Hello all, Are there any suggested strategies for when to use

use_tmpfs

? I tried using it in hopes that it would speed up s3 downloads, but didn't see any significant improvement. How big does the dataset have to be to see gains?