Repository navigation
test_1063_constrained_freeslip_parallel[ti]: topography moved from its recorded value (37.75 vs 37.08) at np 4 #801
Description
Activity
For the record, this is the issue I was referring to at yesterday's (Sept 29) SC meeting @bknight1 .
- added a commit that references this issue
on Oct 4, 2026 Superseded by #826, which is the same defect measured better.
Both carry the same numbers — recorded anchor
37.07799837159, observed37.752547…— but this issue frames it as partition dependence ("at np 4"), and #826 reproduces it on the recording host, on the anchor's own mesh, with the mesh fingerprint matching (434 cells, vol 2.35568541253) and velocity agreeing to 3e-11.So the np=4 framing here is wrong: nothing about rank count is involved. The topography recovery moved against a velocity field that did not. #826 carries that, and it is the more useful statement of the problem.
Closed in the untouched-issue triage; the work continues on #826.
tests/parallel/test_1063_constrained_freeslip_parallel.py::test_constrained_freeslip_partition_independent[ti]fails at np 4:It fails identically (37.752547901985295) on 05d5248 (
feature/log-conformation, PR #795), so it predates #800; the other two cases of the file pass. The 1.8% move is a change of the transversely isotropic case's answer, not partition dependence (the value is the same run to run). Not bisected.Underworld development team with AI support from Claude Code