Jago Strong-Wright

I've played more with this now and think the way we can do this with the minimum code changes is as follows. The only other way I can see is...

Just to clarify, for a Face field in a Bounded direction the first halo points are at 0 and N + 1, or is in N+2 with 1 to N+1...

> ```julia > using Oceananigans.Grids: total_length > ``` This is strange because if I set the Nx + 2 point as the boundary it doesn't end up working (for face...

From what I've read the reason we might need to not just exactly prescribe inflows is because if the interior solution varies from the prescribed inflow (and possibly outflows) and...

Oh I see, that looks good. My confusion was because `length(Face(), Periodic(), 2) = 2` whereas `length(Face(), Bounded(), 2) = 3`. I guess instead of using `tx`/`ty`/etc. here it could...

This seems correct for Periodic topologies but wouldn't it be the case for Bounded too. Do particles get bounced if you put them just to the right of the N'th...

Oops sorry meant to leave my comment in the PR, will change to there

I have seen particles cause out of bounds memory access errors on GPU before. I think it happens because the interpolation tries to index into the array at an index...

I don't think this is the same as my error linked from before.

> > I have seen particles cause out of bounds memory access errors on GPU before. I think it happens because the interpolation tries to index into the array at...