Advancing research often requires computational resources that exceed the capabilities of personal laptops or desktop computers. Whether you're processing large datasets, performing complex analyses, or running simulations, efficient computing resources can drastically reduce your time to results and improve your research productivity.
Common scenarios where enhanced computing resources become essential include:
- Computational tasks take too long:
Tasks such as data analysis, simulations, modeling, or visualization can become computationally intensive. When your local resources are insufficient, leveraging High Performance Computing (HPC) clusters or cloud services allows tasks to complete in significantly less time. - One server is not enough:
Complex research tasks often surpass the capabilities of a single computer. HPC clusters like Cypress and LONI combine multiple servers (nodes) into powerful computing environments that distribute workloads efficiently, enabling researchers to handle computational tasks far beyond the scope of a standard single-server setup.
Not sure which resource is best suited to your needs?
Reach out to our Technology Delivery Team—we’re ready to guide you through available options and get you up and running quickly.
Cypress is Tulane University’s High Performance Computing system and is composed of Dell technology, including a high-speed internal network, Intel Ivy Bridge CPU, Intel Xeon Phi Coprocessor, and Lustre file system running on Dell storage with Intel Enterprise Edition for Lustre (IEEL) technology.
| Node Type | C8220X/E5-2680 v2 & Xeon Phi 7120P |
| # of Nodes | 124 |
| CPUs per node | 2 |
| Total E5 2680 v2 Cores | 2,480 |
| Total Xeon Phi cores | 15,128 |
| Tflops CPUs only (theoretical) | 69.44 |
| Tflops system peak w/ Phi (theoretical) | 369 |
| Memory per node | 12 nodes: 256GB each - 25 nodes: 128GB each - 87 nodes: 64GB each |
| Total Memory | 11.84TB |
| Storage | 1.476PB expandable to 4PB |
Access & Account Requests
- All accounts must be associated with a Tulane faculty member who is the Principal Investigator (PI) of his/her research group. Please have your PI contact us to establish a group on Cypress.
- See the wiki section Getting an account.
Cypress In 7 Questions
- Who provided Cypress?
- Cypress was provided by Information Technology for use by the Tulane research community.
- Services upon request: new account, consultation, software installation, storage increase
- See wiki for Getting an account
- What is Cypress?
Tulane's High Performance Computing (HPC) cluster
- When was Cypress Installed?
Cypress was installed in 2014. At that time it ranked #271st supercomputer site in the world by TOP500
- Where is Cypress?
Cypress is located in the Tulane datacenter in Downtown New Orleans
- Why was Cypress built?
Cypress was built to support computational research across the entire university.
- How were Cypress components chosen?
Cypress components were carefully selected for consistent performance over a wide range of scientific applications.
- Where can I find a summary of Cypress’s features?
For a summary of Cypress features, see wiki About Cypress
User Guide & Documentation
See the wiki page Information on how to use Cypress.
LONI offers computational resources to support research activities. It provides access to high-performance computing (HPC) resources for processing and analyzing large volumes of research data. As of 2024, three high-performance computing (HPC) clusters are available.
Tech Specs
- QB2: Operational since 2014, this cluster boasts a peak performance of 1.5 Petaflops. It features over 10,000 CPU cores composed of Intel Ivy Bridge-EP Xeon E5-2680v2 processors and 960 NVIDIA Tesla K20x GPUs, distributed across 504 compute nodes connected by a 54 Gbps InfiniBand fabric. It utilizes a 1 PB Lustre file system and over 30 TB of fast main memory.
- QB3: Operational since 2020, QB3 has a peak performance of 857 Teraflops, with 9,696 CPU cores composed of Intel Cascade Lake Xeon Platinum 8260 Processors spread across 202 compute nodes connected by a 100 Gbps InfiniBand fabric. There are 16 NVIDIA Volta V100 GPUs available. It utilizes a 1.5 PB Lustre file system and over 37 TB of fast main memory.
- QB4: Launched in 2024, QB4 offers a peak performance of 4.3 Petaflops. It comprises 35,008 CPU cores composed of Intel Ice Lake Xeon Platinum 8358 Processors and 144 NVIDIA A100 GPUs, distributed over 547 compute nodes connected by a 200 Gbps InfiniBand fabric. It utilizes a 6.5 PB Lustre file system and over 136 TB of fast main memory.
Access & Account Requests
Account
All Tulane researchers can obtain LONI account with a valid Tulane email address.
https://allocations.loni.org/
Allocation
Only full-time faculty members or research staff including Postdoc can act as Principal Investigators (PI) and request LONI allocations. Everyone else will need to join an existing allocation of a PI, usually advisor/supervision or course instructor (if your course requires a LONI account).
https://www.hpc.lsu.edu/users/loniaccounts-helper.php
User Guide & Documentation
See web page:
https://hpc.loni.org/resources/hpc/index.php
ACCESS is a program established and funded by the U.S. National Science Foundation to help researchers and educators, with or without supporting grants, to utilize the nation’s advanced computing systems and services – at no cost to you.
Tulane University Campus Champions are available to facilitate proposals for these resources.
ACCESS Resources Training & Workshops Events & Trainings | ACCESS Support
Microsoft Azure (Project based, fee-based)
Microsoft Azure is an emerging service available on a project-specific basis with fee-based chargeback as part of an ongoing IT pilot program. Azure provides scalable cloud-based solutions, offering flexible compute options alongside robust storage capabilities. Azure Storage supports large-scale data storage and management through Blob (unstructured data), and File (file shares, enabling customized solutions tailored precisely to your project's computational and data-storage requirements. The Microsoft Azure cloud computing platform that can be useful for complex workloads or those that require segmentation from Tulane’s network. The cost is based on individual usage requirements.
