When creating a btrfs filesystem on multiple SSDs, using the single -m option will ensure that the metadata is not duplicated. On an SSD, duplicating metadata is thought of as a waste of space and has an overhead that can lessen the life of the disk, as shown in the following code:
# mkfs.btrfs -m single /dev/sdb
The second way is to use the ssd mount option. This option will set a few performance options:
Allows large metadata clusters
Allows more sequential data allocation
Disables leaf writing to match key and block order in the b-tree database
Commits b-tree log fragments without batching multiple processes