Per-region DMI on Antiferromagnet: kernel does not respond to spatially varying D, despite correct storage

[sateesh kandukuri]

Dear mumax+ developers,

I am running an Antiferromagnet simulation on a rectangular track with two sublattices and interfacial DMI applied to sub1 only. The skyrmion responds correctly to uniform DMI changes but does not respond to spatially varying DMI, although the per-region values are stored and read back correctly via eval().  I would really appreciate your guidance on the right approach.

Test results:

1. Uniform D = 0.259 mJ/m² via magnet.sub1.dmi_tensor.set_interfacial_dmi(D): skyrmion diameter = 11 nm 
2. Uniform D = 0.388 mJ/m² (1.5× baseline): skyrmion diameter = 53 nm (kernel responds correctly)
3. Two regions defined via regions=numpy_array at construction  (left half = region 0, right half = region 1). Per-region DMI assigned via setInRegion.  Verification via xxz.eval() returns the correct values (0.259 in region 0, 0.388 in region 1). But the skyrmion nucleated and relaxed in region 1 stays at 11 nm instead of expanding to ~53 nm.
4. Tested with callable assignment magnet.sub1.dmi_tensor.set_interfacial_dmi(lambda x,y,z: ...) — same result. 
5. Tested with direct numpy array assignment magnet.sub1.dmi_tensor.xxz = array (shape (1, 1, ny, nx)) — same result.

I tested with minimize(), relax(), tight tolerance (tol=1e-12), and larger initial nucleation radius — all give 11 nm.

What is the correct way to apply a spatially varying interfacial DMI to a sublattice of an Antiferromagnet such that the kernel responds to the spatial variation? Is there a known limitation with per-region DMI in k_dmiFieldAFM, or am I missing a required step in the assignment pattern?

I can share a minimal reproducer script if helpful.

Thank you, 

Sateesh

[Ian Lateur]

Hi,

It could be that you are running into a bug that has recently been fixed. This fix is included in the latest version v1.2.0 (or on the master branch). Could you update your mumax+ installation?

If this does not resolve the issue, a minimal script would be helpful.

Hope this helps!

Ian

Op wo 29 apr 2026 om 13:52 schreef sateesh kandukuri <sateesh...@gmail.com>:

--
You received this message because you are subscribed to the Google Groups "mumax2" group.
To unsubscribe from this group and stop receiving emails from it, send an email to mumax2+un...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/mumax2/dfb7f85f-2aa3-4955-a98d-158dfc4fcce8n%40googlegroups.com.