Yeah, the magma is just going to plop out of the final pump with exactly the same horizontal momentum as from any of the other pumps, i.e. none at all. It WILL move magma horizontally significantly faster than allowing it to flow naturally, but a single pump at the beginning of the pipe would work just as well without all of the extra pumps.
The horizontally-aligned pumps, on the other hand, WILL increase the "velocity" at which the magma moves (since if each pumps' input is filled, each will teleport one tile of magma per clock tick, adding up to, say, 10 tiles of magma per tick for a 10-wide horizontal strip). The main challenge with this is keeping the inputs fed, as magma does not pressure-path like water does unless pumped (and if you pump it to pressurize it, now you have to keep THOSE pump inputs filled). Even with water keeping them supplied would be difficult, as even a river will not keep the pump inputs permanently filled (I used something similar for the high-pressure water supply of an old water logic design).
If I remember correctly, the horizontal pump design was used with a large reservoir of magma to make the original "magma cannon". It mostly just caused magma to spread in a circle from the output very, very rapidly, without much in the way of shooting forward.
That "gauss-cannon" design looks interesting, on the other hand, as it allows for tapping a magma pipe at multiple levels to provide even more magmatic goodness. The main problem you'd run into there is running out of supply magma, like in all the other designs
I still want to work out some sort of obsidian-casting based reusable cave-in trap. Hmm . . .