Using node-local storage
When Slurm starts a job, it creates a temporary directory on each node assigned to the job. It then sets the full path name of that directory in an environment variable called
Because this directory resides on local disk, input and output (I/O) to it
is almost always faster than I/O to a network storage (/project, /scratch, or /home).
Specifically, local disk is better for frequent small I/O transactions than network storage.
Any job doing a lot of input and output (which is most jobs!) may expect
to run more quickly if it uses
$SLURM_TMPDIR instead of network storage.
The temporary character of
$SLURM_TMPDIR makes it more trouble to use than
Input must be copied from network storage to
$SLURM_TMPDIR before it can be read,
and output must be copied from
$SLURM_TMPDIR back to network storage before the job ends
to preserve it for later use.
In order to read data from
$SLURM_TMPDIR, you must first copy the data there.
In the simplest case you can do this with
cp /project/def-someone/you/input.files.* $SLURM_TMPDIR/
This may not work if the input is too large, or if it must be read by processes on different nodes. See "Amount of space" and "Multi-node jobs" below for more.
Executable files and libraries
A special case of input is the application code itself. In order to run the application, the shell started by Slurm must open at least an application file, which it typically reads from network storage. But few applications these days consist of exactly one file; most also need several other files (such as libraries) in order to work.
We particularly find that using an application in a Python virtual environment
generates a large number of small I/O transactions--- More than it takes
to create the virtual environment in the first place. This is why we recommend
creating virtual environments inside your jobs
Output data must be copied from
$SLURM_TMPDIR back to some permanent storage before the
job ends. If a job times out, then the last few lines of the job script might not
be executed. This can be addressed two ways:
- First, obviously, request enough run time to let the application finish. We understand that this isn't always possible.
- Write checkpoints to network storage, not to
If a job spans multiple nodes and some data is needed on every node, then a simple
tar -x will not suffice.
Copy one or more files to the SLURM_TMPDIR directory on every node allocated like this:
[name@server ~]$ srun --ntasks=$SLURM_NNODES --ntasks-per-node=1 cp file [files...] $SLURM_TMPDIR
Extract to the SLURM_TMPDIR:
[name@server ~]$ srun --ntasks=$SLURM_NNODES --ntasks-per-node=1 unzip archive.zip -d $SLURM_TMPDIR
Extract to the SLURM_TMPDIR:
[name@server ~]$ srun --ntasks=$SLURM_NNODES --ntasks-per-node=1 tar -xvf archive.tar.gz -C $SLURM_TMPDIR
Amount of space
At Niagara $SLURM_TMPDIR is implemented as "RAMdisk", so the amount of space available is limited by the memory on the node, less the amount of RAM used by your application. See Data management at Niagara for more.
|cluster||space in $SLURM_TMPDIR||size of disk|
The table above gives the amount of space in $SLURM_TMPDIR on the smallest node in each cluster. If your job reserves whole nodes then you can reasonably assume that this much space is available to you in $SLURM_TMPDIR on each node. However, if the job requests less than a whole node, then other jobs may also write to the same filesystem (but not the same directory!), reducing the space available to your job.