Research Computing

Research Computing Resources

A variety of research computing resources are available to all graduate students, and visitors in the Department of Statistics. 

Research Computing Consultation 

Our faculty, Ph.D., postdocs, and sponsored visitors have access to a variety of computing resources to aid in their research.  Questions, inquiries, and consultation meeting requests about any of the services listed below should be addressed to Sr. Research Systems Engineer, Tameek Henderson, at th2974@columbia.edu. Tameek’s walk-in office hours are Monday – Tuesday 0800 – 1700. 

Walk-in (in-person or zoom) office hours on research computing are available every Monday and Tuesday from 8:00 am to 5:00 pm. 

Please refer to the department’s computing page for information on office computing such as printing, email, etc. 

High-Performance Computing (HPC) 

All the affiliates of the Department of Statistics have access to the departmental allocation at the Columbia University Information Technology (CUIT) High-Performance Computing (HPC) shared clusters: Ginsburg & Terremoto. The clusters were facilitated by the Shared Research Computing Policy Advisory Committee (SRCPAC), an advisory group with representatives from all HPC user groups, libraries, CUIT, and the Office of the Executive Vice President for Research. Please contact th2974@columbia.edu with any questions.

Terremoto 

On the Terremoto High-Performance Cluster (HPC), the Department of Statistics has dedicated resources including:

Dell C6420 nodes with dual Intel Xeon Gold 6126 Processor (2.6 GHz)

  • 11 Standard Nodes (192 GB)
  • 5x GPU Nodes with two Nvidia V100 GPU modules
  • 6TB scratch space
  • 5 year lifetime

Terremoto, launched in December 2018, expanded in December 2019. Terremoto cluster documentation may be found HERE

Ginsburg

On the Ginsburg High-Performance Cluster (HPC), the Department of Statistics has dedicated resources including:

All servers are equipped with Dual Intel Xeon Gold 6226R processors (2.9 GHz):

  • 15 Standard Nodes
  • 5 GPU Nodes with two Nvidia A100 GPU modules
  • 20TB scratch space
  • 5 year lifetime

Ginsburg, launched in February of 2021, expanded in March 2022. Ginsburg cluster documentation may be found HERE

For access to the departmental allocation on the shared HPC clusters, Ginsburg & Terremoto, please send a request to th2974@columbia.edu for approval. Research Computing Email List Request Form

Suggestions for training workshops are welcome and should be sent to th2974@columbia.edu

CUIT Research Computing Resources

Foundations for Research Computing Resources