CRID Systems and Compute

CRID Team supports multiple platforms, departments for your research computing needs. 

1500+

Hosted Virtual Machines

3

Cloud Service Providers

103000+

Average compute hours per month

Get Help

CRID Team Consultation

Systems and Research Computing Services 

Includes data center hosting services, resource computing, and general server support and administration.

Areas of Service

  • Supported Configurations

    Type:

    Rack Hosted Physical Server(s)

    Virtual Servers

     

    OS Configurations:

    Windows

    • 2019
    • 2022

     

    Linux Supported OS Configuration:

    • Ubuntu
    • Rocky
    • Oracle Linux
  • Features
    • Highly available
    • Scalable
    • Backups available upon request
    • Supported/managed by systems administrators
    • Customizable
  • Service Expectations and Limits

    Max Total Allowable Size for any Single VM:

    vCPU Cores: 12 vCPU

    Memory: 32 GB

    Disk: 2TB

    *Cluster builds available see HPC Consultation

Request this service

  • What's GENIE HPC?

    The GENIE cluster is the first university-wide HPC. It is funded by the Biggs Institute and the Geriatric Research, Education and Clinical Center (GRECC) from Scientific Talent Acquisition and Retention (STARs) awards from Univerity of Texas Systems. Beyond supporting research projects from Biggs Institute and GRECC, the cluster will also provide computational access to researchers from other departments, thereby fostering collaborative partnerships with researchers and each other.

  • HPC Shared Computing Cluster

    Condo Computing:

    This provides the researcher with much greater flexibility than owning a standalone cluster. The Shared Compute Clusters are shared between researchers who buy nodes for the cluster. IMS provides all of the redundant infrastructure, data center facility, core infiniband, redundant head nodes, login nodes, spare nodes, and cluster management. The use of an annual compute-core-hours allocation resulting from their equipment purchases for a period of five years. Researchers may divide their compute-core hours allocation among members of their research group. 

     

    More Information

  • GENIE HPC Support Model
    • Cluster management, proactive monitoring, and support
    • Implement industry best practices
    • Triage and vendor management for hardware and licensed software (Bright management)
    • Training for job, user and queue management
    • Setup test queue to allow image testing
    • Online documentation
    • Monthly or Quarterly utilization reporting
    • Base OS, Scheduler, Cluster Manager (CMS) support and fixes
    • Recommendation for improvement

Request this service

  • Platforms

    VMWare Horizon

    Azure Virtual Desktop

    Citrix

  • Service Options
    • Persistent
    • Non-Persistent
    • RDS Published Application  
  • Benefits and Features
    • Flexibility - Run legacy and homegrown applications from any device; run Windows-only applications on non-Windows devices; specialized software can be added, updated, or removed quickly
    • Mobility - Access from anywhere, any device
    • Availability - Redundant power, network, and systems datacenter
    • Security - Allow secure access to sensitive data; remove content from the endpoint device; reduce threats from theft or compromise from the client device
    • Manageability - Simplify desktop support and management; manage automated upgrades, patches, and version control
    • Extend the life of older hardware - Users don't need to buy their own copies of software to be installed on their devices

Request this service

Testimonials

  • Jaynal is a complete professional. He helped with our projects in a timely and dedicated manner. And was able to find solution for a problem persisting for a long time.

    Muralidharan Sargurupremraj, PhD

Introduction to GENIE HPC Cluster

Genie - Genomics, Epigenomics, Network, Imaging, and Education

 

UTHealth SA faculty members can sponsor their own HPC projects or Storage Allocations, while students, post-doctoral researchers, or research staff must use a UT Health SA faculty member as their Principal Investigator, or PI, as their sponsor.

GENIE cluster consists consists of two head nodes, two login nodes, one storage node (300TB), eighteen compute nodes (20 cores each), two large memory nodes (768GB each), and 10 GPU nodes.

In total, there are 600 Xeon CPU cores, 9 NVIDIA T4 GPU, 9 NVIDIA V100s GPU cards, 12.3 TB aggregate RAM, 57 TB mirrored-SSD as scratch space, and all nodes are connected to an InfiniBand switch device that provides up to 200 Gb/s full bidirectional bandwidth per port (2-4 ports each node).

Additional Information

103000+

Average compute hours per month

101500+

Average Jobs per month

  • Eligibility

    Who Can Use GENIE?

    In July 2020, GENIE, the computational infrastructure for Genomics, Epigenomics, Network, Imaging, and Education was collaboratively funded by the Glenn Biggs Institute and the Greehey Children’s Cancer Research Institute (GCCRI), was established.

    Currently, GENIE users are associated with these two institutions.

    To find out more details, please contact IMS regarding user access and sponsorship.

    There are two types of user accounts;

    1) Users affiliated with HPC Cluster Sponsoring Institutes: GCCRI & BIGGS 

    2) Subscribed University Users

    At this time, we are not creating user accounts for people outside of the university community.

     

    Research Faculty and Staff or Students/Residents/Fellows with Sponsor Approval

    You will need to provide the name of your faculty advisor or sponsor (PI-Principal Investigator) to make a request.

    Once the request is submitted, an email notification will be sent to your faculty advisor  or sponsor requesting approval of your request.

    You will also receive an email notification for your SR request. 

  • Service Level Agreement

    IMS staff will provide support for services during regular operating hours from Monday – Friday, 8 am – 5 pm, with the exceptions of university holidays. HPCF staff will make a reasonable effort to ensure systems and services remain available at all times. All requests should be submitted to My Service Center (IMS TeamDynamix) Service Request. When the user has submitted a service request or trouble ticket within the Standard and IMS support scope, the staff will acknowledge, respond to the request based on the severity level. 

     

    Response time is the time it takes to acknowledge users' issues. It is measured from the time an incident or request is created until the time that the user is advised their incident or request has been received and is being addressed.

     

    Resolution Time is the time it takes to resolve users' issues or fulfill their requests. It is measured from the time an incident record is created until the time that the user is advised their problem or request has been solved. An increase in workloads and complexity in installation, configuration, and troubleshooting may result in an increase in resolution time.

  • Node Details
    Node TypeGeneral Compute NodeHigh Memory Compute NodeGPU Node T4GPU Node V100s
    ModelDell C6420Dell C6420Dell R740XDDell R740XD
    QTY1824 nodes / 2 x T4 1 node / 1 x T44 nodes / 2 x V100s 1 node / 1 x V100s
    Processor(2) Intel Xeon Gold 5215 2.5G, 10C/20T, 10.4GT/s, 13.75M Cache(2) Intel Xeon Gold 5215 2.5G, 10C/20T, 10.4GT/s, 13.75M Cache(2) Intel Xeon Gold 5215 2.5G, 10C/20T, 10.4GT/s, 13.75M Cache, Turbo, HT (85W) DDR4-2666(2) Intel Xeon Gold 5215 2.5G, 10C/20T, 10.4GT/s, 13.75M Cache, Turbo, HT (85W) DDR4-2666
    GPUN/AN/ANVIDIA(R) Tesla(TM) T4 16GB Passive, Single Slot, Full Height GPUNVIDIA Tesla V100S 32GB, 260W, Passive, Double Wide, Full Height GPU
    Memory(12) 32 GB DIMMs(12) 64 GB DIMMs(12) 32 GB RDIMM, 2933MT/s(12) 32 GB RDIMM, 2933MT/s
    Network card(1) Dual Port 10 GbE SFP+(1) Dual Port 10 GbE SFP+(1) Intel X710 Quad Port 10GbE SFP+, rNDC(1) Intel X710 Quad Port 10GbE SFP+, rNDC
    InterConnect: InfiniBand (100 Gb/s)Mellanox ConnectX-6 Single Port HDR100 QSFP56 Infiniband Adapter, PCIe Full HeightMellanox ConnectX-6 Single Port HDR100 QSFP56 Infiniband Adapter, PCIe Full HeightMellanox ConnectX-6 Single Port HDR100 QSFP56 Infiniband Adapter, PCIe Full HeightMellanox ConnectX-6 Single Port HDR100 QSFP56 Infiniband Adapter, PCIe Full Height
    Hard Drive(2) 1.9 TB SSD SAS Mixed use 12Gbps(2) 1.9 TB SSD SAS Mixed use 12Gbps(2) 1.9 TB SSD SAS Mixed use 12Gbps(2) 1.9 TB SSD SAS Mixed use 12Gbps
  • Installed Software Apps

    Software list here

Virtual Desktop Services

VMWare Horizon View Desktop

VMWare Horizon View

The VDI Service provides remote access to users from anywhere and from any device to a full Windows desktop, which is hosted in the Health Science Center data center.

More Information

  • Service Options
    • RDSH Published Application - Request access to University Published Applications like Chrome, FireFox, and Edge or request a client-based application be published using RDSH technology (minimum 50 concurrent licenses)
    • Virtual Desktop Lite - Random/Non-Persistent – Virtual Desktop Lite service provides users with remote access from anywhere on any device to a full Windows desktop hosted in the data center. 2 vCPU / 6 GB RAM / 5 GB Profile Space
    • Virtual Desktop Standard - Static/Persistent – Virtual Desktop Advanced service provides users with remote access from anywhere on any device to a full customizable Windows desktop hosted in the data center. 2 vCPU / 6 GB RAM / 20 GB Profile Space / 20 Hard Disk for Applications
    • Virtual Desktop Power User - Static/Persistent - Virtual Desktop Power User service provides users with remote access from anywhere on any device to a full customizable Windows desktop hosted in the data center. 2 vCPU / 8 GB RAM / Localized Profile Space / 100 GB Hard Disk for Applications
  • Benefits and Features
    • Flexibility - Run legacy and homegrown applications from any device; run Windows-only applications on non-Windows devices; specialized software can be added, updated, or removed quickly
    • Mobility - Access from anywhere, any device
    • Availability - Redundant power, network, and systems datacenter
    • Security - Allow secure access to sensitive data; remove content from the endpoint device; reduce threats from theft or compromise from the client device
    • Manageability - Simplify desktop support and management; manage automated upgrades, patches, and version control
    • Extend the life of older hardware - Users don't need to buy their own copies of software to be installed on their devices
  • Desktop Protection

    Type 2 virtual desktop backups are automated. Backups are kept 3 days, and restoring a desktop from backup returns the desktop to any previous state in the past 3 days. Desktops that have been cancelled and deleted cannot be restored from backup.

  • Disaster Recovery (DR) and Redundancy

    Redundant VDI is available at the secondary data center site to facilitate virtual desktop provisioning in the event of an outage of the primary data center. New DR-ready virtual desktops can be deployed onto end-user devices so users can continue working, keeping the core business alive.

Azure

Azure Virtual Desktop (AVD) service provides remote access to users from anywhere and from any device to a full Windows desktop, which is hosted in the Microsoft Azure Cloud.

More Information

Computing Services

UT Health IT System Operations and Administration offers managed servers, hosted server environment, virtual servers, physical servers, and server software installation services.