fixing alfven and inchworm #58
Conversation
|
@brryan if you're happy with this, let's merge in. |
|
actually, I don't think this is quite working for snake yet. Janiris and I threw something in there, but I think something is off with the snake transformation of the b field. was hoping one of you guys could spot what we did wrong. |
|
Ah ok. I'll look more carefully. At a first glance I didn't see anything wrong. |
src/pgen/linear_modes.cpp
Outdated
| for(int d = 0; d < 3; d++){ | ||
| Bdotu += v(ib_lo+d, k, j, i) * ucon[d+1]; | ||
| } | ||
| Real bcon[NDFULL] = {Bdotu, 0.0,0.0,0.0}; |
There was a problem hiding this comment.
b^0 = B.u/alpha? Also B.u = Gamma B^i v_i but this looks like B.u = B^i u^i?
There was a problem hiding this comment.
Good catch. I think this is likely the last thing needed to get this working.
There was a problem hiding this comment.
aren't those the same in this case? alpha = 1 and u^i = Gamma v_i since this is Minkowski.
There was a problem hiding this comment.
Is this going to eventually be used for snake/inchworm? The alphas are being used when calculating the b^i just below
There was a problem hiding this comment.
this is just the pre-transformation stuff though, isn't it? like no matter what this will always be Minkowski.
There was a problem hiding this comment.
Right sorry yeah I see this is pre-transformation. OK this should be fine... hm...
There was a problem hiding this comment.
Oops right. So this isn't the bug. I need to sit down with this and go through it with a finer comb.
|
I think this works now, but we should do a convergence study of snake alfven and something in inchworm to confirm before merging. |
|
I would like to see convergence of just the entropy mode in snake and inchworm as well |
|
Big effort to dig into this, glad to see it resolved. |
:)