Predicting batch queue job wait times for informed scheduling of urgent HPC workloads
This work addresses the challenge of scheduling urgent HPC workloads for disaster response, though it is incremental as it applies machine learning to an existing bottleneck.
The paper tackles the problem of predicting job wait times in HPC batch queues to schedule urgent workloads, achieving predictions within one minute of actual start times for 65-76% of jobs and showing 3.8 to 18 times better accuracy than Slurm.
There is increasing interest in the use of HPC machines for urgent workloads to help tackle disasters as they unfold. Whilst batch queue systems are not ideal in supporting such workloads, many disadvantages can be worked around by accurately predicting when a waiting job will start to run. However there are numerous challenges in achieving such a prediction with high accuracy, not least because the queue's state can change rapidly and depend upon many factors. In this work we explore a novel machine learning approach for predicting queue wait times, hypothesising that such a model can capture the complex behaviour resulting from the queue policy and other interactions to generate accurate job start times. For ARCHER2 (HPE Cray EX), Cirrus (HPE 8600) and 4-cabinet (HPE Cray EX) we explore how different machine learning approaches and techniques improve the accuracy of our predictions, comparing against the estimation generated by Slurm. We demonstrate that our techniques deliver the most accurate predictions across our machines of interest, with the result of this work being the ability to predict job start times within one minute of the actual start time for around 65\% of jobs on ARCHER2 and 4-cabinet, and 76\% of jobs on Cirrus. When compared against what Slurm can deliver, this represents around 3.8 times better accuracy on ARCHER2 and 18 times better for Cirrus. Furthermore our approach can accurately predicting the start time for three quarters of all job within ten minutes of the actual start time on ARCHER2 and 4-cabinet, and for 90\% of jobs on Cirrus. Whilst the driver of this work has been to better facilitate placement of urgent workloads across HPC machines, the insights gained can be used to provide wider benefits to users and also enrich existing batch queue systems and inform policy too.