How to confirm SSD cache is being used?

Install, Configure and Update
1

I’m installing Cryosparc 4.4.1 in a SGE cluster. Our configuration of SGE hands out a variable called TMPDIR with name of a path on a local SSD on each compute node. TMPDIR changes for each job.

I followed the instructions for dynamic SSD paths and added this to config.sh:

export CRYOSPARC_SSD_PATH="$TMPDIR"

How can I confirm that the SSD cache is actually being used? I see this in job.log when doing 2D classification:

Transferring across 2 threads: (0/20)
Transferring across 2 threads: 001865614607949450149_14sep05c_c_00003gr_00014sq_00006hl_00002es.frames_patch_aligned_doseweighted_particles.mrc (1/20)
Transferring across 2 threads: 000691201326874248749_14sep05c_c_00003gr_00014sq_00005hl_00005es.frames_patch_aligned_doseweighted_particles.mrc (2/20)
<snip>

I’m 95% certain the above lines mean the SSD is being used, but I’d love official confirmation.

2

You could add commands that inspect the properties and contents of $TMPDIR before and after the
{{ run_cmd }} line of your cluster script template.

3

Coincidentally, that occurred to me immediately after making this post. Should have thought of it sooner.