Welcome to the CCR community!


Please watch this Intro to CCR workshop recording to learn more about what CCR is and how CCR's resources can enable your research or aid your business



If you do not already have a CCR account, please set that up first and then come back here to begin.


There are many different resources that CCR supports.  This guide provides links to articles on the topics most critical for new users to understand in order to get started utilizing the CCR HPC (cluster) resources.  If you're looking for information on logging in and using other resources, please see the following links:

Lake Effect Research Cloud

Globus File Transfer Service


The majority of people use one or more of the three CCR high performance Linux clusters (aka supercomputers).  In order to best utilize the CCR clusters, users should know how to use the UNIX command line shell.  Though we do provide easy-to-use alternatives to traditional command line tools, such as Open OnDemand, it would benefit you to learn a few basic linux commands.  Check out our Intro to UNIX guide and downloadable 'cheat sheets' here



CCR Help:


The CCR help portal provides a searchable knowledgebase for answers to many frequently asked questions as well as how-to articles, event notifications, and system alerts.  If you can't find an answer to your question, please submit a help ticket via this portal or by emailing ccr-help @ buffalo.edu




Virtual Workshops:

In 2020, CCR developed a series of virtual workshops to provide information on a variety of topics. We encourage users to watch the recordings of these workshops prior to attempting to use the systems.




Login Options:

OnDemand - web-based portal providing a single access point for all HPC clusters and storage.  Users can transfer files, access a shell environment on the cluster front-end login server, launch interactive and remote visualization jobs, and monitor jobs all without installing any client software or web plug-ins. 


Other Command Line & GUI Login and File Transfer Options

Graphical Display Setup for Command Line Logins

NOTE: If you want to connect using SSH/SFTP/SCP you must upload SSH keys to your CCR account first and make sure to specify your private key when connecting to CCR.



Storage:


Available Storage Options

How to check your quotas

Data Backups

Data Retention Policy



Software:


CCR staff maintain a very long list of software installations for use on our clusters!  Some of these software products are listed here

However, the majority of software applications installed do not have individually documented articles. We use software modules to load the proper environment for each software application and version.
What are software modules and how do you use them?


If you search the available modules and don't find the software you're looking for, you have two options.  You may be able to install the software yourself in your own home or group project directory.  You would then create your own module to load the proper software environment:


Creating your own personal modules

Creating modules for your research group


If you are unable to install the software yourself or you believe it would be something useful to the general CCR community, you may request an installation by CCR.  Please see our software installation policy for more details.  



Running Jobs:


CCR maintains 3 clusters: UB-HPC (academic) cluster is available for use by UB faculty, their students, and collaborators at no cost.  The Faculty cluster houses nodes purchased by faculty and research groups for their primary use.  The industry cluster is for use by business partners of the university.  Idle nodes on the faculty and industry cluster are available for use by academic users via the scavenger partitions.

 

Computing Resources

Batch Scheduler (SLURM)

Submitting SLURM scripts to the cluster

Submitting an interactive job to the cluster

Are you trying to run on the cluster but your job keeps getting killed?  This may be the reason

Checking the status of your jobs

Monitoring your jobs

How jobs are prioritized in the CCR queue

Viewing job history and accounting

NOTE: When running an interactive desktop app in OnDemand, you are running a job on the cluster.  Checking the status, priority, and monitoring the OnDemand job can be done in the same way as if you ran a job with a batch script.



How-To Guides:


Python environments are one of the most frequently asked about topics by CCR users and they can be complicated!  We provide guides for "Using Conda Environments on the CCR Cluster" and "Using Python Virtual Environments on the CCR Cluster"


Job monitoring guides are also provided for serial, parallel, and GPU jobs.  These are extremely helpful to understand what is happening on the node(s) when your job is running and to provide insight into how to make your jobs as efficient as possible.



When you're ready, head on over to:

Advanced User Topics