The Taylor Geospatial Institute Regional AI Learning System 

The Taylor Geospatial Institute Regional AI Learning System (TGI RAILS) offers a high-performance computing and data analysis system, funded by the National Science Foundation and the Taylor Geospatial Institute, and managed and operated at the National Center of Supercomputing Applications at the University of IllinoisThe system is available for use by the regional consortium of academic institutions of the Taylor Geospatial Institute https://taylorgeospatial.org/.  

TGI is a comprehensive and holistic geospatial research community with eight outstanding research institutions in the Midwest central region as founding members, namely Saint Louis University, the Danforth Plant Science Center, Harris-Stowe State University, Missouri University of Science and Technology, the University of Illinois at Urbana-Champaign, University of Missouri – Columbia, University of Missouri – St. Louis, and Washington University in St. Louis.  

Through partnerships with industry, government agencies, and research and educational entities, TGI advances multiple disciplines including artificial intelligence, data analytics and geospatial science with the targeted goals of producing significant benefits in the areas of food systems, health and social equality, smart and resilient communities, workforce development, national security, and economic development. 

TGI RAILS is a high-performance computational and data analysis resource for researchers and students at member institutions who focus on areas of GIS applications, AI, and Machine Learning. A fraction of the system is available through the Open Science Grid. 

The system offers both CPUs and GPUs, along with low-latency, high IOPs storage on a high-performance network and WAN connectivity. 

Technical Specs

Role - QuantityLogin/Utility - 2CPU Compute - 3GPU Compute - 3
CPU2 x 16c/32t Intel Sapphire Rapids2 x 48c/96t Intel Sapphire Rapids2 x 48c/96t Intel Sapphire Rapids
Memory256GB512GB2TB
Local Scratchnone2TB NVME4 TB NVME
GPUnonenone8 x NVIDIA SXM H100
RDMA Interconnect2 x 100Gbe (25GB/s)2 x 100Gbe (25GB/s)4 x 100Gbe (50GB/s)

Storage: VAST + Local Scratch 

  • 1 PB capacity at 2:1 compression 
  • 40 GB/s read, 5 GB/s write 
  • 338,000 4k random file operations/s 
  • AI focused 
  • Highly resilient 
  • Inline Deduplication 
  • Inline Compression 
  • Low overhead administration 

 

Who can use RAILS 

Geospatial related researchers, faculty, and staff at the TGI member organizations are eligible to use RAILS for geospatial-related research. The system can also be used in courses of any discipline. Up to 20% of the RAILS capability is accessible via the Open Science Grid. https://osg-htc.org/.  

Training 

User training for RAILS  – Coming Soon! 

 Documentation 

https://go.ncsa.illinois.edu/TGIRAILSUserDoc

How to Get Started Using RAILSs 

Getting started on RAILS is easy, simply follow these steps. For assistance, contact your local TGI support team or send questions to help+TGI@ncsa.illinois.edu 

  1. Complete the TGI Data Services Request form on the TGI website, select “Data Processing or Analytics” under “What Service do you need?” and answer the subsequent questions regarding the specifics of your request.
  2. Receive confirmation email from DataServices@taylorgeospatial.org after approval
  3. Follow the instruction in the confirmation email to establish an NCSA account and select the time desired on the system.
  4. Access TGI Rails. Upload data, run scripts, visualize data, and more

 

Helpful tips

Once approved in step #2, you will be prompted to create an NCSA User ID:_https://go.illinois.edu/NCSAAccountRequest 

Tip: It is recommended you use your institutional email address to expedite the process and further confirm you are faculty/staff at a member organization. 

Using your NCSA Login, go to the XRAS allocation system to submit your request.  https://go.illinois.edu/TGIRAILSProjectRequest 

The response time is typically 2-3 working days.  Once your allocation has been approved, you can reference the documentation for more details on how to get started using RAIS.  

https://go.ncsa.illinois.edu/TGIRAILSUserDoc/accounts/