Closed MTCam closed 7 months ago
Currently with this PR, the following examples have the warning (Falling back to a slower transformation strategy as some loops are uninferred which mesh entity they belong to.
) still:
Most of the examples were fixed by the changes we made to componentwise_norms
and compare
functions in simutils
.
heat-source switched to use the alrready-fixed componentwise_norms
instead of actx.np.la.norm
advection_diffusion_reaction
, ablation-workshop
are fixed after fixing up 1d normal returns in grudge/geometry/metrics.py
with axis tags.
blasius
is working after local_dt
fixed to do something different in simutils
to avoid the fallback compile.
sod
fixed by avoiding calls to h_min_from_volume
.
Debugging session for fallback warnings in mirgecom
Questions for the review: