Rory Conlin
Rory Conlin
> Currently, the boundary coefficients from the efit_to_desc function and the ones from SangKyeun's file do not match because his coefficients are not from the last surface. @f0uriest I think...
See also #719 for some other related ideas.
> When I try to import the PiecewiseOmnigenity objective, I get the following error: > > ``` > DESC version 0.11.1+555.g6474d1e7b,using JAX backend, jax version=0.4.24, jaxlib version=0.4.24, dtype=float64 > Using...
We probably don't have to find the "perfect" scaling (and there likely isnt one for all cases). We could just allow Psi(rho) to be a profile like the others, defaulting...
Note that DKES is generally not great for eps_eff. Since eps_eff is computed as a limit as collisionality goes to zero, and DKES has problems converging for low collisionality. It...
also need to add the new objective to docs (or do we want to wait until #1290 for that?)
So one option for a unique identifier could be adding a `__new__` method to the `IOAble` base class: ```python def __new__(cls): self = object.__new__(cls) self._id = id(self) return self ```...
I'm a bit confused by the use of gauss-chebyshev above. You factor out the $1/\sqrt(1-x^2)$ but the leftover part still has a $1/\sqrt(1-\lambda B)$ which is still singular at the...
hmmm ok. Once we get the basics working it would be good to compare gauss-chebyshev vs tanh-sinh or modified trapezoid.
Feels really weird that going from m=8 to m=9 makes such a big difference?