You are here:

XDMoD

 XDMoD - XSEDE Metrics on Demand - is an NSF-funded open source tool designed to audit and facilitate the utilization of the XSEDE cyberinfrastructure by providing a wide range of metrics on XSEDE resources, including resource utilization, resource performance, and impact on scholarship and research. While initially focused on the XSEDE program, Open XDMoD, xdmod.sourceforge.net, has been created to be adaptable to any HPC environment.

Use of CHPC's XDMoD Installation

You can can access CHPC's installation at xdmod.chpc.utah.edu. For more information, please see the documentation available at the xdmod.sourceforge.net site. Note that CHPC is still working on implementation some of the features of the program.

On the front page you can see summary information for the CHPC clusters. By default, you will see the previous month’s usage, but you can change the timeframe being reported by either the pulldown menu or by changing the start and end date and then refresh. This view provides statistics on the number of users running batch jobs and the number of groups to which they belong, the number of jobs, the total number of CPU time (core hours), and the average wait time. 

Using XDMoD, you will be able to delve deeper into the usage. To do so, select the Usage tab.

On this page, we will cover examples of ways to get information using XDMoD.

Example 1: How to get a group usage broken down into per user level.

On the Usage page, select the time frame of interest. On the left hand side there are a number of metric choices; in this panel select “Jobs by PI” and expand the menu. Expanding the "Jobs by PI" results in a new set of choices, including “CPU Hours:Total”. On the same level as the time frame choices there is a filter – open this and select your group. This will give you single entry for the group

xdmod1

 

Then on the blue bar – left click and select drill down by user. By the filter these are display choices – you can do a pie chart, a bar graph etc., as well as plotting this data as a time series of different steps. Note that you can also export the data as a figure or a CSV file which excel can read.

xdmod2

 

If instead of selecting drill down by user, you can  select drill by by resources and get a diagram or spreadsheet that gives the usage based on which cluster (kingspeak, ember, lonepeak, ash.tangent), or select drill down by queue and see usage in terms of queue (kingspeak, kingspeak-guest, owner partitions, etc). You can also do multiple drill downs, for example first by queue and then by user to get a result such as:

xdmod4

 

 

Last Updated: 5/4/17