I can see you've got a decent bit of work into those, but unfortunately, I believe those will be the bottle neck in your setup. I think you've backed yourself into a bit of a corner actually if you're going for much higher output. If you're using split port heads, you are best off maintaining the split all the way up through the LIM and UIM. The only reason for blending the two like many do, is for convenience at the expense of some power throughout the low and mid-range especially, but I suspect even at the top end due to lack of any coherent helmholtz tuning.
Ideally, you would want to custom make a manifold, but that's a LOT of work. You could take an SVT or even an SE manifold, and have the plenums cut and rewelded to add a little volume, and then have it double extrude honed to open up the runners a bit to match the larger 3L valves. Of course you'd still want to run an IMRC in that case as well.
Otherwise you're better off starting with some 3L heads, and either running higher compression pistons or decking the head a bit if you are dead set on the higher compression. The pistons would be the preferred choice, as I'm not even sure you can deck these heads enough without screwing up cam timing (I've never checked, so I don't know).
If you go with the oval port, then you would want to use one of the later designed manifolds that best take advantage of the oval port. Like the ST220, or the 05+ Escape, or even the Mazda/Fusion manifolds may work. The runner lengths, runner taper, and plenum volumes are much better on these later manifold designs. The 09+ manifolds are even better, but they need the 09+ heads as well to work correctly, and that will get expensive to sort out.
Personally, I would like to see someone go all out with the split-port design. I think it has better potential, and so did the 2.5L world challenge engine builders at Kinetics.