Which format for external disk for performance?

None of the HFS flavors will offer any performance benefit that's measurable. For that reason, go with the default format disk utility gives you and look to optimize elsewhere.

I have seen encryption slow down some storage medium such as slow USB flash and I would expect journaling there to also be more of a slowdown than on storage with fast cache or more responsive write service times.

I suppose journaling could in some rare circumstances (bizarre edge cases) slow things down as you need to journal certain actions, take the action, and then journal that the action is complete. However, the journaling is for rare events like file creation, file deletion, etc... - some metadata operations are journaled, not all data writes like other filesystems allow which can cause a large performance impact.

Journaling does make it easy to roll the filesystem back to a known state if the system crashes or the drive gets disconnected without changes flushed, so you can save hundreds of seconds to several minutes at the next reboot / mount by journaling.

For best performance, keep the drive approximately 70% full at most as portions of the disk are much slower for random access and the OS uses the fast part of the drive first.

So, in order of fastest to slowest speed transfer:

Connection type: Thunderbolt2 > Thunderbolt > Fiber Channel > SAS > USB 3.0 > FW 800/400 > USB 2.0

Storage Medium: Specialized SLC/MLC Flash with high speed controllers > PCI Prosumer SSD > 15k HDD > Consumer SSD > 10k/7.5k HDD > Fast Small FLASH > 5k HDD > Consumer USB flash storage

Journaled for certain. I'd use the encryption option for security sake, but I don't know what the I/O performance trade off is with this setting. That is all.

For better I/O performance from external disks, purchase a RAID. I use 2 Guardian Maximus RAID boxes from OWC which provides a mirrored (redundant disk) RAID1, but they also sell other RAID boxes with faster I/O.

For true I/O performance, nothing beats SSD, though.