPU resources. Investigators may purchase condos to share access to computing resources with their team.

Benefits of Condo Ownership

• 6 year life cycle - condo resources will be available for a duration of 6 years.
• Access to more cpu cores than purchased - condo will have access to 1.25 times the number of cpu cores purchased.
• Support - CCV staff will install, upgrade and maintain condo hardware throughout its life cycle.
• High job priority - Jobs submitted to a condo have the highest priority to schedule.

Brown Sponsored Allocation

CCV provides several storage options for research data. Brown Faculty members are allocated 1TB of default storage. With an active grant, researchers are allocated an additional 10TB on replicated storage or 20TB of non-replicated storage.

Additional Research Storage

Additional research storage can be purchased. Purchased storage is available in two types: Non-Replicated (a single copy) or Replicated (multiple copies at different locations).

Pooled Storage Allocations

We are able to accommodate group payment plans where individual and grant allocations are pooled under an umbrella such as a group, center, department or institute.

To set up a group payment plan, please provide the following information:

• A list of researchers in the group
• A list of grants with end dates for any PIs in the group
• Who will handle the invoicing

Have questions about our resources & rates or ready to upgrade your account? CCV can help! We offer consultations to determine the most cost-effective resources for your research needs.