The University of Arizona
    For questions, please open a UAService ticket and assign to the Tools Team.
Page tree
Skip to end of metadata
Go to start of metadata

Our Mission


UA High Performance Computing (HPC) is an interdisciplinary research center focused on facilitating research and discoveries that advance science and technology. We deploy and operate advanced computing and data resources for the research activities of students, faculty, and staff at the University of Arizona. We also provide consulting, technical documentation, and training to support our users.

This site is divided into sections that describe the High Performance Computing (HPC) resources that are available, how to use them, and the rules for use.




Quick News




Ocelote was built on CentOS 6, and sometimes the kernel for that version is too old, especially for GPU optimized work.  We have reprovisioned a number of the P100 nodes with CentOS 7.  To use these nodes, the PBS select statement will look like:

select=1:ncpus=28:mem=224gb:np100s=1:os7=True

You can change the other attributes like the number of cores, but os7=True must be present.  The default for the GPU nodes is os7=False
More information is here or can be found here



It is time to move from Python2 to Python3.  We will not be updating the Python module version 2.7.14, but it will not be removed until required, for example by a security exposure.


Anaconda is now available as a module.  It expands the capability of Jupyter with Jupyter Labs; includes RStudio, and the Conda ecosystem.  We recommend you access it through OnDemand and the Ocelote Desktop.  See these instructions.






Quick Links


Calendar

DateEvent

Maintenance downtime is scheduled from 7AM to 6PM on April 29 for ALL HPC services

Maintenance downtime is scheduled from 7AM to 6PM on January 29 for ALL HPC services

Maintenance downtime is scheduled from 7AM to 6PM on October 30 for ALL HPC services

Maintenance downtime is scheduled from 7AM to 6PM on July 31 for ALL HPC services.

We will be patching all the HPC systems and Upgrading R to version 3.6.1 during this time.



Debug queue is added to support testing code or trying script options. It has higher priority but short limits.

46 nodes with Nvidia GPU's are available for standard and windfall use.

The 2012 systems (cluster, smp, and htc) have been powered off as scheduled.

We offer a new web portal called Open OnDemand which includes Jupyter notebooks and a nifty file browser
 
Grand Opening of Ocelote 3pm
 
 Our new cluster is delivered.
Installation will take most of the week
  • No labels