Hprc banner tamu.png

FASTER:Batch Queues

Jump to: navigation, search

Batch Queues

Upon job submission, Slurm sends your jobs to appropriate batch queues. These are (software) service stations configured to control the scheduling and dispatch of jobs that have arrived in them. Batch queues are characterized by all sorts of parameters. Some of the most important are:

  1. The total number of jobs that can be concurrently running (number of run slots)
  2. The wall-clock time limit per job
  3. The type and number of nodes it can dispatch jobs to

These settings control whether a job will remain idle in the queue or be dispatched quickly for execution.

The current queue structure is: (updated on August 17, 2022).

Queue Name Max Nodes per Job (assoc's cores)* Max GPUs Max Duration Max Jobs in Queue* Charge Rate (per node-hour)
development 1 nodes (64 cores)* 10 1 hr 1* 64 Service Unit (SU) + GPUs used
cpu 128 nodes (8,192 cores)* 0 7 days 50* 64 Service Unit (SU)
gpu 128 nodes (8,192 cores)* 10 7 days 50* 64 Service Unit (SU) + GPUs used

Checking queue usage

The following command can be used to get information on queues and their nodes.

[NetID@FASTER1 ~]$ sinfo

Example output:

 cpu*             up     7-00:00:00   1-32        16/61/13/90      1024/3904/832/5760
 gpu              up     7-00:00:00   1-32        25/5/8/38        844/1076/512/2432
 memverge         up     2-00:00:00   1-2         0/0/2/2          0/0/128/128
 fpga             up     2-00:00:00   1-2         0/2/0/2          0/128/0/128

Note: A/I/O/T stands for Active, Idle, Offline, and Total

Checking node usage

The following command can be used to generate a list of nodes and their corresponding information, including their CPU usage.

[NetID@FASTER1 ~]$ pestat

Example output:

Hostname       Partition     Node      Num_CPU    CPUload    Memsize    Freemem    Joblist
                             State     Use/Tot               (MB)       (MB)       JobId User ...
fc004          cpu*          idle      0   64     0.01       256000     253490

Checking bad nodes

The following command can be used to view a current list of bad nodes on the machine:

[NetID@FASTER1 ~]$ bad_nodes.sh

The following output is just an example output and users should run bad_nodes.sh not see a current list.

Example output:

% bad_nodes.sh 
REASON                                                       USER             TIMESTAMP            STATE        NODELIST
The system board OCP1 PG voltage is outside of range.        root             2022-07-11T14:38:07  drained      fc152
testing DIMM B5                                              somebody         2022-09-01T12:31:22  drained*     fc017
Not responding                                               slurm            2022-09-19T12:10:47  down*        fc001


Checkpointing is the practice of creating a save state of a job so that, if interrupted, it can begin again without starting completely over. This technique is especially important for long jobs on the batch systems, because each batch queue has a maximum walltime limit.

A checkpointed job file is particularly useful for the gpu queue, which is limited to 2 days walltime due to its demand. There are many cases of jobs that require the use of gpus and must run longer than two days, such as training a machine learning algorithm.

Users can change their code to implement save states so that their code may restart automatically when cut off by the wall time limit. There are many different ways to checkpoint a job file depending on the software used, but it is almost always done at the application level. It is up to the user how frequently save states are made depending on what kind of fault tolerance is needed for the job, but in the case of the batch system, the exact time of the 'fault' is known. It's just the walltime limit of the queue. In this case, only one checkpoint need be created, right before the limit is reached. Many different resources are available for checkpointing techniques. Some examples for common software are listed below.