Mesh options¶
Option |
Description |
Default |
|---|---|---|
cap_Bp_ylow_xpoint |
Fudge to get rid of minimum in Bpxy.ylow field, because it can cause large spikes in some metric coefficients, which may cause numerical problems in simulations |
False |
curvature_smoothing |
Smoothing for components of bxcv output: None - no smoothing; ‘smoothnl’ - non- linear smoothing using the algorithm from IDL hypnotoad. |
None |
curvature_type |
Expression used to calculate curvature operator ‘bxcv’ |
curl(b/B) |
follow_perpendicular_atol |
Absolute tolerance for following Grad(psi) |
1e-08 |
follow_perpendicular_maxits |
Maximum iterations when following perpendicular gradients |
10000 |
follow_perpendicular_recover |
Recover from follow_perpendicular failures? This will generate points that don’t follow flux surfaces |
False |
follow_perpendicular_rtol |
Relative tolerance for following Grad(psi) |
2e-08 |
geometry_rtol |
Tolerance for checking identities on calculated geometrical quantities (for example the Jacobian) |
1e-10 |
number_of_processors |
Number of processors to use for parallelisable loops. Note overhead of the parallelisation is fairly high (probably due to cost of pickling psi interpolation function, etc.), so increasing the number of processors above 1 is only useful for relatively large grids, or large values of finecontour_Nfine. This option may even slow things down on desktop/laptop computers which can boost clock-speed for single-core operation. |
1 |
shiftedmetric |
Is grid generated for paralleltransform=ShiftedMetric? |
True |
wall_point_exclude_radius |
When adding the wall point to a contour, this is the radius within which existing contour points are removed as too close to the wall point (points being too close might lead to bad interpolation). This parameter should not usually need to be changed. |
0.001 |