Rory Conlin
Rory Conlin
We could maybe add a coil optimization tutorial as part of #579 or #726 (or both), and then use those coils for a free boundary example?
Closing this in favor of #1071
[kruger2022generalizedtorsatroncoil.pdf](https://github.com/user-attachments/files/15976043/kruger2022generalizedtorsatroncoil.pdf)
That looks more like a regular scan of a function over an array. The scan trick I was referring to combines scan + select to apply a sequence of functions...
with the usual automatic scaling the variable the optimizer sees is `x / |f'(x)|` this is basically making a new function `g(y=ax) = f(ax)` so the optimizer sees `y /...
> Slightly tangential comment: `deriv_mode="batched"` appears to compute the Jacobian wrt only the optimization variables. `deriv_mode="blocked"` however, appears to compute each objective's Jacobian wrt all parameters related to that objective...
> I think some of our scales are off. Testing with approximate W7-X values of `Psi = 2` and `a = 0.5`, this results in: > > * B =...
> We need to at least add a correction factor to the existing pressure scale on the order of 1e-2, since the thermal energy is typically only a few percent...
Update title and description before merging
> Are black and linting tests failing due to the high number of lines changed or some server error? I have tried to re-run them several times, but they fail...