Hmmm. That makes sense, since the pet AA should be @ 3s and through testing you do a lot less Burning Strikes than AA, when they should start out equal at base SPD.
That could explain strange results I was seeing when testing @ 354 SPD and 354 CRT (since there was less artificial delay), and the fairly consistent results I had with a higher SPD (started to smooth out that delay somewhat)
So really, we should be seeing for Ifrit that his AA should be going off at the same rate as Burning Strike at base SPD, and have greater gains than it currently does, instead of SPD alleviating that delay somewhat for BS to only catch up to AA frequency.
Definitely sounds like Bug Report time.