Out of idle curiosity I just measured my current machine vice when clamping a 3" wide parallel in the jaws; the jaws are 6" wide. With the parallel being clamped on the top 1/4" or so of the jaws:
'Lift' of movable jaw: ~0.01mm
Movement of top of fixed jaw: ~0.04mm
With the parallel sitting at the bottom of the vice jaws movement of the top of the fixed jaw was less than 0.01mm. In all cases I tightened the vice as much as I could by leaning down on the supplied handle, probably 50ft-lbs or more. According to the manual that corresponds to a clamping force of around 5000lb.
My vice is sold as having minimal jaw lift. It is certainly way better than my previous secondhand machine vice, which had a jaw lift on the order of 0.4mm.
The downside is that my current machine vice (Kurt) is heavy (around 80lbs), large even for a Bridgeport mill and expensive. There are cheaper, and smaller, (Glacern) and more expensive (Orange) lookalikes available in the US, but I have no experience of them, and don't know if they have UK distributors.
If I want to leave parallels in place I tighten the vice a little, tap the work down and then fully tighten. The parallels are then immovable. If I want to remove the parallels, for thru drilling for example, I nip the vice up, remove the parallels and then fully tighten.
I only use the round bar against the movable jaw method when clamping work with one machined and one rough surface. The rough surface being the one against the movable jaw. Using a round bar in this case ensures that alignment of the work is determined by the machined surface against the fixed jaw, not the movable jaw against a rough, and possibly non-parallel surface. Using the method against a machined surface marks said surface.
Andrew