Half a squirt opposite the intake valve opening, half a squirt closer to it. One of the quirts has a bit more time to evaporate in the intake, and perhaps cool the intake valve too?
It's one of those things that seems to be an issue when you think about it, but in the context of it happening dozens of times per second, the precise timing of the injection really doesn't matter much. What does matter more is that the right amount gets injected.
On the MS3X setup I had on the 16V turbo, I wired it up to do timed per-cylinder injection. Then, messing around with the laptop, I could change how many squirts the engine got per cycle, when those squirts were timed, and even have them all squirt together like the stock ECU does.
And really, back and forth, I really couldn't tell much of a difference. Not at higher RPM's, not at idle (where in theory it should make more of a difference). It just didn't matter. 1 squirt, 2 squirts, batch squirts, individual squirts, timed before the intake opens, timed when it opens, timed opposite. I'm guessing maybe a sensitive tailpipe sniffer might tease out some tiny differences in emissions quality, but in the context of doing a shade-tree tune on a Volvo motor, it didn't matter. Getting that VE table tuned in makes a huge difference, when the squirt/s occur makes practically none.