MJFX singular injectors, perhaps even the larger MJ singular injectors, will fit below the manifold even without a spacer... but in my view there is neat while practical (in terms of future servicing etc) and then there is being slightly OTT neat (dismantling needed to access bits like filters etc, take the top off the engine to get to injectors..). It would take about equal time to fit injectors under the manifold as it does to fit them on top. Some engines might be considered a work of art by some people, other engines probably not... I don't think LPG components installed out of sight in the engine bay of a P38 outweighs positives of easy servicing, I don't think there is a technical advantage in fitting injectors under the manifold on a Rover V8.. If I thought there was a combination of neatness and technical advantage I would do it that way, otherwise the accessibility point wins out.
I believe manifold spacers for these engines are available off the shelf anyway? A sometimes slightly negative point of manifold spacers is that they add volume to the intake manifold, which on some engines (those that use TPS and AFM's rather than a MAP sensor as part of load input to the petrol ECU) will mean acceleration fuelling enrichment will then occur slightly ahead of ECU designer anticipated actual air flow increase past the inlet valves. For sure a negligible effect with a short spacer but one which could offset any negligible benefit from shortening LPG injector to manifold pipe lengths.
Seen a bit of rubbing on sound deadening material with injectors mounted on top, seems to be much worse if sound deadening material sags where it shouldn't.. but while we're talking about neatness we wouldn't want sound deadening material to sag either.
Easy to tell if an engine runs group or sequential injection in LPG software, some systems actually spell out what type of injection the engine runs, or another way to tell is by petrol injector duration - a normally aspirated group engine will idle at around 2ms and full load will be at around 8ms, these are half the usual figures for a sequential engine. Got me thinking now whether I have seen group injection on Gems P38s, I have seen sequential Gems, previous to your post I would have been more certain I'd seen group injection Gems on P38s too, still tend to think I have. As said above it's easy to tell if an individual vehicle runs sequential or group anyway. Most Rover V8's running sequential run with about 4ms pinj duration at idle with warmed engine at just over 0.4bar manifold pressure.
Edit - Last week I had a customer here with a P38, not one of my conversions but one that I repaired a couple of years ago (fitted new injectors on). As usual, the original installer had fitted the injectors atop the manifold without making holes for injector pipes to reach to the correct place in the lower manifold (so injectors fired gas into the upper manifold, several inches from the ideal place). Anyway, on this occasion the car came in because it was running very poorly on gas "It always ran beautiful since you fixed it but recently it went bad all of a sudden while I was driving". Turned out the ECU (AFC, which is really just a rebadged KME Diego) was driving a couple of injectors improperly (all peak current, no hold current). This had caused one of the injector coils to melt - easy to replace the injector coil but would have involved taking the upper manifold off if injectors were fitted under the manifold. Effected a temporary cure while it was here just by fitting another injector coil but warned the owner not to run on gas or at least to keep his eye on the 2 effected coils if he did run on gas (don't want it catching fire) before he returns for the ECU replacement (got a few of the more recent KME Nevo's in stock, Diego are being phased out but I'll get hold of one to make for most cost effective repair). A temporary repair made simple by easy access to components. Should he (unwisely) drive on gas in the meantime, should the coil melt again before I fit another ECU, no big deal changing the coil again (and I won't charge for the coil, got loads of loose ones for these injectors knocking about).
Simon