What is the XSEDE Queue Prediction tool, and how do I use it to estimate when my job will run?

The Queue Prediction tool estimates wait and start times for jobs submitted to batch queues on Extreme Science and Engineering Discovery Environment (XSEDE) digital compute services. Accessible via the XSEDE User Portal, the Queue Prediction tool uses Karnak, a prediction service developed by Texas Advanced Computing Center (TACC), to calculate predictions based on your job's resource requirements (e.g., number of processing cores and amount of wall-clock time) and the present status of the desired system and queue.

To use the Queue Prediction tool to estimate how long your job will wait and when it will run on one or more XSEDE digital services:

  1. Access the tool here; alternatively, in the XSEDE User Portal, on the Resources tab, click Queue Prediction.
  2. Under "Job Information", enter your job's resource requirements (number of processing cores requested, wall-clock time needed) and a desired confidence interval size.
  3. Note:
    Although the input field is labeled "Confidence interval size", the field accepts only percentage values (i.e., confidence levels). The resulting predictions are expressed in confidence intervals of plus or minus a certain number of minutes for the confidence level you specified. For example, for the default confidence level of 90%, you may get a start time prediction that looks like this:
      START TIME                            90% CONFIDENCE
      Fri, 16, Dec 2016 18:00:00 GMT         ±00:30:00
    This result indicates you can be 90% sure your job will start some time between 5:30pm-6:30pm GMT (i.e., 6pm GMT, plus or minus 30 minutes) on December 16, 2016.
  4. Under "Prediction Options", use the radio buttons to select Wait Time, Start Time, or Both.
  5. For a prediction based on the default queues of available systems, click Predict.
  6. Alternatively, to select specific system(s) and queue(s), click Show System/Queue Options, use the checkboxes to select the desired resources, and then click Predict.

If you need help or have questions, contact the XSEDE Help Desk.

This document was developed with support from National Science Foundation (NSF) grants 1053575 and 1548562. Any opinions, findings, conclusions, or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the NSF.