If you're willing to target recent x86 exclusively, then you can increase resistance by employing the AES-NI instruction.
That is not one but rather a group of instructions. Perhaps you just typoed the missing 's'. One of those instructions was of particular focus of mine.
I wasn't sure if it was one or multiple, but was too lazy to figure out which:-(
Thanks for correcting...
Hmmmm . . . just reading that iPhones have had AES hardware since the iPhone 5S . . . although I'm reading that the AES chip sits between the flash and the main CPU - perhaps meaning it can only be used to encrypt/deprypt data going to/from the flash. If not, perhaps iPhones could be used for efficient mining . . memorycoin, cryptonight, maybe HODLCoin soon enough

Understand that I am working on
a new concept for unprofitable PoW mining, in which it is not critical that that the PoW hashrate for phones be at par with anything; they will still add to the total systemic hashrate which makes my design more secure.
Note that it would be problematic if phones are 100X slower or less efficient on the PoW than desktops.