Readme.md

Computing in the Department of Statistics at UConn.

• Cluster
  • Links
  • Quick start
  • Developer tools
• Miscellany
  • Off-campus network access
  • Linux workstations
  • Command-line access to UConn SFTP

Other resources

• The Department website has additional information about the cluster and computing resources in the department.
• Wenjie Wang has made some examples for the cluster.
• Henry Linder has made some useful default documents, including Rmarkdown templates and a boilerplate conference website.

Cluster

The subdirectory cluster/ contains resources related to the Statistics Department's Beowulf compute cluster. The cluster is maintained by the Physics Department at UConn.

Purchase of the cluster and related software was partially supported by NSF Scientific Computing Research Environments for the Mathematical Sciences (SCREMS) Program grant 0723557 to M.H. Chen, Z. Chi (PI), D. Dey and O. Harel.

Links

• Basic cluster load summary
• Comprehensive load summary
• Wiki
• Department webpage about the cluster
• HTCondor distributed computing scheduling software used on the cluster

Quick start

The subdirectory cluster/example/ contains a fully-functional example compute job. notes.pdf contains notes and resources about general use of, and access to, the cluster.

Developer tools

If you require access to a more recent version of gcc (e.g., for the package RcppArmadillo), it may be necessary to run scl enable devtoolset-3 bash prior to compiling.

Miscellany

Off-campus network access

Some computing services at UConn are only available from on the UConn campus, including the Statistics cluster and University-hosted SFTP web storage.

To access the University's network from off-site, University affiliates can:

• Use the VPN to log in with a NetID;
• Use the University's Skybox system;

Members of the Department of Statistics may also SSH into the research lab Linux workstations.

Linux workstations

The research lab contains several Linux workstations that are available for use by members of the Department of Statistics. The workstations provide shell access to a Debian-based scientific computing environment, and are accessible via SSH from off-campus.

Contact Professor Yan for access and policies.

Command-line access to UConn SFTP

To access UConn-hosted SFTP storage, first set the environment variables LFTP_USER, LFTP_PASSWORD appropriately. Then, copy all files in the local directory /local/path/ to the remote directory /remote/path/ with lftp, e.g.,

$ lftp -e "mirror -Rv /local/path /remote/path; quit;" sftp://host:22 -u $LFTP_USER,$LFTP_PASSWORD

For UConn SFTP servers, host is something like name.uconn.edu. Some UConn services are exclusively available from the UConn network. The University's allows users to log onto the campus network. Members of the Department of Statistics can also